Medium Frequency ATU design

This is a topic I have covered before, but it is worth doing it again for future reference.  The previous post covered downgrading an AM transmission facility for WGHQ, Kingston, NY.

This is part II of that process.

WGHQ transmitter site, towers 1 and 2 removed
WGHQ transmitter site, towers 1 and 2 removed

The old towers have been cut up and put in a scrap metal dumpster. They are off to China to be melted down and made into a submarine or a missile or a tank or something useful like that.

Towers scrapped
Towers scrapped

The directional array had three towers in a straight line with a common point impedance of 60 Ohms.  Dropping two towers greatly changed the electrical characteristics of the remaining tower, therefore the existing ATU needed a bit of reworking to match the 50 Ohm transmitter output.

The first step, correct a few deficiencies left over from the old array.

Vise grip tower feed
Vise grip tower feed

This vise-grip RF connection has to go. The problem is where the tower erectors attempted to solder the copper tubing.  That tower base plate is pretty big and I would wager they didn’t use enough heat to make the solder connection.  They were probably working in the winter time, thus the “temporary” fix.  This tower was put up in 1993, so that temporary fix lasted 23 years.

I removed the offending tool and soldered the connection to another part of the tower with silver solder.  The smaller crossbar made a good connection point.

RF feed correctly connected to the tower
RF feed correctly connected to the tower

After soldering, I cleaned up and sprayed some grey primer on it to prevent rust from forming where I scraped the paint off.

Next, I made an impedance measurement:

WGHQ tower base impedance measurement
WGHQ 920 KHz tower base impedance measurement

That junk on the upper part of the graph is coming from WHVW on 950 KHz. The tower itself looks pretty good, 77.6 Ohms resistance with 130 Ohms inductive reactance.  Since this is not part of a directional antenna system, the ATU design is pretty straightforward.  Given that WHVW on 950 KHz is located 10.41 miles away, a low-pass filter design is optimum.  A basic low pass filter T network has inductive input and output legs with a capacitive shunt leg to ground.

T network diagram
T network diagram

Each leg is used to match the 50 Ohm transmission line impedance (R1) to the 77.6 Ohm tower impedance (R2) and cancel out the 130 Ohms of inductive reactance.  This is a vector impedance problem, much like a vector force problem in physics.   Some basic arithmetic is required (always include the units):

X1, X2, X3 = √(Zin x Zout)

X1, X2, X3 = √(50Ω x 77.6Ω) or X = 62.28Ω

The value of inductance or capacitance for each leg is calculated using the basic inductance or capacitance formulas:

L (μH) = XL / 2πf(MHz)

And

C (μF) = 1 / 2πf(MHz) XC

Thus the input leg, or X1 = 62.28Ω / (6.28 x 0.92 MHz) or 10.78 μH

The Shunt leg, or X2 = 1 / (6.28 x 0.92 MHz x 62.28Ω) or .0028 μF

The output leg is a little different.  The tower has 130 Ohms of inductive reactance that needs to be canceled out with a capacitor.  Rather than cancel out all of the inductive reactance, then add an inductive output leg, the tower reactance can be used as part of the tuning circuit.  The design calls for 62.28 Ohms inductive reactance, so 130Ω – 62.28Ω = 67.27Ω, which is the value needed to be canceled by a capacitor:

Output leg, or X3 = 1 / (6.28 x 0.92 MHz x 67.27Ω) or .0025 μF

A little Ohm’s law is used to calculate the base current for both the day and night time operations.

Ohm's law pie chart calculator
Ohm’s law pie chart calculator

Thus the daytime base current is I = √(P/R) or I = √(1000 W/77.6Ω) or 3.58 Amps.

Night time base current is I = √(38 W/77.6Ω) or 0.70 Amps

Current handling requirements:

Base current is calculated to be 3.6 Amps at 1,000 Watts carrier power.  Allowing for 125% peak positive modulation makes it 5.7 Amps.  Having a safety factor of two or 11.4 Amps output leg and 14 Amps input leg.

Voltages: 353 maximum input voltage, 439 output.

Thus, 20 amp, 10 KV parts should work well.

The designed schematic for the ATU:

WGHQ ATU Schematic diagram
WGHQ ATU Schematic diagram

Putting it all together.

