Engineers are funny. We all have our likes and dislikes and our reasons for both. I don’t really like Harris products. Even when I was in the military, their stuff seemed a little “light.” I suppose having to deal with an MW-50B transmitter at my first full-time chief engineer gig didn’t help that impression. The MW-50 would “blow up” every six months or so. I say blow up because that is the only way I can describe it, no overload lights or any other indication of trouble until the blue lightning flashes and thunder from the PA section. What a POS.
Other Harris transmitters, such as the SUX-1, FM20H, Gates-1 etc have also left me less than impressed.
In order of preference, my choice of AM transmitters would be:
Any Nautel solid-state unit. Nautel makes good equipment that is well-supported.
Any BE solid-state transmitters. I favor the A model over the E model, but both are good. One condition, they must absolutely be well grounded and all of the toroid filters provided by the manufacturer must be used when installing.
Any tube-type Continental transmitter. There are older units, I believe 816R but they work well and sound good on the air.
We have a Harris Z5-CD transmitter for one of our FM stations. Brand H is not my preferred make, however, it was already installed when we bought the station, so I have to live with it.
This particular site gets hit by lightning strikes often. Normally, it does not affect anything until the transmitter gets turned off for maintenance. Then, almost invariably, when turning the transmitter back on one of the modules will fail. Most often this is manifest when one of the two power supplies shut down causing the transmitter to run no more than 20% power.
The way this is troubleshot is to slide each module out and turn the transmitter back on. When the power supply stays on, the bad module has been located. A confirmation test is to check the MOSFET for a short circuit between Drain and Source. This short circuit condition puts a direct short on the power supply causing it to crowbar and turn off.
So, once the bad module has been located, and the spare module is installed in the transmitter, then what? Most engineers call Harris and ship the module back for repair. Most engineers don’t want to mess with unsoldering a surface mount MOSFET and soldering a new one in. I find it moderately entertaining to fix things myself, so I do not do what most engineers do.
NXP BLF177 MOSFETS
The MOSFET in this particular module is the BLF177, made by NXP. Harris will sell you one for quite a bit of money. You can also buy one from Mouser for about half the cost.
Harris FM Z series transmitter PA module with cover removed
Once the parts are obtained, the worst part of the entire job is unsoldering the old MOSFET. This takes some patience and skill. What I found works best is to melt some solder on the foil leads and get them good and hot. Since this MOSFET is already destroyed, we don’t have to worry about heat, etc. The one thing you do not want to do it actually break the MOSFET open. That is because it contains beryllium oxide, a known carcinogen. Once all the solder is liquid, carefully pry the foil up with a small screwdriver. There are several components that have to be moved to work on this.
Harris Z series PA module with MOSFETS removed
After the old MOSFET is removed, clean up the solder pad with a solder pump and solder wick. I like to use a little liquid flux on the solder wick, it makes things go faster.
Once all the old solder is cleaned off the solder pads, I brush a light coat of liquid flux in the pad. Again, this makes things go faster.
Harris Z series FM transmitter module new MOSFETs waiting to be soldered
The new MOSFETS are very sensitive to static discharge, so I always use a static drain wristband when handling them. I place both MOSFETs onto the circuit board. I then solder them on using as little heat as possible from the soldering iron. Again, the MOSFETs are sensitive to heat and one can easily be destroyed if it gets too hot.
Harris Z FM series PA module repaired
This is the module with the new MOSFETs soldered in. I use defluxing compound to remove all the extra flux. Once it cools off, I test the new module with a DVM:
Harris Z series FM PA circuit board under test, resistance is 3.3 Mohm
If the MOSFETS are good, they will have an internal resistance of around 3.3 MΩ. If the module is bad the MOSFETS will read only a few ohms if shorted:
Harris Z series FM PA module under test, DVM reads 1.6 ohms
That is how you do it. I think Harris charges $775.00 per module to repair. I fixed this one for $240.00, but that is not the reason I did it. I did it for the fun that was in it.
Once upon a time, in the not-too-distant past, all long-distance communication in the US was handled by one company, AT&T. There was no other company that could transmit data over medium to long distances. The breadth and scope of their communications network are not understood by most people these days. Most people know that AT&T handled long-distance telephone calls for the Bell Telephone System until the Bell breakup in 1984. However, AT&T did a lot more than long-distance phone.
