|Farnsworth Peak||146.62 MHz||Operating normally.|
|Scotts' Hill||146.62 MHz||Operating at slightly reduced transmitter power|
|Lake Mountain||146.76 MHz||Operating normally. IRLP node 3352 available to UARC members.|
|UARC HF Remote (transceiver)||HF frequencies||Nominal|
It was reported that the 146.76 repeater on Lake Mountain had almost stopped transmitting during a QSO on Tuesday afternoon, June 15. We'll say “almost” because those in Utah Valley with quiet receiving locations, good antennas, and an unobstructed path to the repeater site, reported faintly hearing the repeater transmitter. But for most purposes, the repeater was out or service.
Suspecting the repeater's outboard GE power amplifier (“PA”), Clint, KA7OEI, had obtained a replacement for the transmitter's own final amplifier. It turned out that Glen, WA7X, had ordered a PA module for an Icom transceiver, hoping to convert it from a medium-power version to the 50-watt version. However, on tearing into the unit to make the modification, he discovered that a prior owner had already made the change and no further effort was required. Fortunately it turned out that his module had the same pinout as the original module in the 76's transmitter, which is a Yaesu Ft-2500.
Clint and the author, K7HFV, headed to the site on Saturday, June 20, and found that, indeed, the outboard PA was the problem. Clint was able to install Glen's module and soon there was a full eight watts going into the transmit feedline. We called the trip a success and headed home.
Clint, KA7OEI, and the author, K7HFV, were free to investigate the latest problem on Monday May 4. We weren't entirely surprised that the tone/weak signal problem had returned. On the previous trip the problem seemed to fix itself when we moved the equipment around a little, so we were never sure of the cause.
On the way up to the mountain we checked how the signal sounded when receiving in AM mode. It was much louder than in FM mode! It clearly had a strong AM component. This made it much more likely that the problem was in the power amplifier (PA) than in the audio processing or the exciter.
When we arrived we discovered that the problem was in a network that is part of the PA's output power regulation system. We certainly did not have a spare network module in hand or were even sure where a new one could be obtained. (The GE PA had not been manufactured for quite a number of years.) Some experimentation showed that a reasonable output power (~15 watts) could be obtained if the pass transistor that ultimately regulates the output power were driven with five volts DC. We hunted for some time for a combination of resistors that would give us five volts but simply didn't happen to have the right combination with us. Then Clint discovered in his gear an even better solution: a 78L05 low-power five-volt regulator! He put the regulator in place and all appeared well.
This stopgap solution does not really regulate the power output as the normal configuration does. It does not sense the output power and adjust it up or down as appropriate. However, we left it in a relatively safe zone for the amplifier, and expect that some modest changes with temperature won't be problematic. We are still pondering on what a good long-term solution would be.
We received a report fron Noji (KN0̷JI) that the tone had returned on the 146.76 repeater. A check showed that the weak transmit signal had also returned.
For several weeks a continuous tone, somewhere in the 800 Hz neighborhood, had been appearing along with the regular modulation on the 146.76 repeater's output. It couldn't be heard on the input or link frequencies, and appeared to be getting generated in the repeater itself. On Friday, April 24, Clint, KA7OEI, and the author, K7HFV, were free to make a trip to the mountain when the weather looked good and there was a good chance that the southern (Mercer Canyon) approach to the site might be passable.
We made it to the site with no trouble and no snow. For a first check, Clint bypassed the outboard power amplifier and keyed the transmitter. The tone was gone! We tried several configurations with and without the power amplifier (PA) and with and without the normal transmit audio conencted and could not get the tone to return. It was nice to have the problem solved, but less than satisfying not to know the cause and not to be able to guarantee it would not return.
Some possible culprits might have been RF getting back into the transmitter, the transmitter (a Yaesu FT-2500) sending a warning tone, some sort of rectification in the unintentional connection caused by the transmitter leaning against the rack wall. The tone seemed to match the frequency of a warning tone that the FT-2500 puts into its speaker when the operator pushes a front panel button while the keyboard is disabled. We may never know the real cause, but just moving the transmitter a few inches cured it.
Having brought lots of test gear along, we thought it appropriate to
make some basic performance measurements:
|Power output into transmit cavity||14.1||Watts|
|SWR on transmit feedline||1.3:1||—|
|Two-meter Receiver sensitivity measurements for 12 dB SINAD:|
|Via ISO-T w/antenna connected||-62||dBm|
|Via ISO-T w/dummy load||-75||dBm|
|Result: degredation due to site noise||13||dB|
|Link Receiver sensitivity measurements for 12 dB SINAD:|
|Receiver through cavity||-99||dBm|
|Via ISO-T w/dummy load||-53||dBm|
|Via ISO-T w/antenna connected||-50||dBm|
|Result: degredation due to site noise||3||dB|
Epilogue: Upon arriving back in the valley we discovered that not only
had the tone gone away, but the repeater's transmit signal strength had also
increased considerably. Our best guess now is that the cause of both problems was
a bad connection between the exciter and the power amplifier (PA). (The very first
thing we did on arrival was to disconnect that pair.) One theory is that
that may have caused a problem with the PA's power regulator and had it go into oscillation.
Another is that RF back into the exciter might have caused it to send a tone that
is normally a warning of a button being pushed when the panel controls are disabled.
(The frequency seemed to match.) We may never know the exact details.
There was a power bump on Farnsworth peak during the evening of 1 March, 2020 and it appeared to have reset the '62 repeater's squelch controller. Despite attempts to remotely control the squelch (Gordon at his house, Clint at his house, Gordon and Clint at Clint's house) we decided that something else was wrong and called the mountain.
Scotts was still working: The COS input on the voter from the local Farnsworth receiver (input #1) timed out after 300 seconds (5 minutes) — longer than the repeater's own time out of about 3:07. It was occasionally possible to reset the COS (probably noise/flapping on the COS line) so the repeater would work for 3 minutes — and then the old controller would time out: After about two more minutes the Voter COS would time out and it would again start working via Scotts (only). It was noted that when telemetry was attempted to be gathered from Farnsworth that all devices (Disciplined oscillator, Voter) would only get a "DEV TO" indicating the loss of serial connectivity.
At about 1215 on the 2nd, we spoke with Ron on the mountain who kindly went over to the Buddha room and verified that the squelch setting on the repeater controller was “1.” Instructing him to press the “Squelch” button (upper right-hand corner) the “Squelch Enable” button did not light up after several attempts of holding the button.
I (Clint) instructed him to turn off the power supply and after about 60 seconds of him telling about the power issues (power bump, breaker after the transfer switch popped, having to “wind up” the transfer switch and restoring power after about an hour or so) he turned it back on and he was able to reset the squelch: We used a setting of “24” (top three lights on) and the repeater seemed to be functional again.
Thanking Ron (he likes cookies) we again tried to control the repeater (sending the “OFF” tone) but were unsuccessful indicating a problem with the control receiver and/or logic.
— We remember that on the last Farnsworth trip (September 2019?) that we could hear no audio coming out of the control RX speaker. Our collective memory is fuzzy when it comes to trying to remember if we verified with the signal generator that the receiver was still working (e.g. audio amp problem) or what.
— Based on Ron's being unable to reset the squelch without power-cycling the controller, it is likely that we would not have been able to remotely control it, anyway.
Due to apparent power failure, Farnsworth squelch was operating fully open. This kept the repeater timed out most of the time, and virtually unusable.
Clint, KA7OEI, was able to procure a power amplifier module for the Scotts' site. He bypassed the final amplifier stage. Keith, KE1THR, and Sam, KJ7AXQ, skied up to the site and installed the temporary module and left with the transmitter putting out about 14 watts. A report from K7DOU in Park City indicated the signal was again usable there. They brought the ailing PA down for repairs.
High minimum bus voltage from Scotts' and reports from Park City indicate that the Scotts' transmitter is off the air. However, signals heard on its receiver are still relayed to Farnsworth.
On the Friday after the power amplifier repair, Keith was once again able to make a trip to Scotts’ Hill site. He reinstalled the PA but was dismayed to find that power readings were down in the milliwatts. Even the exciter output read such low power output that it was suggested the problem might have been in the exciter all along.
Perhaps the problem was just in the metering. Some alternate ways of determining whether the PA was putting out power were suggested. There was power supply current drawn during transmit, and the ability to put noticeable heat into one of the dummy loads in the isolator (a device that lets power go from the transmitter to the antenna, but not from the antenna to the transmitter.) Both methods indicated the something more than milliwatts was making its way out of the PA. With that much assurance, Keith found there was nothing further that he could do and left the site.
It would have been more satisfying if someone in a primary coverage area of Scotts’ had been on the air and could have reported if the signal sounded normal in strength. Then, by happy coincidence, we heard from Gary, KK7DV, in Wellsville, Cache County. Gary lives in one of the few areas where the signals from Farnsworth Peak and Scotts’ Hill arrive at about the same strength. Gary reported he could hear the beat note between the two carriers! This was particularly good news. Because of the FM “capture effect”, there is only a narrow range of signal strength difference between two signals that will produce a beat note. That transmitter had to be putting out something close to normal power. The problem must have been in the power meter.
Well, that was what we thought. When Keith got home to Brighton he checked the meter using his handie-talkie and dummy load. It seemed to be working just fine. So why had he not been able to read the 25 watts that the transmitter had be set to? The best guess was that there was a problem with one of the jumper cables that had been used in the measurement. In any case, Gary's report assured us that the transmitter was operating normally.
On the Monday following the 5-12 trip, Keith made another trip to the Scotts’ site and found that an essential cable, removed to allow PA removal, had not been reconnected to the exciter. Reconnecting the cable brought the repeater's signal back to where it was usable in Brighton and Park City if not in Evanston.
Meanwhile, Clint, KA7OEI, had the ailing PA on a test bench. The final transistor pair had blown and no spares were on hand. It happens that this particular PA is capable of putting out over 100 watts, but we had been running it only in the 20- to 30-watt range. A quick solution to the problem seemed to be, simply, to connect the driver stage through to the board's output. It worked. Measured into a dummy load, the board now put out 25 watts. It was deemed ready to return to the mountain.
Sometime during the week of April 30 through May 6 regular users of the 146.62 repeater on Scotts’ Hill started reporting a weak transmitted signal. Listeners on the Farnsworth half of the synchronous pair reported that the Scotts’ users were getting into the system just fine. They just could not hear it well.
The road to Scotts’ is often not passable for vehicles until sometime in July. We are fortunate to have a ham who can get there using another transportation mode. Keith, KE1THR, is a ski instructor at Brighton and has several times volunteered to ski up to the Scotts’ site when needed. So, to diagnose the problem, on Friday, May 12, Keith went to the site and was joined by Marty, N7XHB, on snowshoes. Possible culprits were antenna, feedline, and power amplifier (PA) module.
The snow was about five feet deep against the door of the building but Keith and Marty were able to shovel enough snow to get in. All looked well on the outside, so antenna and feedline were not likely to be a source of the problem.
Checks with a power meter identified the PA as the problem child, so, following instructions from KA7OEI on the ground, they removed the PA module and connected the GE exciter directly to the isolator/cavity chain. It was expected that the exciter, even though its output was only in the hundreds of milliwatts, would put more signal into the antenna than had the failed PA. Keith and Marty brought down the PA for troubleshooting and repair.
Unfortunately, when folks listened for the Scotts’ signal after the trip, the signal seemed to have gone from weak to non-existent.
On Saturday, October 1, Clint, KA7OEI; and Bryan, W7CBM; made a trip to the Scotts’ Hill site and installed a new exhaust fan, having determined that the old one was beyond repair.
Meanwhile we got word from the club Treasurer that our power consumption seemed to be back to normal, after the August 24 removal of a stalled fan and a bulging battery. Based on two meter readings for the billing cycle and one that we made on our August 24 visit, he calculated the following:
On Thursday, I [Clint, KA7OEI]; Bryan, W7CBM; Randy, a co-worker of Bryan's; and two of their friends went up to Scotts Hill.
There, we managed to get the 1600-ish pound battery bank into the building, first using a frame/gantry to hoist it off Randy's truck at which point we installed the casters and then using a smaller frame/gantry straddling the doorway to move it into the building, over the threshold.
It now occupies roughly the same floor space as the “original” battery bank in the southwest corner of the building. Originally, it had an approximate 1500 amp-hour capacity, but since it is about ten years into its 20-25 year lifetime, it will probably be about 80% of that capacity.
At the time of the same visit I attempted to resuscitate the exhaust fan on the building that had seized up. I was successful in “un-seizing” it, but it is apparent that a winding has shorted, so this fan will have to be replaced — either with one just like it, or with one or more square “boxer” type fans to be mounted on a board that could be placed within the existing filter/cover.
The previous, wooden “battery trolley” is still in the building and I have tentatively offered it to John (K7JL) for use with the set of batteries found in the “other” Scotts' Hill site (e.g. 145.27 and various repeaters).
The “Wednesday Night Hike Group” happened to decide on Scotts' Hill as a likely destination. Upon reaching the radio site (after some of the party had climbed the nearby peak) the party found it only natural to check on the health of the Scotts' repeater.
There was reason to believe something unusual was going on. The power bill for the site had shown higher usage than was normal and telemetry readings had shown unusually high temperatures. Possible culprits were a stalled exhaust fan and a battery with a shorted cell that was drawing more than its fair share of charging current.
It turned out that both guesses were correct. The fan was stalled and very hot, and one battery (the western-most in the upper bank) was bulging and overheating. We had not brought tools, but with Clint's pocket “nerd-knife” it was possible to disconnect the offending battery and remove it from the chain. Also, the fan was unplugged. More attention will be given to the fan on a future trip.
At approximately 9 P.M. the power meter reading was 6314.
We got word from Randy, K7SL, that vegetation near our receiver site was getting so thick it might make getting to the cabinet difficult. Randy was working most days on installing a new FM broadcast transmitter on the site. Firday the 29th seemed like a good day for us to follow him and another KSL technician up to the site.
Our rack and tower were, indeed, getting lost in some kind of shrub. We are still not sure what it is that can grow so verdantly at 9000 feet where there's no water and very little soil on the bedrock. Clint, KA7OEI, had brought his electric chainsaw and generator along, but even with power help, loppers, and hand shears, it took several hours to clear a path and cut down branches that were blocking the path in front of the lower element of our receiving collinear.
