The FAA used the AM-6155 in the 1970's and 1980's with ground-to-air transmitters. They were designed to cover 225-400 MHz at 50 watts AM continuous output. With only a few hours of work, and a few extra parts, they are capable of 400 watts RF output on 144 MHz, 222 MHz or 432 MHz in SSB/CW at Amateur Radio duty cycles. This amplifier used either an Eimac 8930 or an Amperex DX-393 tetrode tube. There are tons of these amplifiers currently in circulation mainly in North America.

ORIGINAL DOCUMENTATION

Web links to the original ITT (the manufacturer) AM-6155/6154 manual and circuit schematics can be found below:

• Schematics
• Manual

MAIN CHASSIS CONVERSION

Basic Chassis conversion

The main chassis (power supply) conversion is described in numerous web locations. Since there were several iterations of the conversion developed over the years, reading the various websites becomes confusing and even contradictory in some cases. Luckily, I found a web link by W3RJW that simplifies everything. As a main reference, use his chassis conversion webpage. But apply the following modifications to his instructions provided in the link:

• In step II-A-2, use 2 x 20 ohm 2W resistors instead ox 2 x 10 ohm. This provides a x10 meter reading factor on the plate current instead of x20. In other words, a reading of 34 would indicate a plate current of 340 mA, much easier to read than a reading of 17.
• Do not perform steps IV-C and IV-D. Instead, use the re-designed grid bias circuit described below. With the new grid bias circuit, these steps should not be performed.

Improved Grid bias regulation

The first tests I performed with my AM-6155 on 432 MHz revealed a major non-linear output response. The output RF power was not tracking the input RF power. The unit picked up gain as the input level increased. I noticed a major sag of the grid bias voltage when keying down at full power. This voltage would go from -78V to -50V in a matter of a couple of seconds. This is due to inadequate grid bias regulation. Apparently, this is worse at 432 MHz due to decreased tube efficiency. If you intend to operate this amplifier in a linear mode (AM, SSB), you MUST perform a grid bias modification. I still definitely recommend that you perform the mod in all cases and for all amateur bands.

A glimpse at the original circuit schematics revealed a resistive divider and a potentiometer as the way to set the bias. Looking around on the web, I found K4HV's bias circuit proposal. That circuit improves the regulation somewhat, but requires numerous resistor values and a 2W pot which is rather expensive to acquire. So I decided to design my own. With my improved grid bias supply, the voltage is now rock-solid and RF response is linear!

The mod consists of disconnecting the existing grid bias circuit and connecting a new regulator circuit built on a small piece of prototype PCB. The new grid bias regulator is based on an LM337T negative linear voltage regulator. The circuit schematic is shown on the right. This circuit allows to adjust the grid bias voltage to between -69V and -95V DC. This is appropriate for setting the idle plate current on both the Eimac 8930 and the Amperex DX-393 tetrode tubes. For the 4CX400A, a different bias voltage range is required. See the section below. This grid bias circuit is quite straightforward to put together. All components can be found at major component suppliers such as Digikey, Arrow and Mouser. Use a small heatsink on the LM337T regulator. A simple, common 1/2W trimming potentiometer (single turn or multi-turn) is used. I have decided not to locate the pot on the rear panel. I find the adjustment pretty stable for a "set once and forget" approach. Besides, long leads on the potentiometer might have promoted instability on the voltage regulator. Of course, this requires that you make the idle plate current adjustment with the amplifier top cover removed and the interlock switch depressed or bypassed. Be careful! Lethal voltages everywhere in there... For mounting the new PCB, I used metal tabs covered with plastic tubing, those commonly used in consumer electronics to keep the wiring in nice bundles. I attached the tab eyelets to the PCB corners using screws and nuts. I then wrapped the tabs around one of the A3 assembly horizontal standoffs as shown on the pictures on the right. Use your imagination and mount the PCB on the A3 assembly your way! 2023 UPDATE I have decided to improve the grid bias circuit for two reasons. First and foremost, I do not like bringing out some high voltage to the rear panel. Secondly the 4CX400A version of the circuit produces some residual plate current (5 -10 mA) when in receive. This is because the most negative grid voltage produced is -67 V, which is not quite sufficient to completely cut off the tube conduction. The new circuit adds complete isolation of the TX-RX switching control by means of a PhotoMOS solid state relay (SSR), a six pin DIP integrated circuit. Any of the PhotoMOS devices will work here as long as it can sustain 200 VDC on its output.  An AQV210, AQV214 or AQV216 is suitable. Applying any voltage between +6 V and +24V to the Tx Keying signal will switch the amplifier grid bias into Tx mode. On the updated circuit, the voltage regulator circuit is always powered up, which improves stability and reliability. When in receive mode, the regulator output is disconnected from the the grid bias line. However a full -100 V of bias is applied through R3, a weak pull down resistor. In transmit mode, the PhotoMOS SSR connects the regulator to the grid bias line.

