Right after Tropical Storm Irene, it was noted that the STL signal strength at the WHUD transmitter site was low. Normally it was 300+ µV, but now reading around 100 µV, which is a problem. Upon further investigation, it was revealed that the STL transmitter on the intermediate hop had higher than normal reflected power.
Time to call the tower crew.
The STL transmit antenna for WHUD’s STL (WPOU464) hop is a Scala Paraflector (PR-950), mounted at the 280-foot level on this tower:
Scala PR-950 on a guyed tower
The fact that it happened after a major storm and the transmitter was showing higher than normal reflected power indicates a problem with either the antenna or the jumper between the 7/8″ Cablewave coax and the N connector on the antenna. A measurement with a spectrum analyzer shows very high return loss:
WHUD STL antenna return loss
This shows the distance to fault 413 feet, with a return loss of -7.4 dB. That distance is either near or at the antenna and -7.4 dB indicates a lot of reflected power. We had the tower climber take apart the jumper connections and terminate the jumper with a known good 50-ohm load. The return loss did not change. We then had him swap out jumpers and reconnect to the antenna. That did the trick:
WHUD STL antenna with new jumper
Much better, most of the power is now being radiated by the antenna, the VSWR is 1.02:1. The impedance bump at 51 feet is a sharp bend in the coax where it is attached to an ice bridge. Reconnecting the transmission line to the transmitter and turning it on confirms that all is normal again. The problem with the jumper was found in one of the connectors, it was full of water.
Water-contaminated Andrew flexwell connector
I cut away the boot, water had entered the connector from the back because waterproofing and tape was not applied all the way to the coax. This was installed in 1998 when the station moved from Peekskill to its current location in the town of Fishkill. The fact that it happened now in the nice weather when Mt. Beacon is still accessible and not in the middle of winter means the radio gods are smiling on us.
The case of the blown fuse, or rather the blown up fuse:
Blown 10 amp fuse on Harris SX5 PA board
F32 is blown into small bits and had to be vacuumed out of the bottom of the transmitter. The reason why is the pair of MOSFETs connected to that circuit were shorted. Of course, the reason for the shorting of MOSFETs needed to be investigated. What I found was on the underside of the PA board where the brass stand-off attacked the toroid combiner board, the nuts attaching the stand-off to the combiner board were loose and there was a big arc mark.
I tightened everything up and replaced the MOSFETS, marking them with a pen in case they short again, in which case the drive section needs to be closely examined.
It was a hot day, it was a cold day. The tube transmitter was running, the solid-state (HD-1) transmitter was off the air. The books show that the company has deep pockets, but the accountant has short arms. And so it goes. In a sordid, yet familiar tale of leaping three-quarters of the way across a river, the builders of this transmitter site seemed to think of everything except the cooling requirements for a 35 KW FM transmitter.
Instead of installing real commercial AC units, someone decided that 34,000 BTU window units were the way to go. At one time, there were eight of those units, all single phase 240 volts sucking down gobs of power and freezing up when the outside temperature dropped below 40°F. This was always a problem but became more so when we took over the maintenance of this site. When there was a full-time engineer, his time, apparently, could be wasted running back and forth turning the window units on or off in the winter as required. Now that a contract company is doing the work, it becomes cost prohibitive to require such things.
Therefore, the time had come to make a change. To that end, six of the 34,000 BTU window units were removed from the building. Two of the existing holes in the wall were used to create an emergency cooling system, consisting of a 4,292 CFM fan and a couple of shutters. Two other holes were used for the new air conditioners and two holes were blocked up. The remaining two window units were left in place in the combiner room, which is a separate cooling zone.
Bard 5-ton wall mount AC unit and cooling fan shutter
The new ACs are five-ton wall-mount Bard units. These are three phase and should be more than enough to keep the transmitters cool. Here is how I arrived at that conclusion:
The entire building load when the main transmitter is running at full power, without the transmitter room air conditioning, is 60 KW.
All of the building loads except the transmitters go through a single-phase panel.
The load on the single phase panel is 10 KW, thus the transmitter load is 50 KW (this 10 KW is mostly the single phase AC units in the combiner room)
The TPO is 32 KW, therefore the transmitter is generating 18 KW of waste heat.
One watt-hour = 3.412 BTU of energy, thus
18,000-watt-hours equals 61,416 BTUs
One ton = 12,000 BTU, thus
61,416 BTU ÷ 12,000 BTU = 5.118 tons
That will take care of the main transmitter waste heat. The HD transmitter generates another 4,000 watts of waste heat or 1.37 tons. The combiner is in another room and doesn’t factor into the calculation. The rest of the equipment is inconsequential, adding up to less than 100 watts.