We then made some receiver sensitivity measurements as follows:
|Measurement||μV for 12 dB SINAD||μV for 20 dB SINAD|
|Without CCITT Filtering:|
|With crystal filter & preamp||0.28||0.54|
|W/ filter, preamp, cavity||0.54||1.14|
|Desensitization by External Noise|
|Conditions||Signal for 20 dB SINAD (dBm)|
|Thru Iso-T w/dummy load||-78.0|
|Thru Iso-T w/antenna||-59.5|
|Site Noise De-sense||18.5 dB|
|Backup J-pole thru Iso-T||-64.5|
|SWR @ 146.02 MHz:||2.3:1|
|SWR @ 144.62 MHz:||2.5:1|
|SWR of backup J-Pole @ 146.02 MHz:||1.4:1|
The outer insulation on the receiver feedline had been breeched. We made a temporary fix by putting electrical tape around the break. (The shield was intact.) We need to replace it with a 12-foot, 4-inch feedline having a male UHF connector on one end and a male type-N connector on the other end. We also need a 6-foot length of ABS with a 90° sweep on top.
|Current squelch setting:||24|
|Setting for full squelch closing:||21|
|Power output at feedline input:||82 Watts|
|SWR at feedline input||1.6:1|
A J-pole link antenna that had been made several years ago was installed. A better location for the transmit antenna was available, but there was not time to make the move. It will have to wait for a trip on a later date.
Starting in late January, remote telemetry revealed a consistent drop in maximum battery and bus voltage compared to typical values over the last several years. It had been customary for the power supply voltage to rise to about 14.5-14.7 volts, disconnecting from the battery bank when it passed 13.8. Now it was not even getting up to 13.80.
The most likely cause was a battery in the bank (of nine) with a shorted cell that was drawing enough current to keep the supply voltage from rising higher. A less likely cause might have been a power supply fault such as one of the two supplies having blown a fuse and dropped off line.
If it, indeed, was a bad battery, it could get worse and possibly draw so much current to take the system down. Keith, KE1THR, who spends his winters as a ski instructor at Brighton, agreed to ski up to the site and investigate. On Tuesday March 1, he skied up to the site to check the batteries and power supplies.
Clint, KA7OEI, reports the measurements made:
The batteries are “numbered” starting on the top shelf, left-to-right, bottom shelf left-to-right with the “open” side of the bank facing.
The first voltage is that under load, the second is immediately after the load is removed:
Power supplies, battery bank connected: Top, unkeyed: 1.3 amps, 14.45 volts - current slowly decreasing Top, keyed: 1.9 amps, 13.16 volts - current slowly increasing. Bottom, unkeyed: 3.0 amps, 14.45 volts - current slowly decreasing Bottom, keyed: 4.0 amps, 12.96 volts - current slowly increasing Power supplies, battery bank disconnected: Top, unkeyed: 0.2 amps Top, keyed: 1.4 amps Bottom, unkeyed: 0.5 amps Bottom, keyed: 12.0 amps Upper power supply only, battery bank disconnected: Bottom supply, unloaded voltage: 14.72 volts Top supply, keyed: 13.4 amps, 12.86 volts Bottom power supply only, battery bank disconnected: Top supply, unloaded voltage: 14.62 volts Bottom supply, keyed: 14.3 amps, 13.22 volts Battery bank voltage, after load testing, 15-20 minutes after having been disconnected: 12.4 volts, using load tester. Note: Load tester calibration should be checked against a known-accurate voltmeter. * * *
Analysis and comments:
- After Keith disconnected/reconnected the power supplies, the load was shared more equally between the two indicating a slight amount of resistance, possibly in the Anderson power-pole connectors on the DC bus.
- It was verified that the low-current float charger appeared to be operational: Red LED on, slightly warm.
- No physical indication of any battery problems (e.g. swelling, detectable temperature difference, etc.)
- The worst battery (#9) was marked has having been originally put into service at USSC in 3/08 and removed from service in 9/12.
- Some of the newer batteries (such as #9) have surface-mount connections rather than "flag" terminals. The small, exposed stainless-steel bolt makes it more difficult to achieve a good, low-resistance connection for the load tester than the large, lead flag terminals, possibly resulting in somewhat depressed "load" voltages. This does not explain the somewhat lower "unloaded" voltage of battery #9, however.
- There is no obvious reason why there has been a change in the voltage as noted on the bus. It was expected that a defective battery (e.g. shorted/weak cell, detectably warm, visibly bulging) would be apparent, but this was not so. Other than having unplugged/reconnected the Power Pole connectors on the DC supplies and a subsequent change in current loading, no changes were made.
- It should be noted that under normal key-up conditions that some "surface charge" is pulled from the battery bank. This fact is obviated by noting that the repeater around 14 amps key-down, but only a total of 6-8 amps was observed being pulled from both power supplies during observation and it was noted that the battery-connected, key-down current was slowly increasing and the un-keyed current was decreasing over time, indicating an exchange of energy. This charge/discharge represents a rather small percentage of the overall battery bank capacity and should not cause appreciable "wear" on the cells over "static" float charge conditions.
Again, thanks to Keith and his friends for going up there to dig out the building a few week ago, Keith for nerding out on skis today, and Gordon for getting the tools to him and helping orchestrate this.
We hope to clean up the typography a little later, but thought we'd get the facts recorded in any old form while they're still fresh. -K7HFV
On Saturday morning, October 18, we enlisted the help of Christensen and Griffith, general contracters, to supply certified tower climbers to make major changes to the antennas on UARC's 146.76 repeater on Lake Mountain.
Here is a narration of the operation by Clint, KA7OEI:
Gordon arrived at my house at 7:45 and we spent a while re-mounting the 22-ish foot long, heavy fiberglass antenna atop the roof of my Jeep, mostly to redistribute the load and prevent it from tearing the roof rack off the top. Leaving my house a bit later than expected, we met the C&G crew at the Smith's at the Crossroads in Saratoga Springs only five minutes later than planned, at which point they offered to transfer the antenna onto the ladder rack on their truck, which we gratefully accepted.
We then proceeded to the south end of Lake Mountain and started up the Mercer Canyon road, dodging many four-wheelers and the like on this, the opening day of deer season, making it to the site at about 10:30. Upon arrival we saw that the C&G truck had managed to lose a rear tire. After unsuccessfully trying to plug it they managed to (eventually) lower the spare and mount it while their other crew member had been rigging the tower to prepare the top antenna for replacement. A while later, during lunch, one of their crew walked over to my car and pointed out that I, too, had a flat — which I promptly swapped out as well.
The first antenna to be replaced, at the 120' level, was the copper J-pole which was installed on a “temporary” basis in 2009. This was quickly lowered to the ground, the clamps swapped to the “new” antenna, a refurbished dual, folded dipole that had been obtained from Dave Williams, WA7GIE, and then elevated into place, secured, connected, and sealed. This exact same model and type of antenna is in use as the transmit antenna on both Farnworth and Scott's '62 and similar types are used as receive antennas in those locations as well. Immediately upon reconnection of the coaxial cable, use of the repeater began again and a quick check of the VSWR indicated that the antenna appeared to be working as it should.
With that part of the job done, the workers re-rigged to remove all portions of the transmit antenna and its “outrigger” mount that included the link antenna. The transmit antenna, a 22-ish foot fiberglass antenna installed in 1997, when last visited 2-3 years ago during the 76'ers party, was showing signs of severe UV degradation with all of the “gel coat” on the south side having sloughed off and it was then that I made the recommendation that we either refurbish or replace it. The antenna was now removed from its mount and lowered and upon inspection it was found to have considerably degraded in the past 2-3 years, with only a few tiny spots of the gel coat remaining and the antenna rapidly shedding fiberglass with the slightest touch (something that is NOT recommended!) In its present condition it is likely that it would have failed either mechanically or electrically within 2-3 (and certainly within 5) years, anyway.
Also removed was the “outrigger,” an approximately 2-foot offset mount that had been installed in 1997 on which the two UHF link antennas were mounted. Upon ascertaining the situation with the on-tower workers making in-situ observations of the blockage caused by the tower structure itself, it was determined that, if possible, we should mount the “fill” antenna as far outboard as possible — farther than this original “outrigger” permitted. Fortunately, I had brought with me a large, galvanized, steel bracket that had originally been used on the UARC COW for mounting its cell antennas. Using a combination of hardware that we had on hand and some that the C&G folks had in their trailer, we rigged this to function as an extension to reach another 26" or so and then mounted the original outrigger to it.
The two UHF Yagi link antennas — one for the Orem auto patch and the other for the Salt Lake auto patch — were removed and three pieces of hardware were installed on the mast. The first was a galavanized “ice breaker” — a bar constructed of angle iron by myself and Dan Kasteler, N7HIW, that would (hopefully) intercept ice falling from above and protect the antennas below it. Immediately below this is a rugged, commercial UHF Yagi — that had originally been on Scott's Hill as the link antenna, but replaced since it was designed for the upper portion of the UHF amateur band (it was supplied by Bruce Bergen, KI7OM) that is now pointed north. At the bottom of the mast on this outrigger is a 3-element welded aluminum VHF Yagi, also pointed north with solid 3/8" elements: This antenna was originally built by Mike Collett, K7DOU (I think), but then completely reworked by myself and N7HIW with a gusset support and much heavier boom material such that only the original elements, gamma match, and N connector were retained.
The hardware was hauled up, with everything mounted, and then wrestled into place and bolted into place. Once that was done I donned the appropriate safety gear and joined the crew at the 60' level with some test gear (an HT, various lengths of coax, a step attenuator and some adapters) and the antenna coupler box that I'd built.
Previously there had been, mounted at the transmit antenna, a Maxrad VHF/UHF diplexer — this, to separate the VHF TX signal from the UHF RX signals from the link Yagis. Enclosed in a plated steel box, this box had almost completely rusted through at this point and would have likely failed in just a few more years without some attention. This new box is a die-cast aluminum box that is sealed along the top with silicone and some drain/weep holes in the bottom, with the components inside liberally coated with clear laquer as well as protected with an internal “drip shield” to protect the components within from condensation — plus, the box itself has sloping sides so that water will tend to run off.
This box does the original job of separating the VHF TX signals from the UHF RX signals, but it also taps off some of the VHF TX signal to go to the “Fill” antenna (the 3-element Yagi) — in this case, at about -8dB from the main signal. Doing so with just a capacitor tapped off the main TX line, it is extremely simple, electrically, and some inductance was added to the main antenna line to cancel out the slight “bump” that this capacitance added. In testing, when the output power of the “fill” antenna port plus the output power of the main antenna were summed, the losses on the VHF side were measured to be on the order of 0.2dB while the losses on the UHF side were around 0.7dB.
Once on the tower I called Phil, K7PB, who lives on the east side of the Salt Lake valley, an area that was known to be only “somewhat” affected by the shadow on the tower. Having him switch his all-mode radio to SSB mode we had him compare the S-meter readings of the (now-new) “main” antenna, and then the “Phil” antenna (Gordon and I jokingly called it), each running separately and he reported an “S8” and “slightly above S8,” respectively. I connected both antennas at once and he again reported a “slightly above S8” reading. Inserting a cable cut to an electrical 1.5 wavelengths at the transmit frequency we ran the same test with both antennas again and he reported a signal level of about “S6.” We concluded that the we'd found something pretty close to “out of phase” with the 1.5 wavelength test cable and that the “original” cable length just happened to be “close enough,” so we left it at that!
Descending the tower I left the crew to tidy things up and seal all connectors while I started gathering things up in preparation to leave. In checking the transmit system VSWR from the end of the coax in the building it was found to be between 1.4:1 and 1.5:1, but this is not too surprising since the transmit antenna, which according to the test data from Sinclair should have a return loss of at least 20 dB at the transmit frequency, is very close and parallel to other similar-length antennas as well as the tower structure — something that would certainly adversely affect the performance of the antenna, but cannot be helped in our situation!
The C&G crew left before we did as it took a few more minutes for us to clean up our tools around the site and pick up a few bits of trash from our having stripped weatherproofing off cables, etc. but we departed the site a little after 4:30, eventually catching up with them near the bottom of Mercer Canyon. Just short of the blacktop we transferred the old antenna to my Jeep, now that we'd be off the bumpy, dirt road, and bade our goodbyes and gave our thanks to their work. From there, I headed to my work to drop off the old antenna while listening to the QSOs on '76.
It was not expected that there would be any change in signal levels in Utah County — and we could detect none. In descending the north side of the point of the mountain it struck me that it was significantly better than it had been in the past, slightly “multipathy” in a few places, but never any difficulty in copy. Exiting the freeway at 7200 south and working my way north along 700 West, I found that direct line-of-sight was occasionally blocked by local obstructions, but the signals were generally good. After dropping the antenna off at U.S. Satellite we headed over to the Sizzler on 7800 South and east of Redwood via 1300 west and passed through an area — the intersection of 1300 west and 7000 south, happening to be within half a block of the QTH of Venus, KB7FXB — that I knew to be particularly problematic for the '76 repeater: There, the repeater had been usually uncopiable, both by being in one of the worst of the null areas and also by being easily overridden by local powerline noise. There, the signals were generally good, now.
Phil, K7PB, has also been driving about in the valley and has reported that a few “known-bad” areas are much-improved, particularly in areas in the middle-west portion of the valley where he'd known the repeater signal to be overridden by powerline noise. He also reports that on the east side and downtown, there were some improvements in the signal coverage of '76, not something that we were intending/expecting, but are gratified to know about nonetheless.
Based on our observations we expect improved signals in Davis and Weber counties as well, keeping in mind the limitations of geographic coverage. We did hear someone near Ogden make a comment that he thought the signals were better, but we need to hear more observations before coming to conclusions. We are also interested in reports of users much farther north, such as Brigham City to see if they note any change.
To summarize the changes:
|Two-meter receive||Top of tower||Dual folded dipole||Replaced “temporary” J-Pole|
|Two-meter transmit||60-foot level||Sinclair fiberglass 8 dBi||Replaced similar but weather-worn antenna|
|Fill (additional transmit)||4 feet out from 60-foot level||3-element Yagi||To cover western Salt Lake Valley|
|Link receive||4 feet out from 60-foot level||5-element Yagi||Needed for control and IRLP|
Since a spring ski trip when Keith, KE1THR, brought down a malfunctioning power supply from the Scotts' Hill site, the site had been running on one of the original two power supplies plus a temporary switching supply put straight across the battery bank. On Monday, October 6, Clint, KA7OEI, and I (K7HFV) went to the site with the plan of reinstalling the now rebuilt supply and rebuilding the other one on site. Clint was able to reinstall the one, but we discovered we did not have the new resonating capacitor for the other.
We took the supply back to Clint's house in West Jordan where he was able to complete the update. The update included replacing the capacitor and replacing a DC circuit breaker with a large fuse. We returned to the site, installed the second supply, and removed the temporary switcher. Parameters looked resonable including a receiver sensitivity degredation due to site noise of about 2 dB.
Starting on March 4, we started getting low voltage readings for the battery bank and the power supply bus on the Scotts' Hill site, which houses one of the synchronous pair of repeaters on 146.62. (See telemetry readings.)