Since the tower looks fairly broad at 920 KHz, we are going to attempt a nice broadband ATU to match it.  This station is currently programmed with a classic country format, and I have to tell you; those old Conway Twitty, Merle Haggard, Patsy Cline, et al., songs sound pretty good on the old AM radio.  The Subaru stock radio has HD, which also has a nice broad IF section, thus allowing all those lovely mid-high-range frequencies through.

This is the existing ATU, which I believe was built by Collins in 1960:

WGHQ Tower 3 ATU
Existing WGHQ T network ATU

The ATU building is a little rough, but the ATU itself is in remarkable shape for being 56 years old.  The input leg inductor is in the center and will be reused as is. The large Jennings vacuum capacitor at the bottom is a part of the shut leg.  Its value is 2000 pF at 15 KV.  The top vacuum capacitor is a series output cap, its value is 1000 pF at 15 KV.  The basic plan is to move the upper cap down in parallel with the bottom cap.  The shut leg inductor will be kept in place to tune out any access capacity.  For the output leg, I have a 2500 pF mica cap and a 10-100 pF variable cap connected in parallel.  The inductor on the output leg will be removed.

After some re-work on the ATU components, I tuned everything up.  The easiest way to do this is to disconnect the legs, measure them individually, and adjust them for the desired reactance, which in this case is 62.28 ohms or thereabouts.  The output leg was measured with the tower connected since the tower reactance is a part of the tuning circuit.  The input leg was right about 10 μH.  The shunt leg turned out to be about 0.002 μF.  This is often the case, theoretical values are slightly different than field values due to stray capacitance and inductance in the connecting straps, etc.

This is the load, as measured at the output terminals on the transmitter:

WGHQ tower load as measured at the transmitter output terminals
WGHQ tower load as measured at the transmitter output terminals

Slightly asymmetric on 910 KHz, but overall pretty good. There is a fair amount of phase rotation in the transmission line due to the length from transmitter to the tower (855 feet, 260.6 meter), which works out to be 0.93 wavelength allowing for the 86% velocity factor of the transmission line.

Time to pack up and go home.

Downgrading an AM radio station

WGHQ in Kingston, NY has been downgraded from a 5KW DA-1 to a 1KW non-DA system.  This was done because two of the three towers in the directional antenna array dated from 1960, were in very rough condition and needed to be replaced.  The remaining tower (furthest from the transmitter building) had been replaced in 1994, is in good condition, and is being kept as the non-directional radiator.

Here are a few pictures:

WGHQ 3 tower directional antenna array, Port Ewen, NY
WGHQ 3 tower directional antenna array, Port Ewen, NY
More deferred maintenance
More deferred maintenance
RF and tower light feed disconnected from tower base
RF and tower light feed disconnected from tower base
Second tower base vegetation not as bad, tower disconnected
Second tower base vegetation not as bad, tower disconnected
WGHQ transmitter and original Collins phasing cabinet
WGHQ transmitter and original Collins phasing cabinet

First tower video (sorry, I appear to have no idea what I am doing with the camera):

Second tower video, this one is better:

Towers on the ground:

I made measurements on the third tower and constructed a temporary ATU with parts on hand to get the station back on the air. They are now running 1 KW day, 38 watts night, as per their CP. I will be going back up to finish the job once the brush has been removed from around the existing tower and the ATU building has been repaired.  The coverage with 1 KW is not bad, actually:

Predicted coverage map from FCC website
Predicted coverage map from FCC website

The translator is on the way.

Whatever happened to the CFA?

Remember way back when, perhaps in high school or college, you met this really cool person who seemed to be wonderful in every way? Yeah, then you got to know them a little better, and, well, those first impressions changed a little bit.

Crossed Field Antenna, Courtesy of Wikipedia
Crossed Field Antenna, Courtesy of Wikipedia

The Crossed Field Antenna (CFA) sort of reminds me of my first prom date.  There was a lot of promise there, but plans fell through.

From a 1999 Radio World article:

Imagine an AM antenna one–fiftieth of a wavelength long, that needs no radial ground system, occupies a small parcel of land, produces little or no RFI (Radio Frequency Interference), has great bandwidth and performs better than a full–sized vertical radiator.