Click on the images to enlarge them. 2023 version circuit 2008 version circuit

Once the PCB is assembled, follow these wiring steps to install the new grid bias circuit:

1- On the existing A3 assembly printed circuit board, cut one of the two leads of the following resistors and lift up their end so that the cut off lead no longer makes contact:

• R14, 1.5K, 2W
  
• R16, 22K, 2W

2- Optionally, on the existing A3 assembly printed circuit board, replace VR1, the 100V zener diode, with a new 100V 5W one such as a 1N5378B. Since this is an old amplifier and we are revising it, I figured it would not hurt to replace that one. Up to you.

3- Solder the new grid bias circuit input wire to the TP4 post on the A3 assembly printed circuit board.

4- Solder the new grid bias circuit output wire to the E10 post on the A3 assembly printed circuit board. Do not disconnect the existing orange wire form the E10 post.

5- Run the T/R Keying wire to the back panel as suggested in Step IV-D of W3RJW's chassis conversion webpage. Applying any voltage between +6 V and +24V to the Tx Keying signal will switch the amplifier grid bias into Tx mode, which allows the idle plate current to flow.

The grid bias adjustment is made with the amplifier powered up and keyed in Tx mode (no RF at the input). Adjust the trimming potentiometer for an idle plate current of between 60 and 90 mA. It should normally represent ~ 20% of the maximum plate current. I set mine to 80mA.

That's it. Solid output and good linearity.

Chassis Re-Wiring to 240V AC

I suggest re-assigning the amplifier so that it runs on an AC mains input to 240V. Considering that we run the amplifier beyond its specified limits, this will give some relief to the transformer primary winding. The switch to 240V is done by re-positioning jumpers on the two TB1 terminal blocks. There is one terminal block on top of the A3 Assembly (hidden under a removable "Caution High Voltage" plate) and one block next to the power transformer under the long high voltage hinged cover. Jumper assignment for 240V is the same for both blocks:

RF PLUGIN CONVERSION

For the RF plugin conversion, make sure you use the circuit schematic that corresponds to the right type of plugin you own. There exists a VHF (AM-6154) plugin and a UHF (AM-6155) plugin out there. Select the proper schematic page, otherwise it will be confusing when doing the grid compartment component removal process.

RF Input Compartment Conversion

For Input (Grid) circuit conversion, I basically followed W3RJW's modification page, steps 1 to 10, with the following adaptations:

In step 6, I failed in trying to remove the input pedestal inductor shaft from the input cavity. The set screws were frozen hard. Instead, I unscrewed the pedestal from inside the grid compartment. I then cut off the threaded shaft as far as I could using a Dremmel router tool with a cutting disk. I then covered the input cavity entry inside the grid side compartment with a piece of adhesive aluminum tape, the type used to seal ventilation ducting. In step 7, I used a Dremmel tool to cut off half of the Input tuning capacitor. This gave a neat result with no bent or damaged plates. Use a bench vise to keep the capacitor firmly in place while cutting.

Click on the figure to enlarge it

• I used 240pF silver-mica capacitors instead of the suggested 200pf values. This is what I had on hand and it worked fine.

• I adjusted the position of RF input injection on the brass grid line to get a good input VSWR. The sweet spot where to connect the input capacitor ended up much closer to the end of the grid line than the 5/8" suggested. I now get a VSWR of less than 1.5:1.

RF Output Compartment Conversion

I did not perform steps 11 to 20 of W3RJW's modification page. These are the more difficult to accomplish "HV Teflon Cocoon" mods. Instead, I performed the simple plate HV injection approach by K4HV. As mentioned by K4HV in his Loading Capacitor Modifications section, do not forget to take the black plastic washer out from the plate coupling gearbox. This allows to increase the travel of the plate disk away from the plate ring. Otherwise plate coupling will be excessive and you will not get the full output power. See the figure to the right.

Click on the figure to enlarge it

Output Modules Removal

Remember to disconnect and remove the low pass filter module and the directional coupler module from the output coaxial line. Replace them with N-Female/N-Female adaptors.

4CX400A TETRODE SUBSTITUTION (OPTIONAL)

I have also implemented and tested an improved grid bias circuit to support the Svetlana 4CX400A (GS-36B) tube requirements. Since the original Eimac and Amperex tubes are becoming scarce, this tube is a good substitute. I purchased a pair of New Old Stock (NOS) GS-36B (4CX400A) on eBay for around $100 per tube, delivered. My initial thoughts were that I would get additional output power over the Eimac 8930 just by comparing the specs.