The best explanation seemed to be that one cell of one of the batteries might have shorted. If the short continued to get worse, it could drag down the power supply voltage and ultimately take the whole repeater off the air. Some emergency intervention was indicated.
The Scotts' Hill site isn't entirely trivial to reach in the winter, sitting at about 10,000 feet elevation and often buried in several feet of snow. (See photos). We had one lucky card up our sleeves: Keith Stebbings, KE1THR, works as a ski instructor at Brighton (within sight of the Scotts' site) and is a very experienced cross-country skier. Keith volunteered to make a ski-trek to the site and try to disconnect the uncooperative battery. He had one free Saturday available, March 8.
Clint, KA7OEI, passed on to Keith the basic tools that would be needed plus a battery load tester. Keith headed out with a non-ham collegue, Chris, and they made the trip to the repeater site in about an hour. The door to the building was behind about five feet of snow, so Keith put his avalanche shovel to work. The next problem was that ice may have sealed the door to the jam, making it impossible to open. Keith's avalanche shovel came in handy again.
The repeater runs on two ferro-resonant power supplies whose outputs are normally connected in parallel. Keith read the voltage and current on each supply and found an odd situation:
Clearly, the load was not being equally divided. Keith kept in touch with Clint via radio and they decided to check the batteries next. Oddly, no battery seemed to be enough below par to indicate a shorted cell. The voltages that Keith measured with the load tester were:
|Upper bank l. to r.: 1-5 |
Lower Bank l. to r.: 6-9
The battery measurements failed to turn up a culprit, so attention turned to the power supplies. The almost non-existent current in the upper supply seemed suspicious. Clint suggested that Keith fiddle with the DC output breaker on the upper supply. At first, it seemed to make no difference whether the breaker was on or off, but then Keith discovered that if he pushed hard on the breaker handle, the picture changed. The upper supply came on line and current picture now looked much different.
A note was made to change the breakers on a future trip. It is likely that the poor performance of the lower supply was due to a faulty capacitor in the resonant circuit. That also should get some attention on a future trip.
Keith made a log entry, gathered up his gear, and, together with Chris, made his way back down. The return trip took only about 20 minutes.
Alas, when telemetry readings were taken on the morning of Monday, March 10, the bus voltage appeared low once again. It is likely that the upper supply's breaker has tripped out again. At least with no serious battery trouble it is likely that the repeater will remain stable until the site is more easily accessible.
On the Monday morning after our 10-26 Farnsworth visit I (Gordon, K7HFV) got a phone call from Clint, KA7OEI. It seems that while writing up the events of the past weekend Clint had realized there was a problem that might cause grief if allowed to stand. When building the new battery controller for the Scotts' Hill site, Clint had placed an electrolytic capacitor between the gate and source of one of the switching FETs. He installed it with the normal polarity — positive on the gate and negative on the source. But now it occurred to him that the device in question was a P-channel FET and the gate was more negative than the source. The capacitor was in backwards. The unit was working fine, but a reverse-polarized electrolytic could very well short over time and take the whole site off the air for lack of power.
To further complicate the situation, a storm system that threatened to close the road for the season had already dumped its first bit of snow. If the problem was to be fixed, it had to be done immediately.
Clint arranged to get away from work for a few hours and we both headed for Big Cottonwood Canyon. For safety we took both of our four-wheel-drive vehicles in case one needed rescuing by the other or one became stuck badly enough that the other was needed to take us back to civilization.
When we got to the high ridge the wind was howling with a vengeance, but road conditions were not much different from what they had been on Saturday. There was only a light layer of snow that had come in overnight.
Clint made very quick work of replacing the capacitor with a new one, installed this time with the correct polarity. (The original one may have been weakened by the time it spent enduring reverse voltage.)
We left quickly and hoped the repeater would be happy until the road became driveable in 2014.
Having found on Friday that audio on the link from Farnsworth Peak to Scotts' Hill was intermittent, and the season for easy access to repeater sites would soon be coming to a close, Brett, Clint, and I (Gordon) spent most of our Saturday at the Farnsworth site.
The behavior of the problem suggested that a simple loose connection might be the culprit. Clint started checking audio cabling on barrier blocks and in less than a minute found a spade lug that visibly moved when touched and made the audio problem appear and disappear. A fraction of a turn of the screwdriver fixed the problem. Other screws in the block seemed to be less tight than they should have been. They might have loosened from thermal cycling, they might have been too loose to begin with back in 2009, or they might have been victims of a combination of the two. With the end of the easy-access season likely approaching, a large number of folks who maintain installations on Farnsworth had been active doing last-minute maintenance. This burst of activity may have jostled our rack enough to cause the final disconnection.
Having fixed the problem quickly, we looked into another issue. We had heard that one of our antennas would have to be moved. Danny, one of the KSL employees, was on duty and he showed us the problem. There was a ham 70-cm antenna right in the path where a microwave dish, yet to be installed, would need to shoot. It turned out that the antenna in question was not ours, but was one we were borrowing.
At one time Glen Worthington, WA7X, had operated a repeater on 449.1 MHz from the same room where the transmitter portion of UARC's '62 repeater is located. Glen's repeater had been taken off the mountain for revamping and a possible frequency change. It had been off the air for several years. The antenna blocking the new microwave path was Glen's. On a trip during 2012 we had discovered an intermittent type N connector on our Heliax feedline to the link antenna that communicates with Scotts'. Not having a spare heliax connector in our tool kits, we simply connected to Glen's antenna, as it was not otherwise in use. Now it was time to get things back to normal.
Clint replaced the bad connector on our own feedline. We rerouted all three of our feedlines (Scotts' link, control link, and two-meter transmit) as well as a spare RG-8X run to the transmit antenna, to a new cable entry point using Microflect hardware. This required chiseling out some sort of resin that a contractor had used to seal the old (and rather haphazard) entry. Brett took on this job. Danny told us that space could be made available for Glen's antenna on the short tower where the '62 antennas are located. This would happen after the road opens for the 2014 season and a large TV receiving antenna could be replaced with a smaller UHF-only version. We understood that the planned reactivation of Glen's repeater was also not expected until the 2014 season. We brought down Glen's antenna and feedline and stored them for use when the new space became available. The antenna was suspended from a conduit behind the racks.
Brett supplied an updated copy of the club station license and attached it to the rack as required by the rules.
Finally, we took the repeater's vital signs and found the following:
|Power into UHF Link Feedline||4.5 Watts|
|SWR into UHF Link Feedline||1.3:1|
|Power into 2-meter Transmit Feedline||72 Watts|
|SWR on 2-meter Transmit Feedline||1.65:1|
|Received Signal Strength from Scotts' Link||-87 dBm||At cavity output|
|Measurements at the Two-meter Receiver Site|
|Two-meter receiver sensitivity at bandpass cavity input||0.30 µV||for 12 dB SINAD|
|Receiver sensitivity at input to crystal filter||0.20 µV|
|Receiver sensitivity at input to preamp||0.15 µV|
|Receiver sensitivity at input to receiver||0.50 µV|
|Local oscillator frequency||250 Hz low||Nominal = 135.319 MHz|
|Sensitivity thru Iso-T w/dummy load||-85 dBm||For 12 dB SINAD|
Iso-T loss = 38.5 dB
|Sensitivity thru Iso-T w/antenna connected||-75 dBm*||Site noise degredation = 10 dB|
|Inferred effective receiver sensitivity||-113.5 dBm
|Receiver feedline SWR||2.2:1|
|*There is some question about this value. It was recorded as -95, but that's backwards from what it should be relative to the dummy load reading. Clint thinks he remembers it as -75 which is in a believable range.|
It seemed to be some kind of record that the three of us had visited all three UARC repeater sites in a period of just eight days. Just a short time before it had seemed we might get through 2013 without a need to visit any of them.
Clint, KA7OEI, had constructed a new charge controller and monitor for the backup battery bank on the Scotts' Hill 146.62. It appeared that a run of unseasonably warm October weather was about to end, and if the new box was to be installed before site access got more difficult (and a lot less comfortable), it would have to be done soon. So on Friday, October 25, Clint, KA7OEI; Brett, N7KG; and the author, K7HFV, made a trip to the site in two vehicles, just to enhance the chance of self-rescue if needed.
The four-wheel-drive road was in reasonably good condition, although there were some very slippery spots and one rut that was deep enough to swallow smaller vehicles. We made it to the site without serious trouble and a minimum of expletives.
The installation went smoothly except that a small “wall-wart” power supply had been left behind. Its purpose was to allow direct monitoring of a-c line voltage. Most of the functionality of the box, though, would be available without it.
Clint tells us that features of both old and new units include:
The new unit added the following capabilities:
With the box installed and the repeater powered back up, we tested for basic repeater operation. Although we could key up the repeater with an H-T, Clint noticed that we were not getting the “Scott's Beep” by which the voter unit indicates that the last transmission came from the Scotts' receiver rather than the Farnsworth receiver. This should have been coming in on the eastbound link from Farnsworth. A check on the link frequency with an H-T showed it was missing there as well. I wondered if any audio from Farnsworth was arriving. Clint checked and found that the link audio arriving was intermittent and sometimes present but low in level and missing its highs. Apparently it had been necessary to go to Scotts' to detect a problem on Farnsworth. We decided to make a Farnsworth trip the following day if possible.
While up there we also load-tested some of the older batteries (ones that were there since the site was brought online in 2009) and found that one of them went open-circuit when it was loaded at about 100 amps. This was replaced with a newer battery that I had brought with me. The other three batteries in the “original” portion of the battery bank should be replaced next year as it was evident that their internal resistances were increasing.
Having done what we could, we packed up and went back down the canyon.
FCC rules require that an automatic identifier using the International Morse Code send no faster than 20 words per minute. An electrolytic capacitor nearing its 40th birthday had lost enough of its original value that the '76 IDer was running at 23-24 words per minute. This did not seem like a crucial problem, but because Brett, N7KG; Clint, KA7OEI; and I (K7HFV) were going to Lake Mountain anyway to help boy scouts on a JOTA event, we thought we might as well drop by the repeater site and replace the capacitor.
In preparing for the trip and determining what component(s) we should take up we discovered that our documentation package did not have a schematic of the digital portion of the ID generator. The ID/Timeout/COR unit was constructed in the 1970s under the supervision of the late W7EU. It used a commercially available PC board based on an article in 73 magazine. The schematic we had of a portion of the board was a badly faded electrostatic copy of the 73 article. The IDer was mentioned as coming from an earlier article, but which issue of 73 was not mentioned. Fortunately, Clint had images of 73 articles from the 70's as they had once appeared on the web. We were able to determine that the IDer had appeared in the February, 1973, issue, and the remainder of the board in the March issue of the same year. Clint printed off fresh copies of both articles. The baud rate generator was an oscillator made from two cross-coupled sections of an SN7400 quad NAND gate. The capacitor in question was a 100-µF.
When we arrived on site we discovered that the capacitor actually in the board was a 50-µF unit. Apparently it had been hand-picked for optimum code speed. After being removed from the board it measured at 62 µF. Installing the new 100-µF component brought the speed down to about 17-18 WPM. We decided that would be adequate and allow for some further aging of the capacitor. We can't guarantee another 40 years, but who knows?
We had also planned to mount a terminal strip on the link receiver more permanently than the tie wrap we had used when the receiver was installed in 2011. It turned out that the strip was sitting directly on the Motorola chassis and drilling screw holes was inadvisable. Brett had a variety of glue that seemed appropriate for the job.
Clint had brought one of his service monitors along and wanted to check basic parameters of the repeater, so the following were measured:
|Transmitted Forward Power from transmitter||21.2 Watts|
|Transmitted Power from cavities||11.0 Watts||Cavity and isolator loss = 2.85 dB|
|Link signal required for 20 dB quieting||-93.5 dBm||(Quieting guessed by ear)|
|Link received signal level||-80.0 dBm||From Murray control point|
|UHF Link SWR into feedline||1.4:1|
|UHF Link SWR into cavity||1.5:1|
|VHF Transmit Feedline SWR||<1.2:1|
|VHF Receive Feedline SWR||1.2:1||at 146.16 MHz|
|VHF Receiver Threshold at cavity input||-119.0 dBm||for 12 dB SINAD|
|VHF Receiver Threshold at receiver input||-123.0 dBm||Cavity loss: 4 dB|
|VHF Receiver Threshold thru Iso-T||-81.5 dBm||Dummy load on antenna port|
|VHF Receiver Threshold thru Iso-T w/ antenna||-66.5 dBm||Site noise degradation: 15 dB|
Miscellaneous data to remember:
This is a little embarrasing. It's October of 2013 and we have never written up the conclusion of the incident below. Details in memory are a little sketchy, but here's the basic outline.
Observation: Clint, KA7OEI was able to correlate the interference with two FM broadcast stations, one of which was on Lake Mountain, and the other many miles away. Apparently the distant station was mixing with the near one, possibly in the near one's power amplifier, possibly in something external such as a guy wire, producing an intermodulation product on our link frequency.
First Trip: Brett, N7KG, and I (K7HFV) attempted to make a trip to the mountain to see if the problem could be solved by just tightening the link receiver's squelch. There are two roads to the Lake Mountain repeater site. The north or Israel Canyon road is the shorter one, but has the steepest grades and is a northern exposure, causing it to retain snow and ice quite late in the season. The southern or Mercer Canyon route has a much gentler slope but is considerably longer, especially for those coming from the north end of Utah Lake.
We first went to the southern road thinking it might be driveable with a four-wheel-drive vehicle and chains. We found the road blocked by a new locked gate with a sign indicating it would be closed from “November 1 to mid-May.” It appeared we would need to hike. The northern route, being shorter, is the preferred hiking route, so we drove back around to where it begins. Unfortunately, the problem had gone away on its own by the time we arrived, so there was no point in continuing the trip.
Second trip: A few weeks later we tried again. A party consisting of Brett; Clint, KA7OEI; Randy, KG7GI; and the author, hiked up to the site. We carried snowshoes, but never needed them because the road was so well packed from travel by snow cats and snowmobiles. We tightened the link receiver's squelch by a small amount, determined that there was still adequate margin to hear the link transmitter in Murray, and went on our way. There were no further reports of the problem.
Lake Mountain Auxiliary Link Gets Interference
The “76'ers” group that uses the 146.76 repeater heavily reported that there had been a problem with the usual Thursday morning “Say Good Morning with Radio” net, a nationwide net held on an IRLP reflector. Doug, KE7BBQ, called me (K7HFV), and said that while the IRLP connection had been made all right and the remote stations were coming through fine, stations on the reflector could not hear the Utah stations and no one had been successful dropping the connection.