For example, if you watched the network news or network TV show anytime before 1980, it was likely brought to you via AT&T microwave system, known as AT&T long lines. Listen to the news on the radio, same deal. Before the widespread use of communication satellites and fiber optics, the AT&T microwave relay network was the only way to get various types of electronic media signals from one place to another.
Beginning in the late 1980s, competing local and long-distance telephone companies began installing fiber optic cables between company offices. That coupled with the increased use of satellite systems for mass media video and audio delivery services made the huge AT&T microwave network obsolete. Some of the old microwave sites that are located in downtown areas have been reused by local phone companies and cell phone providers. Many of the rural sites now sit empty.
ATT long lines microwave site with towers
This is the former AT&T microwave relay site located near Kingston, NY. It is now owned by American Tower, Inc. There are two towers behind the building, only the tower on the right has a few active communications antennas on it. The taller tower is 190 feet tall and was built in 1957. The shorter tower is 120 feet tall and was built in 1961. Both towers and everything on them were made by Western Electric, the same company that manufactured the telephone sets. Chances are, Western Electric contracted the actual manufacture of equipment out to others, then billed AT&T, their parent company a markup. Something that would make all MBAs proud.
Western electric 190-foot tower, built in 1957
This tower was built in 1957. The structure and galvanizing are still in excellent condition.
The large antennas you see on the towers are microwave horn antennas. They are no longer in use. Several transmitters and receivers would have been connected to each one of these antennas by use of RF multiplexers. Each microwave transmitter/receiver would have had several data channels. Generally, this was C Band microwave equipment, so it was in the 4, 6, and 8 GHz frequency range.
Western Electric KS-15676 microwave antenna
All of this telephone traffic was transmitted on digital data channels unencrypted. Many have argued that this allowed the government (most notably the NSA or National Security Agency) to intercept and listen to most domestic long-distance telephone calls within the US. There is a book called Puzzle palaceby James Bamford if you are interested in NSA history. It was written more than 20 years ago, so it doesn’t really apply today, but it is an interesting look at what the government was up to.
The building itself is huge, the first floor is 16,000+ square feet and the second floor is 10,000+ square feet. Only about 1000 square feet of this space is actively being used.
I believe this building was built in the late 1940s or early 1950s, just as Kingston was growing into a major IBM manufacturing site. It has remnants of the ATT coaxial-based system that was used prior to microwaves. The IBM buildings are located a few miles to the southeast of this location, they are another cold war relic for discussion later. The IBM buildings were a major computer research and development site in the 1950s until it closed in 1992. It was assumed that the Soviets had several spy satellites trying to steal secrets from the area, and the IBM facility was a primary nuclear target.
Blast baffle for generator cooling air intake
The microwave relay site has 12-inch re-enforced concrete walls. The ventilation air intakes have blast baffles to prevent a pressure wave (from a nuclear explosion) from blowing the ventilation equipment off of its mounts.
pneumatic actuator panel seals all outside openings with steel blast doors
All of the outside openings were able to be sealed with steal blast deflectors using a pneumatic control panel located in the control room. There was a five-minute timer, presumably to allow the HVAC units to be secured before the doors were closed. They were heavy gauge steel shutters designed to deflect the pressure wave of a nuclear explosion. Since this is an earlier building, it is likely that it is built to a 2 PSI pressure wave spec. Newer buildings were built to 20 or even 50 PSI. This microwave relay site would not have withstood a direct hit from a nuclear warhead, especially the higher-yield warheads that came later on.
Water chillers for HVAC system
There were three large water chillers to provide cooling to the HVAC units. Since this was the 1950’s all of the electronic equipment would have had tubes, which would have generated a lot of heat while operating. There were two loops in the HVAC system. The refrigerant loop, which ran between these units and the huge condensers on the second-floor roof, and the chilled water loop which ran between these units and the air handlers located in various parts of the building.
There is a bomb shelter in the basement. I found a couple of olive drab cans of civil defense water laying around. The lights were not working at the bottom of the stairs, so I chose not to go into the bomb shelter itself.