Before we go any further, let me warn you that you MUST CONDITION any new transmitter tube before applying high plate voltage to it, otherwise the tube will arc internally and this will damage the tube and your amplifier. I learned this the hard way! The conditioning procedure consists of running the tube with the amplifier powered on (filament on, fan running, no plate voltage, no RF at the input) for a few days (I suggest 4 days). This allows the getter to capture any remaining gas molecules inside the tube. Feel free to search the web for more information on power tube conditioning and the getter electrode.

Anode Ring Modification The anode heat dissipator on the 4CX400A is slightly smaller than on the original tubes. To accomodate this smaller diameter, you must bend the amplifier anode ring fingerstock towards the center by about 0.5mm, one finger at a time. In order to accomplish this right, I suggest you take the anode ring out of the cavity. Only two screws need to be removed to do this. Improved Grid bias regulation For the 4CX400A, the grid bias voltage is higher (less negative) than for the original tubes. The circuit shown to the right does the trick for the '400A tube. Follow the procedure described above to modify the AM-6155 chassis to the improved grid bias circuit, but instead, build the improved grid bias circuit shown to the right. 2023 UPDATE As mentioned above, I have decided to improve the grid bias circuit, and this benefits the 4CX400A even more. With this improved circuit, -100 V DC gets applied to the grid when in receive mode. This completely cut off the tube conduction. With the previously proposed circuit, some residual plate current (5 -10 mA) will flow when in receive. Although not serious, this is not ideal.

Click on any image to enlarge it. 2023 version circuit 2008 version circuit

Chimney and Cover Re-Assembly

One thing I noticed while re-installing the chimney and cavity cover is that the chimney protrudes outside of the cavity by a millimeter or so. Once installed, the 4CX400A tube stands probably slightly taller than the original tube. This causes the cover to not quite sit on the cavity flange. When tightening up the four cover screws, go gradually on the tightening, working diagonally and trying to maintain equal tension on the four screws.  Do not overtighten, as this force gets transferred to the tube itself. Once complete, there will still be a small gap between the cavity and its cover, but certainly not enough to disturb cooling. If RF leakage is a concern to you, you may want to use an RF gasket to fill the crack, or simply use aluninum tape all around to cover the gap. I did not care since the RF plug-in sits inside the chassis.

Results

Well, the bottom line is to not expect a big improvement over the original tubes with the 4CX400A. From what I could see, you can get maybe 25-50W more, so 350-400W but the plate current is significantly higher too. This shows an even less efficient configuration compared to the original tubes, which is not high to start with. The power supply is the main limitation in trying to push more output. Cooling is also limited with the stock blower.

While testing, I pushed the input power a bit too far (more than 15W), and blew up the three panel-mounted zener diodes ($10 each) that set the screen bias and a few other resistors blew up as well. Simply put, the AM-6155 chassis is not designed for such stress. I suggest you settle for 300-350W output in CW.

(2023)  ALTERNATE APPROACH FOR TX-RX KEYING (OPTIONAL)

This web page and many other modification sources propose steering the grid bias for doing the TX-RX switching. While consulting the old documentation that came with the purchased AM-6155s, I came across an alternate approach for doing the TX-RX Keying: Steering the screen bias instead. It was proposed by AF1T back in 1984. See pages 2 and 3 of his notes (PDF document to the right). His original notes are provided here with his permission.

AF1T's Tx-Rx Switching Approach

In a nutshell, Dale suggested killing the screen bias instead of steering the grid bias. That voltage is +390 VDC. Being positive, it is inherently more stable to work on, as a failure of that screen voltage will most likely mean the voltage going to zero, essentially eliminating any possibility of DC plate current. In retrospect I think this would have been a better approach. Now it involves a relay which can withstand 400 VDC, which is not cheap or common. However some solid state relays can do that. Here is one available at Digikey for less than $7 US: The CPC1983Y. That solid state relay and a high value pull-down resistor on the output, like described on page 3 of AF1T's document, would do the trick. A +12 to +14V signal applied to the SSR's input through a limiting resistor would do the witching, while providing total isolation, a plus compared to the previous grid bias solution.

This approach does not eliminate the need for the improved grid bias circuit described above, but that grid bias circuit would be always on, permanently tied to the grid bias line, and no TX-Rx keying would be needed on the grid bias circuit.

I would definitely like to hear back from those who try this approach!

OPERATION AND PERFORMANCE

With the above conversions performed, I get the following performance on 432MHz:

Original Eimac or Amperex Tube
CW/SSB: 15W input yields 325W output @ 300mA of DC plate current. I can push to 375W output for short bursts.
JT-65: 250W output @ 250mA of DC plate current is the maximum I can run in 50-second continuous transmissions without the overheat protection tripping.

Svetlana 4CX400A (GS-36B) Tube
CW/SSB: 15W input yields 350W output @ 350mA of DC plate current. I can push to 400W output for short bursts.
JT-65: Like with the original tubes, reduced output power by at least 100W in order no to trip the overheat protection.

AC Mains Current
In the 3A to 4A range @ 240VAC.