I listened and found that the squelch for the link receiver (which receives the IRLP audio from the node in Murray) was not closing when the link carrier went off between IRLP transmissions. My first thought was that the squelch needed to be tightened. Squelch thresholds sometimes drift with temperature and component aging. One other thing that can cause a squelch to open is weak carrier from the outside.
Repeater trustee Brett Sutherland and I made arrangements to rent snowshoes and prepare for a trip to the mountain on the following Saturday.
11-??-12 Miscellaneous Maintenance on Farnsworth
Clint, KA7OEI, and Gordon, K7HFV, made a trip to the Farnsworth Peak site of the 146.62 repeater. Below are Clint's notes.
|Transmitted Forward Power from PA||107 Watts|
|Transmitted Power from duplexer||67 Watts||Duplexer loss = 2.0 dB|
|VHF Transmit Feedline SWR (at duplexer output)||1.65:1||Antenna originally for 155-165 MHz|
|UHF Link Transmitter Forward Power||5.0 Watts|
|UHF J-Pole VSWR||1.3:1||Intermittent|
A bad “N” connector on the LDF4-50A feeding the UHF link J-Pole was noted causing intermittent connection and will need to be replaced. When the connection was properly made, the VSWR at the indoor end of the coax was 1.3:1 (on UHF J-Pole). Because of this intermittency, the UHF link to/from Scott's was moved to the WA7X antenna.
A 30K (at nominal room temp) thermistor and a 56K resistor were connected in series and this combination was put across pins 5 and 6 of the NE567 in the 3.2 kHz tone detector on the Scotts-to-Farnsworth UHF link. Before, it went up in frequency with decreasing temperature, but now it goes down — but not by as many PPM/deg C. Based on the results of the winter of 2012-2013, it seems to have prevented this decoder from drifting off frequency and becoming unreliable. A similar thing should be done at Scott's.
Attempted to modify voter code to add some usage timers so that the voting beep reflected how much time was actually spent on each receiver rather than simply the last COS — this, to make it more accurately depict how much time was on a particular receiver. For some reason, it didn't fix the problem, but the code was left on the voter since it didn't seem to break it, either.
Farnsworth Squelch Battery Replaced
Randy, K7SL, needed to make a trip to Farnsworth Peak for his employer and remembered that we usually put a new 9-volt battery in the '62 squelch logic at this time of year. He purchased a new battery (being careful to avoid Eveready) on his way up and then made the switch.
The repeater's squelch (or “carrier operated switch”) can be set remotely to one of 32 settings in order to match changes due to temperature or site noise. An SSI 4000-series CMOS register holds the current value. A 9-volt battery retains the current setting during any short power failures that might occur.
The setting that was made last August seemed to be a good one — weak signals could get through, but there was no problem from keyups due to site noise. Unfortunately, we didn't know just what that setting was. Randy read it back from the panel as “23.”
Randy also reported that there was generally about three inches of snow on the ground, but he had to “power through” a few deeper drifts that were up to his grille. The season of easy access to this mountaintop is likely to be ending soon.
Min/Max Bug Fixed
Clint, KA7OEI, made a trip to Scotts' Hill on the morning of the 23rd to fix a bug in the software that records minimum and maximum temperatures. Now, minimum and maximum are kept in non-volatile flash memory, so that they will survive not only a “warm reset” but also a complete power failure.
Miscellaneous Jobs Completed on Scotts'
With the season of accessibility by road close to ending, Clint, KA7OEI, and the author, K7HFV, went to Scott's Hill to finish up several projects related to the 146.62 repeater there. Fortunately, the snow that had caused so much consternation on the 13th was almost entirely melted.
First was an attempt to further isolate the source of a noise floor increase when the 145.27 repeater was keyed up. This increase was measured on the October 13 visit. Clint set a signal generator on the repeater input at a level low enough that the noise floor increase would cause the repeater squelch to close. This made the repeater effectively a detector of the noise floor. Then he connected a 50-watt transmitter on 145.27 to a Yagi and tried “spraying” the '27 signal at different parts of the site. The one area that seemed to produce noise floor was the Forest Service building, one of the three buildings on the site. Further investigation will have to wait until contact can be made with the folks who service the Forest Service communications equipment.
The next project was to install the additional batteries that were missed on the 13th. That was accomplished without difficulty. We should now have at least 250 amp-hours of capacity.
The final project was to install a temperature sensor that could be read remotely. This involved connecting an LM-335 to an unused A-to-D input in the squelch unit and adding software to read it. It took a few hours of debugging to get everything working, including a feature that Clint thought of adding after we were on the site. That feature was to record a maximum and minimum temperature that could be read remotely. Ultimately, the idea was extended to include maximum and minimum supply voltage.
With that all apparently working we headed back to the valley. We queried the temperature and discovered that the maximum and minimum were being reset periodically. Oops. Maybe there would be time for Clint to return on the following day (Sunday). Otherwise, at least the readout of current temperature was working and that was all we had originally intended.
Progress Made Reducing Scott's Interference
After we had narrowed down the source of our 146.02 interference (see the 9-29 entry) to a particular building, John Lloyd, K7JL, arranged for us to meet with representatives of the site owner, and Utah Broadband, the owner of the equipment suspected of being the interference source. Joe Nelson and Brandon from Utah Broadband met us that morning and were very helpful and cooperative.
Getting to the site was a bit more difficult than expected. Recent snows had left the road snow-covered in some areas and icy in others. Branches had fallen across the road where it was cut into a steep slope, resulting in limited overhead clearance. All the vehicles made it to the base of the hill where the radio sites are located, but only Clint, KA7OEI, got his vehicle to the top; the rest of the party walked up the final climb, about a five-minute endeavor. (Our intent to install some heavy batteries gave Clint some extra incentive.) Clint's route involved going through the sagebrush to avoid a badly drifted section of the road.
The interference was not nearly as strong as it had been on September 29 (likely because it was warmer then) and we were not able to recreate the beautiful spectrum analyzer pattern we had seen before. The spectrum analyzer sat out in the open for much of the day as a reminder that we tried. However, the signal was easily audible in the 2-MHz range on a portable receiver with a BFO. It had a very distinctive sound (very far from a T-9 note).
Turning off one of Utah Broadband's several “power over ethernet” (POE) power supplies caused the signal to disappear. It was first thought this power supply was the culprit, but replacing it did not solve the problem. The offending supply was finally determined to be one on the tower powering an IP link to Farnsworth Peak.
It was possible to wrap a number of turns of the ethernet cable that powered the power supply in question around a television flyback core. This addition made a noticeable reduction in the interference. Additional ferrite cores were added in other places for some additional insurance.
It turned out that the problem link was a backup for other paths and was not necessary in normal operation, so the Utah Broadband folks agreed to keep it turned off except when needed. (It could be remotely controlled.) A more permanent solution can be investigated next summer when the site is again accessible. The Utah Broadband folks departed and Clint and I turned to some other projects.
Clint installed the new “battery trolley,” a roll-around cart built by Bruce Bergen, KI7OM, and capable of holding a large number of batteries. Clint placed several new batteries on it and wired them into the backup power array. (Three more batteries were waiting in my car back at the end of the pavement, but a second trip along the dirt road to retrieve them proved infeasible.) This should improve the run-time during a power failure.
John Lloyd, K7JL, came up to join us. He decided not to try the icy road, and instead walked in from the pavement (which takes a little over an hour). John helped us investigate another problem: when the 145.27 repeater in a nearby building was keyed up, the general noise floor on two meters came up and our effective sensitivity on 146.02 went down. No final diagnosis was made, but the problem had some puzzling characteristics. It was present when '27 was transmitting, but not when '62 was transmitting. The '62 system has a '27 notch cavity just to avoid desensitization by '27.
|Antenna||No||-54 dBm||Degredation of 5 dB from site background noise|
|Antenna||Yes||-46 dBm||An additional 9 dB of degredation when 145.27 is keyed|
The noise floor while '27 was transmitting, measured with a five-element Yagi about midway between the two buildings, came in at -104 dBm, unacceptably high.
We noticed one additional phenomenon: There was a pattern of carriers characteristic of a 61 kHz switching power supply coming from the Forest Service building. These signals dropped off quickly with distance from the Forest Service building and were not an immediate source of concern, but likely contributed to the general site background noise.
At that point we decided to leave. That turned out not to be as easy as one might have hoped. The route that Clint had taken up the hill when we arrived turned out not to be reversible. Clint tried powering through the snow drifts but, even though he was going downhill, he became high centered. It was late afternoon and the other vehicles that had come in with us had already left. John figured that there was little he could do to help, so started the walk back to the pavement after suggesting some alternate routes that Clint might have taken.
The only shovel along was of the portable, folding variety which can easily lift tablespoons full of snow at a time. Clint tried digging snow out from under the vehicle, but after a long period of shoveling, we were still stuck. We called for help. Randy, K7SL, was willing to come to our aid from where he was working at the KSL radio transmiter.
Randy was over an hour away, so Clint continued shoveling and finally managed to get enough snow out that his vehicle could be backed out of the snow drift. This gave us freedom to travel back and forth over the last quarter-mile of road between the snow drift and our building, but no way to get to the outside world.
Randy arrived and after looking over the options decided that he could winch Clint's vehicle up on to the original route that Clint had used to bypass the snowdrift. Randy guided Clint onto a safe approach angle, ran the winch cable, and spent a full ten seconds winching before Clint could continue above the road under his own power. We headed down the mountain and I had a great excuse for being late to the Homebrew Night UARC meeting.
Interference Source on Scott's Located
The 146.62-MHz repeater on Scott's Hill had been receiving intermittent interference during the early fall season. The interference was manifested mostly as squelch breaks from a weak carrier on 146.02. The problem with any kind of intermittent interference is that, to isolate the source, one must be present when the interference is occurring, and that may not be predictable.
On Thursday evening, September 29, the problem began happening and Clint, KA7OEI, and the author, K7HFV, were available to go to the site and investigate. We had been on the site while the problem was happening several weeks before during a Wednesday night hike, but had not been well-equipped to pinpoint the source. From what we could tell from handie-talkies, the strength of the extraneous carrier did not vary much as we checked it in different parts of the site. That behavior implies the source is some distance away. Other radio sites such as Lewis Peak were discussed as possible sources.
On the 29th, however, Clint went equipped with a spectrum analyzer and service monitor as well as a Yagi. A look with the spectrum analyzer showed a pattern of radiation peaks every 280 kHz, alternating high and low in strength. This is the pattern of harmonics produced by a near square-wave with fundamental frequency of 280 kHz. The most likely culprit for such a signal is a switching power supply.
The Yagi showed the signal to be coming from a neighboring building on the same site. The site had multiple towers, but as best we could ascertain, the signal peak came when the Yagi was pointed at a tower on which indicator lights could be seen, apparently on a piece of equipment mounted right on the tower.
We could do little else that night, so the next step was to arrange a meeting with people from the company that owns the site and people operating the equipment most likely to be the source of the problem.
New Link Receiver Installed on '76
On Saturday, August 20, Clint, KA7OEI, and the author, K7HFV, replaced the UHF link receiver, a GE MasterPro, with a Motorola Motrack. The MasterPro worked just fine, but had a very wide IF bandwidth intended for 15 kHz deviation. In an area with a bandplan based on 5 kHz, this made the adjacent channels unusable for any potential user in the same geographic region. The change should make the frequency coordinator happy.
Clint had put the new receiver on frequency and made sure the audio and COS outputs were compatible with the rest of the '76 repeater. We made adjustments in the repeater's controller unit so that audio from the new receiver arrived at both the two-meter transmitter and the DTMF decoder at proper levels. In the process we discovered that the new receiver had a higher level audio output than its predecessor and that this higher level was just adequate to deliver enough audio to the DTMF decoder with the adjustment all the way up. Apparently we had been running a marginal level to the decoder for some time. (That would be since 1977, based on labeling on the MasterPro.) The new situation may help the problem of occasional missed keep-alive tones used to make the control system fail-safe.
We measured audio frequency response from the new receiver to the two-meter transmitter and also from the two-meter receiver to the two-meter transmitter and obtained the following curves:
We also made sensitivity measurements for both receivers. The numbers are the signal required for 12 dB SINAD (ratio of signal to signal plus noise and distortion).
|Conditions||UHF Sig||2-m Sig||Comments|
|Measured from receiver side of cavities:|
|X||X||-41 dBm||Strange! See below.|
|Signal generator directly into receiver:|
|-99 dBm||-100 dBm||Iso-T loss: 58 dB|
|X||-113 dBm||2-m preamp adds 13 dB|
|Measured from antenna side of cavities:|
|X||X||X||-44 dBm||-53 dBm|
|X||X||X||-46 dBm||-48 dBm|
|Computed effective sensitivity:|
|X||X||X||-104 dBm||-106 dBm|
It's a bit of a puzzle why the measurement for the UHF receiver on the receiver side of the cavities and diplexer (or is it duplexer?) shows better sensitivity on the antenna (-51 dBm) than on the dummy load (-41 dBm). Normally, attaching the antenna brings in outside noise and reduces the effective sensitivity. The answer probably lies in the cavities presenting a peculiar, partly reactive, impedance.
Transmitter output power into the cavities was measured at 19 watts.
|UHF antenna thru cavities and diplexer:||1.7|
|Two-meter transmit antenna thru cavities, diplexer, and isolator:||1.15|
|Two-meter receive antenna thru cavities:||1.2|
|Two-meter receive antenna direct:||1.3|
|UHF Squelch Threshold (fully tight):||-85 dBm|
|UHF Squench Threshold (as adjusted):||-107 dBm|
|Received signal strength from IRLP site:||-90 dBm|
To remember next time: Clint attached a barrier block to the back of the new receiver using a wire tie. On a future trip we should remember to bring a drill and appropriate bits, screws, and nuts to mount the strip more permanently. It would also be good to get schematics and service documentation to place at the site or in the official “documentation sack,” or both. A copy of the current license document would also be a good addition.
Our trip coincided with a picnic by the "76ers" group. Doug, WE7BBQ, took the following videos available on YouTube:
A collection of still photos is available here on the UARC site.