This potential new antenna was all the rage during the early 00s or whatever you call that decade.  I remember thinking to myself; I will believe it when I see the test results.  At one point, there was a battery of tests run in the installation in Egypt and China.  The test results are spotty at best, however, none of these installations performed up to expectations.  While it looks like a cool idea, and it would have been great to see it succeed, it seems that sheer willpower alone will not make a particular system work outside of the laws of physics.  There are a few of these still in operation out in the wild, mostly in Egypt.

Lightning protection for WLAN links

More and more wireless LAN links are being installed between the transmitter and studio.  Often these links are used for network extension, remote control, site security, VOIP telephony, and sometimes even as a main STL.  These systems come in several flavors:

  • Moseley LAN link or similar system.  Operates on unlicensed 920 MHz (902-928 MHz) band.  Advantages: can use existing 900 MHz STL antennas, can work reliably over longer distances, transmitter, and receiver located indoors.  Disadvantages: slow, expensive
  • ADTRAN TRACER or similar system with indoor transceivers and coax-fed antenna systems.  Operates on unlicensed or licensed WLAN frequencies.  Advantages: fast, transmitter and receiver located indoors, can be configured for Ethernet or T-1/E-1 ports.  Disadvantages; expensive
  • Ubiquiti Nano bridge or similar system where the transceiver is located in the antenna, the system is connected via category 5/6 cable with POE.  Operates on unlicensed or licensed WLAN frequencies.  Advantages; fast, relatively inexpensive.  Disadvantages; equipment located on the tower, difficult to transition base insulator of series fed AM tower.
  • Ubiquiti Rocket or similar system where the antenna and transceiver are separate, but the transceiver is often located on the tower behind the antenna and fed with category 5/6 cable with POE.  Operates on unlicensed and licensed WLAN frequencies.

For the first two categories of WLAN equipment, standard lightning protection measures are usually adequate:

  • Good common point ground techniques
  • Ground the coaxial cable shield at the tower base and at the entrance to the building
  • Appropriate coaxial-type transmission line surge suppressors
  • Ferrite toroids on ethernet and power connections

For the second two types of WLAN equipment, special attention is needed with the ethernet cable that goes between the tower and POE injector or switch.  Shielded, UV-resistant cable is a requirement.  On an AM tower, the shielded cable must also be run inside a metal conduit.  Due to the skin effect, the metal conduit will keep most of the RF away from the ethernet cable.  Crossing a base insulator of a series excited tower presents a special problem.

The best way to get across the base insulator of a series excited tower is to use fiber.  This precludes the use of POE which means that AC power will be needed up on the tower to power the radio and fiber converter.  This may not be a huge problem if the tower is lit and the incandescent lighting system can be upgraded to LEDs.  A small NEMA 4 enclosure can house the fiber converter and POE injector to run the WLAN radio.  Some shorter AM towers are no longer lit.

Another possible method would be to fabricate an RF choke out of copper tubing.  This is the same idea as a tower lighting choke or a sample system that uses tower-mounted loops.  I would not recommend this for power levels over 10 KW or on towers that are over 160 electrical degrees tall.  Basically, some 3/8 or 1/2-inch copper tubing can be wound into a coil through which a shielded ethernet cable can be run.  Twenty to twenty-five turns, 12 inches in diameter will work for the upper part of the band.  For the lower part, the coil diameter should be 24 inches.

In all cases where CAT 5 or 6 cable is used on a tower, it must be shielded and the properly shielded connectors must be used.  In addition, whatever is injecting power into the cable, ether POE injector or POE switch must be very well grounded.  The connector on the shielded Cat5 or 6 cable must be properly applied to ensure the shield is grounded. 

In addition to that, some type of surge suppressor at the base of the tower is also needed. Tramstector makes several products to protect low voltage data circuits.

Transtector APLU 1101 series dataline protector
Transtector APLU 1101 series data line protector

These units are very well made and designed to mount to a tower leg. They come with clamps and ground conductor designed to bolt to a standard copper ground buss bar.

Transtector APLU 1101 series dataline protector
Transtector APLU 1101 series data line protector

There are various models designed to pass POE or even 90 VDC ring voltage.

Transtector APLU 1101 series dataline protector
Transtector APLU 1101 series data line protector

This model is for POE. The circuit seems to consist mostly of TVS diodes clamping the various data conductors.

As more and more of these systems are installed and become a part of critical infrastructure, more thought needs to be given to lightning protection, redundancy and disaster recovery in the event of equipment failure.