Stairs going down to the bomb shelter
“Okay everybody, the missiles are on their way, so let’s head down these stairs and pray”
There were two diesel generators, one was 325 KW which could run the entire building. The other was a 200 KW which could run the critical building functions. The fuel storage consisted of two 10,000-gallon tanks buried in the ground outside. Each steel fuel tank had a cathodic protection circuit. Basically, a small negative electrical current was passed to the steel tank to keep it from rusting. Apparently, it worked because when the tanks were removed in 2000 after 45 years in the ground, the primer was still on the outside of the tank.
Electrical switch gear, part of power company sub-station
The building has its own power substation. The electricity from the utility company comes off the pole at 13,800 volts and goes to a large step-down transformer on a pad outside. From there 480 volts is fed to this switch panel, where it is routed to motors loads or other step-down transformers within the building.
Frame room floor, equipment removed
On the main floor, there were rows and rows of wire terminal equipment, microwave transmitters, receivers, and data and RF multiplexers in racks. The room in the above picture is about 10,000 square feet, there is another 6,000 square feet beyond the plastic heat barrier. This microwave gear received and transmitted data from Albany and Germantown to the north; Poughkeepsie, Putnam Valley, Ellenville, and Spring Valley to the south. All of that equipment is gone now, replaced by empty space.
Now the whole place is a little creepy.
There are about 500 copper wire pairs of telephone cables that came into various parts of the building to carry the DS-1 and DS-3 circuits that interfaced with the TELCO office in Kingston.
All in all, this was a serious building, no expense was spared in the construction and equipment outfitting. The entire building is shielded with copper mesh screens embedded in the concrete walls. There were redundant systems on top of redundant systems, something that you do not see these days, even in government buildings such as emergency operation centers (EOCs) and 911 call centers.
Most (if not all) radio engineers cringe when they hear a clap of thunder. Then the waiting begins. What are we waiting for? The cellphone to start ringing, of course. Over the twenty or so years I have been doing this, I have learned a few things. One of them is you cannot overground something.
That being said, you can, of course, ground something improperly.
The worst areas we have for lightning damage are the Gainesville/Pensacola markets. Those places are in the lightning capital of the US. Time was our class C FM station was getting knocked off a couple of times a month.
US Thunderstorm Days map
There is hope. When we upgraded the stations and installed new transmitters in 2004 I insisted that the tower and building be properly grounded. I even got into an argument with the CFO about the “mission creep” as he put it. Never mind that I put $20K in the initial work specification for grounding.
There are a couple of strategies to use when dealing with lightning at transmitter sites:
Grounding: First, foremost, and always. Grounding should consist of multiple ground rods driven as deeply into the earth as possible. At the Trenton Florida transmitter site, we used 20-foot-long ground rods driven in 20 feet apart all the way around the building and in five 60-foot spokes around the tower. All of these ground rods and tower bases were bonded with #2 solid copper wire CAD (exothermically) welded to the ground rods. All turns were kept to a large diameter radius to keep inductance down. When lightning strikes the tower, this creates a large electron sink to dissipate the strike energy into.
Bonding: All equipment cabinets, racks, and everything metal is bonded together and to the same ground point presented by the grounding system. When lightning strikes, often the ground cannot dissipate the energy fast enough. When this happens, the entire ground area around the tower gets charged up. Current will only flow down a less resistive path. If everything is bonded together, the potential between any piece of equipment or component is the same, even if that potential is +10,000 volts. No flow of current means no damage.
The transmitter building is located away from the tower. At almost every FM and TV transmitter site I have visited, the building is right smack at the tower base. By moving the building away about 100 feet or so, the EMP from the tower strike has dissipated (log function) significantly before it passes through the transmitter building. It is a little more expensive to install due to the added transmission line lengths and losses, however, it works.
I have been at the Trenton Florida transmitter site when lightning struck the tower. The result, not even a transmitter overload. Nothing was noticed on the air, no damage was sustained by any equipment. For the last five years, there has been no off-air time due to lightning damage at this site.
Studio building with lightning rod, Gainesville, Florida
The studio site has a similar story. We built a new studio building in 2005, there is a 100-foot monopole that holds the STL antennas. You know that it gets hit during a storm. I remember the manager and IT guy from Pensacola commenting about how nice the new SAS Rubicon consoles were. Both of them also said that they wouldn’t last through the first summer because of lightning damage. Four years later, not a single incident of damage to the consoles, computers, or anything else in the building because we grounded everything as I described above.