We were also visited by Scott, KB7YOT, and Keith, KB7M, who were on the mountain to work on the 147.08 and 147.28 repeaters, respectively, not to mention Bruce Wilson, AD7GO, who just happened to be up exploring on his ATV. It was a busy day for hams on Lake Mountain.
|01-24-11 Power Outage on Scott's|
Friday, January 21, a snow slide caused a break in the power line to the Scott's Hill site. We understood through John Lloyd, K7JL, that the power company found a section of the access road that traverses a very steep slope to have too deep snow to be safe for a snow cat. The backup batteries apparently ran down sometime on Saturday. Power appeared to be restored by mid-day on Monday the 24th. Clint, KA7OEI, reports that there were multiple breaks in the line and that Rocky Mountain Power managed to effect the repair using “a helicopter and a cadre of snowmobiles.”
|10-16-10 Miscellaneous Changes on Farnsworth|
Clint, KA7OEI, and the author, K7HFV, went to Farnsworth Peak on Saturday, October 16, 2010. It was getting close to the time of year when the road might close for the season at any moment and several issues really needed to be taken care of. Work done included the following:
New antenna and feedline: Clint's measurements from the valley of transmitted signal strength seemed to have dropped by 6 dB over the winter. Clint had located a Sinclair model SRL 222*5 (yes the model number has an asterisk in it) dual-dipole antenna which should have had about 3 dB gain over the J-pole that had been serving us for the last 10 years. The RG-8X feedline had been out in the sun for the same length of time and was due for replacement. Power and SWR seemed to vary at the antenna end of the feedline implying that something was not as it should have been.
The new antenna and Belden 9914 feedline were installed. We are still not sure what caused the original change, but measurements from the valley after the replacements seemed to show that the 6 dB had been regained. Part of the improvement may have been that the highest part of the dual-dipole sits higher than the old antenna. The 9914, as well as having a bit less loss at the outset, includes a foil shield which should make it less prone to dielectric contamination than the old 8X.
Some parameters measured came out as follows:
|Power Output at Transmitter||65 W||65 W|
|SWR at Transmitter||1.35||1.8(!)|
|Power at Antenna||35 - 40 W||—|
|SWR at Antenna||1.5 - 2||1.6|
In the process of mounting the new transmitting antenna, Clint discovered some PVC lying on the roof and, with a little closer inspection, found it to be our control receiver antenna! For a temporary fix, he broke the shield for about the last quarter-wave of feedline to give us a quarter-wave whip.
Deaf link receiver: A problem that we had been experiencing all summer was that the 70-cm link receiver, used to receive the signal from the Scott's Hill site, would sometimes lose sensitivity due to detuning of the local oscillator chain. The adjustable element of the filter at one stage of the chain consisted of a variable capacitor made of a disk on the end of the tuning screw and a flattened bit of wire at the end of the associated coil. When the the screw was moved in far enough to tune down to the 70-cm band, it was so close to the coil end, that tiny changes, including temperature changes to the coil, caused relatively large capacitance changes. Clint's cure was to add a small piece of brass to the coil end of the capacitor, thus increasing the overall capacitance and allowing the screw to be further out where the capacitance would be less sensitive to small dimensional changes. He applied the change to two other stages besides the one that seemed most sensitive. It may be next season before we know if this has worked.
Mysterious transmitter failures: We had been having a problem that occasionally — maybe every few weeks — the transmitter would mysteriously go off the air and return in five to fifteen minutes. A thought was that there was a self-resetting thermal chemical fuse in line with power to the IPA (intermediate power amplifier) and the fuse might behave badly if it were inside the the box where the ambient temperature might be high. This would explain the restoration after a cooling-off period.
Alas, the fuse turned out to be outside the box on a terminal strip. The IPA put out 1.8 Watts while drawing 800 mA of DC current. It seemed stable and when tuned, input current seemed to track output power.
We ran the transmitter for a considerable length of time, hoping to catch the problem happening, but were unsuccessful. The only thing that seemed at all flakey was a connecting cable that caused a bit of scratch in the signal when wiggled. We could find no position where it cut out, though. So the mystery continues.
The ever-present Eb: Since the installation of the new, frequency-stable, exciter, part of the synchronous linking system, the transmitter had developed a low-level tone around 300 Hz. Clint installed an L-C notch filter to eliminate it. The filter seemed to bring the tone below the level of audibility.
|07-19-10 Link from Scott's to Farnsworth Brought Back to Life|
Clint, KA7OEI, and the author, K7HFV, made a trip to Farnsworth Peak on Monday afternoon (and evening), July 19, 2010, to fix several believed problems. First on the list was the inability of the Farnsworth site to hear the westbound link transmissions from Scott's Hill. This prevented users who could access only the Scott's site from being heard on the system.
The problem had come and gone at least twice during the previous few weeks. Clint could hear the transmissions from the valley, so there was no question that Scott's was transmitting. Possible explanations included receiver problems, duplexer problems, and local interference on the link frequency. The antenna was shared with the eastbound link transmitter, and its transmissions could be heard just fine, so the antenna was not one of the suspects.
As soon as we arrived on site, Clint was able to hear the Scott's link transmitter on an H-T, a discovery which pretty well eliminated interference as the culprit. Next step was to check the link receiver's sensitivity with a signal generator. It turned out that it took a signal level of about -80 dBm to acheive a 12 dB SINAD (ratio of signal to noise and distortion). A normal expectation for the GE Master II receiver would have been about one microvolt or -107 dBm for the same criterion. So the receiver was needing more than 25 dB more signal than it should to deliver copiable audio.
Receivers of this series have a reputation for detuning themselves by growing metalic hairs (dendrites) on their cavities. However, the susceptible surfaces had been treated with a coating that normally solves that problem. Clint first tested for the dendrite phenomenon by giving the unit a sharp bash, something that will normally break unwanted hairs free and restore normal performance. No particular change was noted.
Sensitivity was brought back to normal by aligning the eliptical filters in the local oscillator chain. Why was this necessary when the receiver seemed to be performing adequately up to a few weeks ago? No one really knows the answer at this point, but we hope it is that the tuning was close to the edge of the acceptable range and the change to summer temperatures pushed it onto a steep slope. More observations will be necessary to confirm it is now in a stable state.
The next problem to investigate was a possible loss of one of the two final transistors in the transmitter's power amplifier. There had been several symptoms noted that seemed to point to such a failure. As part of a program of regular measurements from the valley of various repeaters' signal strengths and sensitivities, Clint had noted that '62 seemed to be about 6 dB weaker than on previous measurements. Additionally there had been some reports from the north of difficulty hearing the repeater. Finally, we had gotten another person at the Farnsworth site to read the power supplly current during transmission. It was about five amps lower than expected. The indirect evidence had been compelling that there was a problem.
The next test showed that there's really no substitute for direct evidence. A wattmeter in the line showed over 70 watts emerging from the cavities, about what would be expected if the amplifier were outputting the nominal 100 watts. The PA was just fine. The feedline to the antenna showed an SWR of 1.4:1, so there was no apparent problem with the transmitting antenna either. The other “evidence” was attributed to a change in position of the measurement antenna, an ammeter that may be aging, and normal propagation changes. Conclusion? No action required.
The next item on our list was a small problem that was created by the change last fall to the linked, voting system. If someone made a transmission when the repeater had gone unused for more than five minutes, an ID would run on top of the transmission rather than politely waiting until it was over.
For those that would like the technical details, it goes like this: The signal indicating presence of carrier on the input frequency (called “Carrier Operated Switch” or “COS”) was now routed through the voting controller before returning to the orignal controller. It turns out that the return driver was an open-collector transistor working into a .004 µF feed-through bypass capacitor. On the other end was a CMOS input expecting a fast rise from a totem-pole driver. It was now getting a slow rise determined by the time constant of the pullup resistor and the feedthrough. The slow rise was causing two different gates with different thresholds to respond in the wrong order.
We needed to speed up the rise time, either on the sending end or the receiving end. First we tried the sending end, dropping the value of the pullup resistor, but that didn't speed things up enough to solve the problem. We moved to the receiving end and thought maybe two previously unused inverter sections could be cascaded to process the signal and speed its rise time. Contrary to the diagrams on hand, however, only one unused section existed. At Clint's suggestion we created a second inverter with a plain old NPN transistor and used it to drive our one remaining inverter. It worked. (Some day maybe we should find out and document where all those other sections are used.)
Clint had an idea on how, using a very few components, we could send a tone during the repeater's hang-time which would indicate the user had talked so long as to time the repeater out. We created a number of creative noises on the repeater while trying out various configurations. Unfortunately, this attempt was foiled by some basic characteristics of the repeater. The hang time has components determined by the original controller, the voting controller, and the disciplined oscillator. Unfortunately, the original controller has no hang time after a timeout, the voter's contribution is minimal, and the disciplined oscillator does not pass audio during its contribution. Oh well.
Finally we thought we would complete a project that we had started the previous fall but did not have time to complete. Just in case we should have to bypass the voting controller and return the repeater to its pre-link configuration, we had added a front panel switch to change hang time values. The switch remained unconnected to anything. We decided to bring the necessary leads out from the original controller and connect it. In the process we decided to increase the hang-time contribution of the original controller, to be offset, eventually, by changes in the disciplined oscillator software and the controller at the Scott's Hill end.
This was accomplished without much difficulty. In the process, we had to bring some new signals off the controller card, but all the pins on its edge connector were used. Clint solved this by adding some stick-on copper foil to the non-copper side of the PC board and making use of the other side of the edge connector. Leads to a previously installed “keep-alive required” switch were brought out in the same manner (replacing a pair of Anderson connectors). Some day we must remember to document what the new pin numbers (or letters) are.
Another observation was made. Clint writes:
Before it was powered down I noted that the Disciplined Oscillator showed an up-time of 14,787,144 seconds — 171 days, 3 Hours, 32 minutes and 24 seconds, accurate to within a fraction of a second per month as this clock [is] based on the same oscillator that holds the transmit frequency to within 1-2 Hz. This would have been the last time that the Farnsworth controller had lost its power — and we think that that was back when the chip was changed in the winter to modify the hang time in lieu of our adding of the “hang time select” switch. (If someone wants to back-calculate from 7/19/2010 at about 4:30pm — and take into account Daylight Saving time to see when it was last powered-up, be my guest!)
Several times in the past we have wondered (while in the valley, 17 miles or so from the repeater site) exactly how much rack space was used by the repeater components and how much was available. I thought I would document it, thus ensuring (as per Murphy's law) there will never be a further need.
|Unit (bottom to top)||Space (inches)|
|Vocomm Power Amplifier||7.00"|
|Control, ID, and Audio Panel||8.75"|
|Astron RM-50 Power Supply||5.25"|
|Link transmitter, receiver, and duplexer||5.25"|
|On Rear Rails|
|AC supply for remote two-meter receiver||3.50"|
Note: The terminal strip that terminates the cables to the remote receiver is mounted at the top of the left-hand rear rail.
|11-05-09 Made Enhancements to '62 Voting System|
Clint, KA7OEI, and the author, K7HFV, made a trip to the Farnsworth Peak and accomplished the following:
Link Transmitter: Shortly after Clint made some adjustments on October 26, the link transmitter from Farnsworth to Scott's stopped transmitting. Clint was able to re-establish the link temporarily by operating a cross-band repeater from his home in West Jordan. This took the signal from the Farnsworth '62 transmitter and relayed it on the 70-cm band allowing linked operation to continue essentially as before.
On this (November 5) trip we found the oscillator in the GE 70-cm transmitter simply not oscillating. The first thought was that the crystal was defective. This series of GE transmitters uses a module called a "channel element" containing the crystal and the oscillator. After some swaps of crystals and channel elements and some cleanup of the solder connections on the original channel element, the oscillator started working. It was tested through the entire frequency range the crytal would allow, and showed quick starting and good stability throughout. This was not how a marginal crystal would be expected to behave. So the transmitter was placed back into service. We continued to other projects, and the transmitter stopped working a few minutes later. Some time later, it started working again as mysteriously as it had stopped. When we left the site it was working, but we are unsure if there is a second problem that may reappear.
Noise Detector Gain: It was thought that some anomalies in the voting (occasionally choosing a somewhat noisy signal from Scott's when a better one was available from Farnsworth) might be the result of not enough audio level reaching the noise detector. The circuit was modified to deliever a higher level and initial tests seemed to indicate it was working well.
Remote Control: The Voter and Disciplined Oscillator units in the new incarnation of the '62 repeater were designed so that a laptop computer could be plugged in at the site and used to read and set various parameters affecting hang times, frequencies, voting performance, and other characteristics. On this trip Clint added the YAFB (“Yet Another Feature-filled Box”). This new box that Clint built connects to the local data ports and allows them to be queried via touch tones (aka DTMF) and to report back via MCW or 110-baud ASCII. This should make it possible to fine-tune the voting performance as necessary without making trips to the mountain.
Phase Reversal: Tests indicated that the audio phase between the Farnsworth and Scott's transmitter were opposite, that is, voice peaks that caused the Farnsworth transmitter to deviate up from the center frequency would cause the Scott's transmitter to deviate down. This situation could cause bad results for mobiles traveling through areas where the two signals overlap. An amplifier in the Disciplined Oscillator was re-configured from an inverting amplifier to a non-inverting amplifier. It is believed this will improve performance in overlap areas.
Frequency Offset Change: Brett Allen, W7DBA, lives in Huntsville, an area where the '62 signals from the two mountains are just about equal. This can cause problems if the two signals seem to interfere with one another. Brett happened to be on the air the night we were making adjustments. When the two transmitter frequencies were within about 2 Hz of each other, they would slowly drift in and out of phase. This would cause loss of sylables during the period when the two cancelled each other out. On the other hand, if the two signals are too far apart in frequency, they can create a tone in the receiver equal to the difference in carrier frequencies. (We're still arguing about whether, on FM, this can be called a “heterodyne.”) With Brett available to judge, we set the offset in the 40-60 Hz range as a good compromise. Fortunately, the YAFB will allow adjustments from the valley so that we can conduct further experiments without the need to make trips to the mountain.
|10-16-09 through 10-19-09 '62 Linked Pair Put Into Operation|
During a very long weekend, the synchronous linked system — two overlapping repeaters, both on 146.62 MHz — was first put into operation. Clint, KA7OEI; John, K7ALA; and the author, K7HFV; installed and adjusted new equipment on Farnsworth Peak to complete linking to the Scott's Hill site. We can't begin to list everything that happened, in fact we probably can't remember everything that happened, but here is a summary in more-or-less chronological order:
Friday, 10-16: We installed and cabled new equipment on Farnsworth Peak which included:
A few obstacles encountered included 10-32 screws for a 10-24 rack and a lack of fuses to go in a new fuse holder.
Once the new electronic equipment was in the racks and interconnected, a great deal of adjustment was required. Audio levels required by the voter and the new exciter were different from the old equipment. Deviation going in and out of the system needed to be the same regardless of whether the chosen input was from Scott's or Farnsworth and whether one was listening to the Farnsworth or Scott's transmitter. Clipping to protect against overdeviation needed to happen at the proper place.
Things were looking pretty good until we used our ears instead of the test equipment to evaluate input versus output with voice modulation. The Farnsworth output seemed distorted and lacking in highs. The remotely located receiver seemed to be putting too high an audio voltage on the cable for the electronics at the transmitter site to process without clipping.