The solar gain is more difficult to calculate, as it is based on the building structure, the type of construction, and the heat gain (loss) through the walls and doors. This building is concrete block, insulated, and has no windows. It is unshaded, however, it is painted a light color. All in all, the solar gain should be less than two tons on a hot day. Therefore the total AC load should be 8.25 tons or less.
Bard 5 ton wall mount AC unit
All that is left now was to install the things. Just pull up the truck and use a crane to lift them in place, except, no; that plan won’t work. This is the transmitter site at the power plant and the 138 KV lines overhead precluded any lifting with a crane. We instead had to build ramps and move things around on large-hand trucks. One unit is installed on the rear of the building, the other on the front. It required several days to make the ramps and four people to muscle the things into place.
The bottom air intake holes needed to be cut out for the new units. Cutting into the concrete block while the BE FM 35A was running proved to be another challenge. We used several sheets of plastic, shop vacs, and extra air filters on the transmitters to keep the concrete dust out of the PA cavities and motor bearings.
Plan B cooling consists of a 4,292 CFM Venturi fan mounted on the rear wall of the building. The fan is controlled by a ceiling-mounted thermostat set to 95 degrees. If the AC’s fail, the ceiling temperature will rise and the fan will turn on.
Transmitter site emergency cooling fan
The room volume is about 3600 cubic feet, therefore this fan will change the room air about once every 60 seconds or so. It is not the best plan to move humid, potentially dirty outside air through a building, but it it keeps the station on the air while the main AC units are being repaired, then so be it.
The entire system went on line last week and is working well.
I give you joy, the unmitigated joy, and sheer pleasure of the Harris SX 5 AM transmitter. This particular unit dates from 1984 and is installed at WUPE in Pittsfield, MA. It has a few issues of late.
Harris SX 5 medium frequency AM transmitter
The first of which is the unbalanced or out-of-ratio condition of the PA current and voltage. When changing power levels, the PA current, and voltage are supposed to track together. When they do not, it is an almost sure sign that one or several of the MOSFETS in the PA are shorted. Shorted IRF-350 MOSFETS are indicated by blown fuses on the PA boards and should be replaced in pairs. The reason for the damaged devices also needs to be investigated. It is entirely possible that the site receives a lot of lightning, which can cause this damage. It could also be heat related, as the site is not currently air-conditioned. The other possibility is under drive conditions.
The MOSFETS turn on and off at a rate of 1/(carrier frequency (hertz)) times per second. If they are under-driven, they will go full-on and short-circuit. The minimum drive is 27.5 volts peak to peak, anything less than that is marginal and can lead to the destruction of the PA devices. Underdrive indicates an issue with the oscillator, which has its own set of peculiar failure modes.
Since this is an older unit, all of the large electrolytic capacitors are also suspect and need to be replaced. There are three power supply capacitors at the bottom of the transmitter, two 76,000 μF 40 VDC for low voltage and one 7500 μF, 350 VDC for high voltage. The modulator section also has six 5100 μF 350 VDC capacitors, collectively known as “dynamite sticks” due to their explosive potential if installed incorrectly.
Harris SX 5 modulator section
While replacing the dynamite sticks, I noticed this PDM pickup board has a whole burned through it. This is a part of the modulator section and if it burned completely open, would likely cause all sorts of problems with this transmitter, likely spurs all around the dial, distorted modulation or perhaps overload and fail altogether.
Harris SX 5 transmitter damaged PDM pull-up board
I managed to fix it with a jumper between what is left of the circuit board trace and the capacitor mounting bracket. I soldered the jumper to the board face and soldered the wire lug. After scraping all the oxidized metal off of the capacitor mounting bracket, I attached with a screw. The board itself needs to be replaced, if it is still available from Harris, which it may not be as support for this transmitter was dropped in 2008.
Harris SX5 PDM pull up board temporarily repaired with wire jumper
The worst, and I mean worst possible situation with these transmitters is some type of control malfunction. The control boards and oscillator are in that large vertical pull-out drawer. God protect and preserve the digital control and S and M boards, as they are a major headache to troubleshoot. They have 7300 TLL (5-volt logic) that controls all functions and only a little problem will cause the entire transmitter to shut down.
Other SX series transmitter tips can be found here.
I didn’t get to replacing the blown devices because of a looming electrical storm, which precludes working inside of transmitters. I’ll get back there next week and finish the job.
Update: I finished the repair job today 8/24. There were 16 blown MOSFETS on the PA boards. I checked the drive levels on the input side of the RF torrid load resistors and it is with normal range. I also found this snake in the bottom of the transmitter across the HV shorting bar.
Small grey rat snake, electrocuted by Harris SX 5 transmitter
Could have been chasing diner. Overall, the site needs help. The air conditioner is coming next week.
Harris SX 5 transmitter fully operational
After repairs, the transmitter is back at full power and modulating +125% again.