We went to the receive site and reduced the audio level driving the cable by about 6 dB, taking some time to help John with the new receiving antenna installation while we were there. After more fiddling with levels, we decided there was still some distortion and loss of highs when the audio was going through the new voter, but it was tolerable and there didn't seem much we could do about it immediately. After final checks on the receiver performance with new antenna and cavity, we headed home.
|Frequency of highest audio output from receiver:||90 Hz|
|Receiver Audio Output to cable w/ 5.4 kHz deviation
at a modulating frequency of 90 Hz:
|< 4 V P-P|
|Apparent gain of new receiving antenna over J-Pole:||7 dB|
Saturday, 10-17: Saturday morning, as we listened to the new system, it became apparent that the audio problems were totally intolerable. Our listening from the transmitter room had been limited by the two-inch speaker on an H-T and local QRM from Channel 18's fans. But now, listening in quiet locations with good speakers, quiet environments, and no multi-path distortion, we could see that another trip was necessary. Clint and I headed back up to Farnsworth Peak.
[To be continued]
|09-04-09 Misc. Tasks Performed on '62|
Clint Turner, KA7OEI; Bruce Bergen, KI7OM; Don Rawlins, N7YUQ; and the author, K7HFV, went to Farnsworth Peak to accomplish several things, mostly related to the future addition of a synchronous '62 repeater on Scott's Hill.
Don and Bruce cut two new holes in the wall of the building to provide an entry for additional feedlines. The holes were fitted with large pipe and capped with Microflect “donuts.” Copper ground strap was added.
Clint and the author modified the control unit to add inputs and outputs to be used later for a voting switcher that would choose the best of two signals, one from the Farnsworth receiver, the other from the Scott's site. We need to remember that a 1 kHz tone with 3 kHz deviation into the receiver produces an output level to the voting switcher of 672 mV rms.
We also investigated some apparent desensitization of the '02 receiver by the '62 transmitter. First, measurements were made at the transmitter site:
|Transmitter output power:||115 W|
|Power delivered to feedline (after cavities and isolator):||66 W||(+48 dBm)|
|SWR on feedline:||1.35:1|
|Noise floor relative to transmit carrier:
(limited by spectrum analyzer's dynamic range)
Further measurements were made at the receiver site:
|Raw receiver sensitivity||-117 dBm||0.31 µV for 12 dB SINAD|
(Cavity, preamp, filter included)
|Sensitivity through Iso-T||X||-55 dBm||Iso-T loss = 62 dB|
|Sensitivity with antenna connected||X||X||-35 dBm||20 dB degradation due to site noise|
|Inferred effective sensitivity||X||-97 dBm|
|Start of de-sense range:||X||X||-50 dBm|
|End of de-sense range:||X||X||-42 dBm|
| “Ant” = antenna connected;
“Iso” = Signal generator inserted through Iso-T
The single receive pass cavity was set for greater rejection (and, hence, greater loss). The screws that secure the coupling loops had become rusted in place, and two screw heads were broken off in the process of loosening them. We should re-tap these on a future trip. Performance was measured again after the loop adjustment (and retuning of the cavity center frequency).
|Raw receiver sensitivity:||-114 dBm||0.45 µV for 12 dB SINAD|
|Sensitivity through Iso-T||X||-54 dBm||Iso-T loss = 60 dB|
|Sensitivity with antenna connected||X||X||-44.8 dBm||9 dB degradation due to site noise|
|Inferred effective sensitivity||X||-104.8 dBm||(-44.8 dBm - 60 dB)|
|Start of de-sense range:||X||X||-50.8 dBm|
|End of de-sense range:||X||X||-43 dBm||Still about 8 dB of de-sense range. Ungood.|
Finally, measurements were made of the strength of the signal from the 146.62 transmitter at several points in the receive chain.
|Arriving on receive antenna feedline:||-16 dBm|
|Inferred isolation between transmit and receive antennas:||64 dB|
|At output of cavity:||-32 dBm|
|At output of crystal filter:||-86 dBm|
|At output of preamp:||-71 dBm|
Conclusions: Apparent desense is normally caused by one of two things (or a combination of both). The first is desensitization of the receiver by a strong signal from the transmitter. This is true “de-sense” and is a failing of the receive system. The second is a noise floor radiated by the transmitter over a range of frequencies including the receiver input frequency. This is a failing of the transmit system.
In this case, true de-sense seems quite unlikely. A signal of -86 dBm arriving at the preamp input should be way too little for a signal outside the receiver's IF passband to cause a problem.
Looking at the other possibility, we see that the transmitter noise
floor is at least 70 dB below the carrier. We start with the carrier at
+48 dBm, so the noise might be as high as -22 dBm when applied to the
transmit feedline. The 64 dB of isolation between the antennas brings it
to -86 dBm arriving at the receive antenna.
That is 19 dB stronger than -105 dBm, the weakest signal that could
produce a good (12 dB SINAD) signal. It is 25 dB stronger than the
-111 dBm (referred back to the cavity input) where de-sense “kerchunking”
first appears, assumed to be the squelch-break level. So it appears quite
possible that such a noise floor could cause the (now) 8 dB of apparent
|06-23-09 '76 Back to (near) Normal|
The '76 receiving antenna was broken into four pieces by a lighting hit in late April. A temporary indoor antenna was put into service in early May, but its performance was poor enough that only the stronger stations were able to access the repeater.
We thought we had procured a replacement antenna, a two dipole array, courtesy of Dave Williams, WA7GIE. However, when the antenna was tested it was discovered that the phasing harness had gotten water in it and was unusable. UARC President John Hardy, K7ALA, was anxious to get '76 on the air in time for UARC Field Day for talk-in to a Utah County site, and there wasn't time to acquire a replacement harness. So Clint, KA7OEI, agreed to build a J-pole of copper pipe to get us through temporarily.
The J-pole was built and tested on Friday night, June 19. Meanwhile, UARC Repeater Engineer, Randy Finch, K7SL, and John Clark, N7SFN, constructed a mount for the J-pole that would fit the clamps on the tower.
On Monday, June 22, Randy and John along with John Hardy, K7ALA; Bryan Mogensen, W7CBM; and the author, K7HFV, went to Lake Mountain and installed the antenna. Thanks particularly should go to to Bryan and John Hardy who did most of the acrobatics at the 120-foot level.
SWR measurements showed the antenna was well matched and the feedline had not suffered damage. We had hoped to do a sensitivity measurement with site noise taken into account, but that turned out not to be possible because of a malfunctioning signal generator of the author's.
Theoretically, the J-pole is close to unity gain over a dipole
which would put it 5-6 dB worse than the old Sinclair antenna. However,
if the system sensitivity is limited mostly by site noise, extra gain
may serve only to bring the noise up as much as the signal. So, in
practice, the performance may not be much different.
|04-27-09 '76 Off the Air|
It was observed Saturday, April 25, that the repeater was no longer
responding to signals on the 146.16 input. The transmitter could be keyed up
on the control link manually or in response to an IRLP connection.
A report from another user is that there is lighting damage to one of
the antennas. Repair may take awhile while a new antenna is procured
|05-30-07 '76 Back to Normal|
Clint, KA7OEI, and the author, K7HFV, were able to make a trip up to Lake Mountain last night, Tuesday, May 29. As it turned out, the problem was entirely in the two-meter receiver. A mounting screw had shorted to a supply voltage bus that ran at least two IF stages, and this had caused a decoupling RF choke to open. That, in turn, left at least two IF stages shut down. With no signals getting to the discriminator or the noise detector, the system decided there must be a full-quieting signal present. Clint tracked the problem down despite the lack of any documentation on the Motran receiver. As if to demonstrate that IRLP was working well, we barely had the cables connected back up and the repeater turned on when VK3AML in Melbourne, Australia, connected to '76 via IRLP.
We believe that everything is now working normally.
|05-28-07 New Strange '76 Problem|
A strange phenomenon was observed this morning (Monday, 5-28-2007). At regular intervals the repeater behaves as if it were being keyed up by a carrier on the '16 input frequency. It stays in this state until the repeater times out plus another minute or so. Then things return to normal for a few minutes. The cycle repeats about every 10 minutes (a value that changed radically through the day).
The source of the problem could conceivably be an external signal,
a problem with the receiver/preamp combination, or a problem with
the repeater logic. An external signal seems unlikely because no
heterodyne could be created against the phantom signal with 70 watts
from the valley. There are some arguments for and against each
possibility, so we probably won't know the cause until we can return
to the mountain. Wednesday currently looks like the first opportunity.
|05-26-07 '76 PL problem fixed|
Clint, KA7OEI, and the author, K7HFV, went to the Lake Mountain site to see why a PL (or CTCSS) tone seemed to be missing on squelch tails and IDs. The PL tone is used to identify times when the repeater's carrier is on, but there is no signal on the repeater input. Those times include situations such as IDs, squelch tails, and IRLP transmissions originating from Internet. We found that the PL was, in fact, present, but its level was low, about 350 Hz deviation.
One problem seemed to be a strong RF sensitivity in the audio board. Clint added some power supply bypass capacitors across the op-amps and reduced the problem considerably, but there was still a lot of distortion on the PL tone when the level was brought up.
Clint also noticed from the scope patterns that most of the distortion was coming on the negative swinging side of the output amplifier. Further examination showed that a transistor in a mute circuit could have gone into conduction on the negative half-cycle. We added a diode in the collector circuit and that appeared to cure the problem, or at least allow a greater level before the distortion occurred. The tone deviation was brought up to about 550 Hz, and tests with IRLP indicated it was being detected properly back in the Salt Lake Valley at the site of the IRLP computer.
Clint had brought along a homebrew receiver preamp to replace the one we had brought down last December. Measurements indicated that the site noise had increased enough that the preamp only provided about a 3 dB improvement, but it was left in the circuit.
Here are some sensitivity measurements that were made:
|Signal in dBm required for 12 dB SINAD|
Finally, for good measure, we checked the SWR on each of our feedlines. They came in as follows:
Apparently there are no particular antenna or feedline problems.
|02-06-07 '76 Back on the Air|
Sometime early in the week of January 28, the '76 repeater seemed to lose its transmit audio. Don Rawlins, N7YUQ, volunteered his snowmobiles to make a trip up the mountain. On Tuesday, February 6, Don and the author, K7HFV, made the trip. Snow depth was low enough that we were able to drive almost all the way, needing chains on the front end for only about the last 50 yards.
The culprit turned out to be an electrolytic capacitor in a 5-volt
power supply. Some TTL one-shot flip-flops were triggering on every
cycle of the more than one volt ripple, and this managed to keep the
audio muted. A 2200-microfarad 35-volt replacement brought the supply
to 4.96 volts with only 2 milivolts of ripple. A QSO with N1BL seemed
to confirm that all was back to normal.
|12-10-06 Adjustments to '76|
Clint, KA7OEI, and the author, K7HFV, made a trip to Lake Mountain on Saturday, December 9, to make a few adjustments to the '76 repeater. One of the digits of the touchtone decoder needed to have its sensitivity reduced, and the audio level from the IRLP link needed to be increased. We went up via the south (Mercer Canyon) road and were surprised how little snow there was for this late in the season, although the top of the mountain did have a general covering of several inches.
We made the intended adjustments, then checked the receiver sensitivity. We saw -94 dBm requried for 12dB SINAD, some 25 dB worse than it had been on our October trip. The problem seemed to be that the ARR preamp was intermittent, likely the result of a dirty contact in one of its bypass relays. We jumpered around the preamp and got the repeater back to about -107 dBm. We brought the preamp back with us for repair. Transmitter output was about 18 watts, which is within the normal range.
Clint noticed a crackling sound that often appeared on the signal when there was no 2-meter input signal (such as when an Internet transmission from IRLP is in progress). Clint had observed the problem before, but only during cold weather. It was traced to occasional high noise spikes from the receiver getting past the squelch gate. The problem was solved by adding a small reed relay in series with the audio line from the 2-meter receiver.
By the time we were ready to leave, it was dark. Loading up our vehicle for the return trip, we had to put up with a strong, cold wind, a precursor to the snowstorm of the next day, and the screeches from some raptor that had made a nest in the community tower.
As follow-up to this trip we need to:
|10-04-06 '76 Back On|
Clint, KA7OEI, volunteered to look at the '76 transmitter, a Yaesu FT-2500. On the first key-up on the bench, it showed a brief burst of output power and then dropped to zero. Clint whapped it on the table a couple of times and it started transmitting full power. Further whaps, probes, and long transmit periods would not get it to fail again. The problem was suspected to be in the power control circuit – a feedback loop that senses output power and makes adjustments to the collector supply voltage of a driver stage in the RF chain. Just in case it should fail again, we added a few wires from the bottom of the PC board to give us test points that could be accessed while the board was in service. We recorded the normal voltages on the test points.
We did not feel totally good about re-installing a transmitter with a likely intermittent problem, but did not have a better approach. So on Wednesday, October 4, Repeater Trustee, Brett Sutherland, N7KG; Tom Wing, W7ETR; and the author, K7HFV; put the transmitter back on the mountain. Everything appears to be working fine. If it fails in the near future, at least we will have some voltages we can check quickly on-site.
We checked the receiver performance while we were there. It looked
quite good at 0.25 µV for 20 dB of quieting. Cavity loss was about
|10-01-06 '76 Off the Air|
On Saturday, September 30, Clint, KA7OEI, and the author, K7HFV,
returned from a trip out of town to discover that '76 was off the air.
We don't know just when it went off, but believe it was operating on the
previous Wednesday. Brett, N7KG, and I made a trip to the mountain on Sunday,
October 1, and discovered that the transmitter was putting out practically
zero power. We brought the transmitter down (having neglected to take
the service manual with us) and hope to get it repaired and back to the
|11-03-04 A Few Squelch Problems|
There was an outage on '62 starting about 11 P.M. on Saturday night, October 30, and ending mid-afternoon the 31st. Another happened this morning, apparently starting sometime in the wee hours and ending about 9 A.M. We hadn't reported them in this space because they could be cured quickly by remote control, and it seemed more efficient to spend our time getting the repeater on the air than to use the time explaining why it was off the air. But some have asked what happened, so here is the story.
The '62 repeater has a remotely-controllable squelch which can be both a benefit and a liability. It allows us, without having to make a trip to the mountain, to adjust for conditions, particularly between summer and winter, and that ability helps make it practical to maintain the repeater with carrier squelch rather than requiring PL tones. However, it also means that the squelch setting can be lost if a handful of flip-flops lose power. They are backed up using a 9-volt battery which is usually adquate for the job. But if those flip-flops reset to zero, then the squelch will be fully open and the controller logic interprets that as a continuous carrier on the input. That will cause the repeater to time out after three minutes and stay off until the squelch closes.
On Saturday night/Sunday morning KSL made some major power distribution changes on Farnsworth Peak, converting from a 240/120-volt system to 480/277/208/120. The repeater, as best we can reconstruct the situation, was powered off for something in the neighborhood of six hours. We suspect that killed off the battery. It had not been changed for two years or more, and had had to supply power during some long periods during our power supply troubles of last summer. When power was restored we simply reset the squelch remotely and the repeater was back on, but it was vulnerable to brief power interruptions. Apparently one of those happened in the wee hours of Wednesday morning, November 3rd. Brett, N7KG, reset the squelch on his way into the office.
Our repeater engineer, Randy Finch, K7SL, tells us that arrangements
have been made to swap out the battery. We hope that will minimize
|07-29-04 Reconsideration on '62's Amplifier|
We arranged to have the amplifier brought down from the mountain. On Tuesday, the 27th, Brett Sutherland, N7KG, repeater trustee, and the author, K7HFV, powered it up and ran it through its paces. No short circuit was apparent. We were expecting to see fried semiconductors, but instead the amplifier powered up without any current spikes and was well-behaved. Driven with about three watts from a handheld, it produced about 90 watts of output and drew approximately 11 amps from the supply. Inside, there was no indication of anything overheated. As far as we could tell, this amplifier had nothing at all wrong with it.
On Thursday, July 29, we sent the amplifier up with Don, N7YUQ, who was making a trip to the mountain for other reasons. Guided by Repeater Engineer Randy Finch, K7SL, by telephone, he reinstalled the amplifier and it seemed to work just fine. So, at this point, the repeater is back to normal output.
So why did the fuse blow immediately when everything was hooked up on the 26th? That remains a mystery. Maybe there was some unseen short in the hookup. Maybe we had a defective fuse. Maybe there is still some latent problem in the power supply. Maybe the mountain Gods were having a bad day. All we can say is it seems to work fine now.
We'd like to remember that the RM-50 power supply takes 20-mm
glass fuses, 10 amps.
|07-26-04 (evening)'62 Amplfier is Also Damaged|
Murphy's Law is still in effect. When the power supply was returned to the mountain and connected, it immediately blew a fuse. When the RF power amplifier was disconnected, the power supply ran happily. Apparently the amplifier was loading the supply with something close to a dead short. So now the amplifier is on its way to the valley for diagnosis and repair.
The '62 transmitter consists of separate exciter and amplifier units. The repeater continues to be on the air running the exciter only with about 3 watts of output.
Repair of the amplifier will almost certainly require ordering
new semiconductors, so it is likely to be a week or more before the
repeater is restored to full power.
|07-26-04 '62 Power Supply Believed Repaired|
The power supply is repaired and waiting for a ride up to the repeater site. Clint, KA7OEI, writes:
The [Astron] RM-50M power supply for '62 has been fixed.
Both diodes (originally 1N1184A -- 40-Amp, 100-Volt -- and had been previously replaced with their NTE equivalent) were shorted.
These diodes were replaced with type "70HFR40" which are 400-volt, 70-amp units. Note that these diodes are "R" types with their packages opposite polarity from the standard, so the wires connected to the diodes were appropriately modified.
After replacing the diodes, John Lloyd observed that the fuse popped after about 10 seconds of operation. The pass transistors were removed and checked -- but no problem was found.
The supply was left on (unloaded) over the weekend with no problem. It was then load-tested to about 25 amps and displayed an intermittent dropoff of voltage - down to about 10 volts or so. John quickly related this to flexing of the circuit board.
Re-flowing of all solder joints on the circuit board seems to have fixed this problem and the supply was loaded to about 30 amps with a voltage sag of about 30 millivolts and about 5 millivolts of AC ripple.
New silicone was globbed over the (external) original diode stubs to reduce the likelihood of shorting. Note that if that were to happen now with these "reversed" diodes the short would occur on the transformer (AC) side of the circuit instead of the rectifier side and would probably merely blow the line fuse.
It is worth noting that the original diodes could have been easily rewired such that their cases would be grounded, using the centertap as the "+" lead.
We may want to remember the following replacement options, should the need arise sometime in the future:
|NTE-5982||100 V||40 A||$4.73|
|1N1186||NTE-5986||200 V||40 A||$4.83|
|1N2131A||NTE-6026||200 V||60 A||$7.25|
Package is 1/4-20 stud mount, DO5.
|07-22-04 '62 Squeal Gone, Transformer OK|
The '62 repeater is now stealing power from the 449.1 repeater instead of getting it from the temporary supply. This seems to have cured the squeal. There is still not enough power available to run the amplifier, so the repeater continues to run in the 3-watt mode.
The power supply arrived in the valley and John Lloyd, K7JL, has been checking its operation. The transformer apparently survived and only the main rectifier diodes failed. They will be replaced with ones having higher ratings.
A gentleman named Dave has been making bogus
announcements about the repeater's condition and his part in fixing
it. Unfortunately, most of his information is wrong. Please ignore
statements about the repeater from this person; he is not associated
with the repeater committee and, as far as we know, has never even seen
the '62 repeater.
|07-20-04 (evening) '62 Back On At Reduced Power.|
We were able to determine that the power supply failed and will have to be brought down for repairs. We don't yet know if the power transformer survived. If not, we may have to wait for a new one to be ordered in.
The repeater is now running on a temporary
supply and is running without the power amplifier, meaning transmitter
output is about 3 watts. The temporary supply seems to have some kind
of oscillation that is putting a squeal on the transmitted signal. The
repeater is usable even though the squeal is a bit annoying.
|07-20-04 '62 Goes Off the Air|
At about 7:20 A.M. on Tuesday, July 20, '62 went off the air. The ATV
repeater, which runs on the same 12-volt power supply as '62, is also
down. This implies a problem with primary power to the rack or a problem
with the power supply. One possible scenario is that the crowbar
circuit in the power supply was tripped by line transients. With the
crowbar trying to short the output but the regulator still intact, the
regulator would go into foldback current limiting. If this is the case,
the cure would be simply to turn off the supply to reset the crowbar
|02-26-04 '62 Power Amplifier Back in Service|
On Saturday, February 7, Brett, N7KG, and the author, K7HFV, went over to KA7OEI's house and pretended to supervise while Clint changed all three transistors in the power amplifier's RF chain. We thought that running a little less output power might give us better life from the finals. The amplifier is rated at 100 watts but we had been running it closer to 130 watts output. We found a tuning compromise that gave us good efficiency and good balance between the two final transistors as well as output power near the rated value. The only remaining problem was finding a way to get the amplifier back on the mountain.
For more information about the amplifier and its failure, follow this link.
On Monday, February 23, Randy, K7SL, our Repeater Engineer, found a ride for the box and several handoffs were made. On Thursday, the 26th, the repeater went back on the air at full power. Readings at US Satellite showed a received signal level of -58 dBm after the change compared to -74 before. If we assume the amplier is now putting out 105 watts (which is close to the measurement made on the 7th with the amplifier on the bench), that would put the exciter power (the output level we had been running for about the last month) at 2.6 watts.
We can't help commenting that we hear a lot of operators say
"power doesn't matter," but there sure were a lot of complaints
when the power was down. The repeater had been running a power level
comparable to a handie-talkie, so this should have given H-T users
a chance to hear what they sound like at the repeater from various
|01-31-04 '62 Power Amplifier Needs Transistors|
The power amplifier for '62 has been brought down to the valley
and put on the bench. It appears that all three RF devices will have
to be replaced. There is an MRF-239 driver transistor and a pair of
MRF-224 final amplifiers. Also, some resistors in the 100-ohm 2-watt
neighborhood will be needed for the collector balancing circuit.
Clint will order the parts and we'll take further steps when they arrive.
|01-17-04 '62 Power Amplifier Dies|
Sometime around midday on Saturday, January 17, the repeater output dropped to the point where it was just barely copyable on a good fixed station antenna in the Salt Lake Valley. Further investigation showed the power amplifer had failed. We arranged to get the exciter patched through to the antenna and the repeater was usable once again. However, it may be noticeably weaker in many marginal areas. The system is probably delivering about two to three watts to the isolators and cavities instead of the normal 120.
We hope to be able to find a way to retrieve the amplifier and repair
it in the next several weeks.
|08-16-03 Maintenance on '76|
Clint Turner, KA7OEI, and the author, Gordon Smith, K7HFV, made a trip to Lake Mountain on the 16th to do four things:
The strange-sounding ID turned out to be caused by one of three capacitors we had suspected before going up. The electrolytic on the output of an LM-309K had lost enough of its capacitance to allow the regulator to oscillate. There was a 3-MHz signal running around the board that caused the behavior to change radically depending on what sort of conductor was close to the board and at what distance. Changing the capacitor cured the problem.
A modification was necessary to make "autopatch mode" (which would also be used for uplinking IRLP audio) resettable. A 2N3906 was added to the 555 timer which causes timeout resetting of autopatch mode. This made the circuit into the famous "missing pulse detector" allowing the timer to be reset before timing completely out.
The power supply that was removed two years ago (see the 7-7-01 incident) had been repaired and was restored to service, freeing the 20A Astron supply for later use on Scott's Hill.
Performance measurements came out as follows:
|Two-meter receiver||Sensitivity||0.35||µV||12 dB SINAD|
|Ch element #||TLN1086A|
|Crystal Frequency||12.846666||MHz||High side, X12, 8 MHz IF?|
|70-cm receiver||Bandwidth||26.4||kHz||6 dB down|
|Bandwidth||47.0||kHz||40 dB down|
|Sensitivity||1.0||µV||12 dB SINAD|
|Transmitter||Power output||17||Watts||Medium setting|
|Two-meter receive feedline||SWR||1.10:1|
|Two-meter transmit feedline||SWR||1.30:1|
|70-cm receive feedline||SWR||1.15:1|
|Autopatch timer||Delay time||5:17||min:sec||Initial run|
|Delay time||4:30||min:sec||After reset|
|06-21-02 Interference Has Not Returned|
It's probably time to tell the end of the interference story. We didn't really know exactly when the story was complete, but now that we've gone two months with no problem, it's probably safe to say it's over.
The interference stopped within a day or two of the time John contacted Mr. Pluim. That was good news, but there was no guarantee the relief would be permanent.
We found out about the next chapter in an odd way. Another ham, Don Wood of Bountiful, who runs a service shop, contacted John after Mr. Pluim brought in the offending radio for checkout. It was a 35-watt mobile strip turned down to about three watts. That, alone, is enough to cause cleanliness problems in many radios. However, Don (who knows his business well) found nothing wrong with this one. When the strip was operated into a 50-ohm load, there were no harmonics or parasitics that were above normal limits.
Presumably, the radio was re-installed, and the problem went away. We don't know exactly why. It is likely that something in the antenna system was presenting an odd, reactive load to the transmitter and causing the problem. Mr. Pluim might have found and fixed a corroded joint, or perhaps it was a connector problem and just the act of disconnecting and reconnecting the radio was enough to remove corrosion and make the connection solid again. A connection problem to the power source is another possible culprit. In any case, we are simply happy that we no longer have the interference.
|04-04-02 Licensee of Interfering Transmitter Contacted|
John, K7JL, was, indeed, able to confirm the source of our interference and get more information. The signal emanates from a small Yagi (about five elements) on top of the southern water tower and aimed south-southwest (very close to straight at Farnsworth Peak). The primary frequency (that is, the one on which a signal is supposed to be transmitted) is 173.2625 MHz. The licensee is listed in the FCC database as the Hooper Water Improvement District.
Spectrum analyzer measurements showed not only the spurious signal near 146.02, but another, even stronger spur at 159.645 (a region allocated to public safety communication). On some bearings, the 159.645 signal was actually stronger than the one on the intended frequency.
Here are some of the measurements: (Note that, because the numbers are negative, smaller values indicate stronger signals.)
|Approx. Bearing from Yagi|
|173.2526||-40 dBm||-36 dBm||-62 dBm|
|159.645||-60 dBm||-42 dBm||-55 dBm|
|146.02||-70 dBm||-60 dBm||-64 dBm|
John tried calling the licensee of the station and, after several referrals, spoke with Mr. Wim Pluim of US Filter Company. Mr. Pluim seemed to appreciate the problem and said he would order a new transmitter unit immediately and replace the existing one as soon as possible. There is hope!
|04-03-02 Interference Source Located|
After an evening of roaming the streets of Roy, we believe the source of the "sweeping kerchunk" has been located. Most of us had given up trying to hear the signal from somewhere other than the mountain. Attempts had failed in the Salt Lake valley, Utah valley, and Tooele valley. But, after mentioning the problem on the Sunday night UARC Information Net, we got a tip from Kim, N7WLW, that he could hear the signal full-scale near his home in West Haven (west of Roy).
Clint, KA7OEI, had a newly procured (and modified) Dick Smith Doppler Direction Finder. Armed with that unit and Kim's tip, Clint and the author headed north. We heard nothing until we were past Clearfield, but then, sure enough, the sweeping signal appeared. The sweep seemed to be remarkably well-centered on 146.02, starting at about '01 and ending at about '03.
We followed the freeway into Ogden, but found the signal-strength peak to be near the Roy exit, so we went west into Roy.
We finally determined that the signal appears to be coming from one of two water tanks just east of Roy West Park (about 4600 S. 2900 West). That area is just a few blocks west of Roy High School. The transmitter is likely used to telemeter such parameters as water level, flow rates, and valve positions. Apparently it has at least one output on a frequency other than the intended one.
The next step will be to correlate the spurious signal with a primary signal and try to locate the licensee. John Lloyd, K7JL, plans to take a spectrum analyzer into the area today and check out a frequency in the FCC's database that seems most likely to be the culprit.
|03-28-02 "Sweeping Kerchunk" Signal Observed Keying '62|
Starting about Friday, March 22, users of 146.62 started reporting a strange, recurring signal keying the repeater. Generally the problem manifests itself as a brief keyup of the repeater every 25 to 45 seconds. The interference is active most of the time, although it sometimes disappears for several hours.
So far, no one has succeeded in hearing the signal on the 146.02 input frequency. Numerous stations in the Salt Lake Valley have listened (some with Yagis and very sensitive receivers) as well as others in Utah Valley and in Tooele County. This behavior seems to suggest about three possibilities for where the source of the signal might be:
Heterodyne tests show an interesting property of the signal: it is not on a stable frequency but is sweeping through the repeater's input frequency. It starts below 146.02, moves up through the repeater's passband, and then leaves the passband as it moves up in frequency still further. The frequency vs. time characteristic is very linear. We have no way of knowing just how far below 146.02 it starts and how far above it goes.
Another intersting property is that the signal sometimes seems to have weak data modulation. It sounds somewhat like AFSK with resting frequencies near 1100 and 1300 Hz. On some passes, only the higher tone is heard. On others, a transition from high to low is heard. On still others, there seem to be actual data being sent.
Observers disagree on whether the signal strength is constant. The sweeping property of the signal causes the repeater quieting to vary as the signal moves through different parts of the passband.
The signal does not seem to affect the 146.94 repeater which is only about a quarter-mile south of the '62 site. This could mean the signal source is very close, physically, to the '02 receiver, or it could simply mean the sweep doesn't extend that far in frequency.
So far, no input observations have been reported from Davis County and points north. If anyone happens to be up that way while the signal is apparent on the '62 output, listen on '02 and let us know if you can hear anything with the same pattern (about a one-second burst every 25 to 45 seconds).
The interference is not strong enough nor long-lasting enough to impair communications through the repeater. It is mostly an annoyance to those monitoring '62. So the solution may have to wait until summer when we can drive to the site with DF gear and identify the source of the signal.
|03-13-02 '62 Returned to Normal Status|
Thanks to our generous benefactor, I (K7HFV) was able to get another snowcat ride up to Farnsworth Peak. We had considered snowshoeing up, but we could not find a volunteer to carry the 40-pound power supply in his pack. Conditions on the ridge were rather exciting. Fresh deep drifts sometimes made the road difficult to find and the winds were fierce, making the snow come at us virtually horizontally.
KA7OEI had repaired the ailing power supply. The 35-amp diodes had both shorted. I was able to install the repaired supply and put the amplifier back in line. We now show about 80 watts going into the feedline (after cavity and isolator losses).
|Normal exciter readings:||RF = 0.4||DC = 0.7|
|Conditions between bandpass cavity and notch cavities:  ||Power = 100W||SWR = 1.85|
|Conditions between last notch cavity and feedline:||Power = 80W||SWR = 1.90|
|Normal power supply status in transmit:||13.67 Volts||22 Amps|
|Remote squelch times:||Enable: 13 seconds||Disable: 50 seconds|
|Power feed to receiver site:||53 Volts AC|
|Fuses required for power supply:||Littlefuse GMA-10|
|Original power supply rectifiers:||1N1184A||100 V @ 40 A|
|03-05-02 Accident Puts '62 Power Supply Out of Service.|
The ATV gang was doing some work to restore the 23-cm receiver. In the process, a cable shorted to the rectifier diodes in the '62 power supply (they are exposed on the back of the Astron unit) and took out the supply. Temporarily, enough power was stolen from the 449.1 repeater to run '62's exciter but not its amplifier. So '62 will be in the QRP mode until the power supply can be repaired and another trip arranged. Perhaps this will give H-T users a chance to see how they sound from various locations, as the repeater will be about the same strength coming back to them.
A problem with the ATV 23-cm receiver has been fixed and the transmitter side of the ATV repeater is again being used for NASA Shuttle Mission coverage. The '62 transmitter was put back on frequency. (It was running about 1 kHz high.)
|02-27-02 '62 Restored.|
In a roundabout way we found out about a ham who was making a snowcat trip to Farnsworth for his employer. With some clever juggling of load, they were able to make space for the author, K7HFV. Time on top would be limited, however.
I would like to thank this ham profusely, but for reasons that I'll leave to your imagination, it is probably best if we don't mention his name here.
The culprit was found to be a Sola constant-voltage transformer which feeds the step-down transformer which sends power down to the receiver site. It was putting out about 13 volts instead of the customary 120. The Sola is used to protect against power line spikes, brownouts, and over-voltage, but is not necessary for the repeater to operate. In fact, '62 operated for over a year without it at the time we biggie-sized our 12-volt power supply to accommodate ATV. So, I just bypassed the Sola and the receiver came back to life. We will have to decide if the transformer is worth replacing. (Anyone out there with an excess Sola? It only needs to handle about 100 watts.)
|02-16-02 '62 Fails.|
Sometime between about 1800 and 1900, the 146.62 repeater stopped repeating. Some quick checks with the personnel on the mountain indicate that the power supply and transmitter are working normally, but there has been a failure in the receiver, the cable from the receive site, the receive antenna, or the squelch logic.
Further diagnosis and repair will have to wait until we can make a trip to the mountain. Those trips aren't exactly trivial this time of year. Bad weather is predicted for the next several days. After that, we may be able to work out a snowshoe or snowmobile trip.
|07-07-01 Hum on '76 Cured.|
A bad 120-Hz hum troubled '76 for most of the past week. Clint, KA7OEI, and the author, K7HFV, went to the site and replaced the power supply with a 20-amp Astron that was removed from Farnsworth a few years ago. We then brought the existing supply back to the valley for troubleshooting.
The supply that had been running the '76 transmitter and receiver for the last several years was built by Steve Kleinlein, AE7G, in the late 70's for use on Farnsworth Peak. (The later addition of a larger amplifier and the ATV repeater made a larger supply necessary on Farnsworth.)
Steve's supply was designed for maximum reliability on a site where winter access was somewhere between difficult and impossible. It had a large main supply and a small backup supply adequate to run the old '62 without the power amplifier. The two supplies were diode OR'd together.
The problem turned out to be that the big supply had its current limiting misadjusted so that it could supply only a few amps. The two supplies had been working jointly to supply the load. This worked fine up until this week when a filter capacitor in the backup supply failed.
The main supply is now working fine on the bench, and the backup will be repaired as soon as we can locate a replacement capacitor (20,000 uF at 30V) that will fit the mounting clamp.
|10-23-00 '62 Finally Back to Full Power.|
We made another attempt to install the power supply and checked several times before leaving home that we actually had the new supply with us. It installed without difficulty. Retuning the amplifier, we were able to triple, approximately, the power output. (Checks in the valley later showed a 5-6 dB improvement.)
Clint added to the ATV repeater a circuit that allowed sending the audio at narrow deviation on the video carrier. (This is in addition to regular wide-FM audio signal, 4.5 MHz above the video carrier.) This feature should be helpful during Shuttle mission retransmissions. It allows the audio to be received on a regular communications transceiver tuned to the video carrier frequency at 426.25 (or perhaps 426.30 or 426.35 -- tune for minimum sync buzz).
The ride back down from the mountain was a bit exciting. It was about 9:30 P.M., quite dark, moonless, and it appeared that the author's (K7HFV) vehicle had lost its alternator. We decided that preserving locomotion was probably more important than having the best of light. So we doused the headlights and got along with an LED flashlight powered by three AA cells, duct-taped onto the winch control box. It was supplemented by a lantern-style flashlight held out the window by the author to highlight a few of the larger obstacles. Amazingly, we got down without running off any of the 500-foot cliffs and without getting collared by the Kennecott security guard who must have been sure we were trying to slip out with a pilfered elk.
|10-16-00 '62 Transmit Signal Now a Bit Stronger|
We headed for the '62 site with the intent of finishing three jobs: installing a new transmit antenna, changing the final transistors in the two-meter amplifier, and installing a new power supply adequate for running the ATV transmitter and '62 at full power.
We were successful at the first two. A newly constructed J-pole went into place when its mount was supplemented by some scrap rubber pieces found on the site. Clint successfully changed the final transistors, and it became immediately apparent that the balance was greatly improved.
When it came to changing the power supply, though, we discovered we had neglected to bring along one component required for the job: the power supply. We were still able to leave the amplfier putting out more power than it had when we arrived. Later measurements in the valley indicated about a 6 dB improvement between the antenna and the the amplifier.
|08-20-00 '62 Back to Normal|
We tried to blame it on the thunderstorm, the ATV repeater, and the repeater's horoscope, but none of those worked. The problem that took the repeater off the air sometime on the 18th turned out to be a cold solder joint between a connecting wire and the terminal of an edge connector. It had held together for 20 years but picked this summer to break loose.
We determined that an important control detection signal was being created properly on the audio board, but never arriving at the logic board. We went to place the scope probe on the terminal and found the wire that used to connect to it floating in the air. Somewhere, we seemed to remember an axiom that wires in harnesses are usually supposed to connect to something. The solution required only a few minutes with a soldering iron.
Clint (KA7OEI) then spent a couple of hours devising a way to reduce the heat generated inside the ATV repeater's audio transmitter box. The final solution involved a string of diodes from the author's (K7HFV) toolbox used as a distributed power resistor. The aural transmitter now seems to be much happer than before.
|08-19-00 '62 Control Problem|
The thunderstorms on Friday night apparently damaged something in '62's control system. The repeater itself works fine, but it is running open squelch. The control logic interprets this to mean a signal is continuously present on the input, so the repeater stays timed out. We hope to be able to go up sometime Saturday and see exactly what is going on.
|07-28-00 '62 Back On; Problem May Recur|
Getting the repeater back on was remarkably easy: we just turned the power supply off and back on. However, deducing why that was necessary is a little harder.
We (KA7OEI, NY4I, and K7HFV) found the power supply in the state where the crowbar circuit had tripped and foldback current limiting was holding the output to about two volts. (Crowbar circuits are safety circuits used in linear power supplies to keep from destroying the equipment they power in the event the pass transistors short out. The crowbar circuit shorts the output voltage if it senses it has passed a high threshold point.)
Just why the crowbar tripped (with the pass transistors still doing just fine) is a bit mysterious. Such a problem is not uncommon with linear power supplies, but this particular one had never done it before in the ten years or so it has been operating '62.
Several factors make us suspicious that heat may have been part of the answer:
We checked the repeater's power amplifier and replaced a bad transistor used to sense input RF and start the amplifier's fan. In the process, Clint noticed that there is considerable imbalance between the two transistors that comprise the amplifier. This suggests loss of emitter sites in at least one of the transistors. It also means the transistor carrying the larger part of the load is in danger of overheating and failing.
To (we hope) reduce the danger, we left the amplifier running somewhat less power (about 40 watts instead of the normal 130) until a new set of matched finals can be obtained. We left the balance somewhat better than it had been when we arrived.
The reduction of output power also reduces the load on the power supply, and, hence, the amount of heat it must dissipate. If our crowbar trip was, indeed, heat-related, this change should make it less likely to happen again.
Of course, this isn't a final or satisfying answer to our problem. With luck, we will be able to take several steps before the season is over:
Clint added a 23-cm preamp to the ATV repeater and adjusted levels. It was then able to relay a good-quality copy of a test signal sent up on 23 cm from the US Satellite site. This makes the machine virtually ready to relay shuttle transmissions later this year. (The 70-cm receiving system has yet to be installed.)
|07-27-00 '62 Off; Power Supply Suspected|
The 146.62 repeater went off the air at about 1708 on Wednesday the 26th. The partially-installed ATV repeater also appears to be off and probably went down at the same time. Since the two share a power supply, this probably implies a problem with the supply or some module having shorted and forced the supply into foldback limiting.
KA7OEI and the author, K7HFV, will try to arrange to go to the site after work this evening and find out what happened.
|07-07-00 Control Problems Possibly Resolved|
The author, K7HFV, and Clint Turner, KA7OEI, spent most of Thursday evening at Farnsworth Peak. We changed some ICs and made some adjustments to the '62 control system. We'll know after a day or two of operation if it is now more immune to outside influences.
The site was home to firefighters watching the brush fire on the west side of the Oquirrhs. The crew for the television station told us they had had to evacuate the site the previous night when the fire appeared to be getting too close.
Installation of the Amateur Television (ATV) repeater is underway. Some adjustments to the control system were made and a link was installed to receive the FM TV signal from the University of Utah Amateur Radio Club (UUARC).
|07-04-00 Control Problems|
The ATV gang spent most of Monday and Tuesday installing equipment for the ATV repeater. This necessitated rearranging some of the pieces of equipment that compose the '62 repeater. This rearrangement (and possibly differences in the amount of incidental RF getting back into the box) conspired with a wiring mistake on the control board to produce problems with the remote control system. The main problem is that the repeater keeps going off, partly due to suspected interference from another service on the control frequency. During hours when no one is at the control point to reset it, the repeater remains off. We hope to be able to go up Thursday night (7-6) and readjust things to solve the problem.
|06-23-00 Deaf Receiver?|
Starting on Tuesday, June 20, the '62 repeater seemed to go mostly deaf. Fixed stations with 50 watts and a good antenna could get in, but most others could not even key the repeater up. We speculated on failures of the receive antenna, the preamp, the receive cavity, or just noise from some other radio service.
When the problem persisted through Wednesday, we made plans to go up to the site Thursday night and see what the problem was.
I (K7HFV) left work on Thursday with the intent of quickly loading some test gear and heading for the mountain. But then I heard a station who barely gets into the repeater under good conditions make a full-quieting transmission on '62. At that point I realized that something must have changed.
When I got home, I called Clint, KA7OEI, who was still at work and had access to calibrated signal generators. He measured the amount of signal that '62 required to produce reasonable quieting and confirmed that it was back to normal and that two other nearby repeaters that had suffered about a 10 dB degredation were also back to normal. Interference to our control receiver on another band had gone away as well. At that point it was pretty clear the whole problem had been interference and not anything wrong with the repeater.
We got Randy, K7SL, to call the operator on the mountain and find out what had just changed. The story was interesting.
Among the many multi-kilowatt stations on Farnsworth there is also a 50-watt "low power" FM station. Their call is KUUU and they operate on 92.1 MHz playing rap. Clint has suggested a better callsign would have been KRAP -- to be pronounced, of course, Kay-Rap. However, during a check earlier in the week, the FCC had decided that the other pronunciation was more appropriate. Their spurious emissions were so bad they were ordered to shut down until the problems could be fixed. You might think their shutting down would fix our problem, but that is actually when it started.
They just turned the transmitter off and left their solid-state final (now an unbiased collector junction) connected to the antenna. One of the problems in their station design was they had no cavity or bandpass device of any type between the transmitter and the antenna. In addition, they had no isolator despite site regulations. So their final amplifier stage made a great high-level mixer. It happily mixed all the kilowatts of FM and TV signals creating a plethora of new frequencies. Unimpeded by a cavity, these new signals were sent directly back to the antenna to be re-radiated around the site.
We estimate the interference was so strong it was about 40 dB above the normal site noise floor. That means that '62 users had to have 40 dB more signal to get into the receiver. (A 40 dB improvement on the transmit end is fairly easy to achieve -- just run 10,000 times your normal output power.)
On Thursday they switched to their new antenna (and, hopefully, cavity) just about the time we heard the first good '62 transmission. Magically, the noise disappeared.
So there you have it -- the incredible self-repairing repeater.
|For more information about the UARC repeaters and their sites as well as technical information about how they work, go to The UARC Repeater Page.|
Page last updated: 06-21-2020