Thursday, December 10, 2009

Improving AM Performance of the Heathkit MT-1 Cheyenne Transmitter

[Note (10 July 12):  The schematic shown below has an error in it.  The 1uF cap paralleled with the 25K pot added to the circuit should connect between pin 8 (cathode) of the 6DE7 and ground, not between pin 5 and ground.  (I would fix this drawing, but I no longer have the original).  - Jeff]

[Update (2 January 2010): New information on my Cheyenne can be found here]

I picked up this transmitter at a swapmeet earlier this year. Although the front panel was in nice condition, the topside of the chassis itself had oxidized quite a bit (as had the cabinet, which was in dire need of rust removal and repainting), and it was not very appealing. Never the less, the price was right, and I thought it might be a fun project to get on the air during the cold winter months!

The Cheyenne -- case removed for repainting and
needing an original knob (hint hint) for the Drive control.
(Click on image to enlarge.)

Hmmm...under the hood, not so pretty

The Heathkit Cheyenne transmitter was a mobile AM and CW transmitter that Heathkit marketed in the late 50's and, I believe, early 60's (its matching receiver was the Heathkit Commanche (MR-1)). With a design including a single 6146 PA, 12AX7 Mic amp, and 6DE7 as a "controlled carrier" modulator, it is similar (although not identical) to the later DX-60 series transmitters.

I powered up the transmitter with an HP-20 power supply and quickly discovered that its audio in AM mode left quite a bit to be desired -- noticeable distortion and a restricted audio passband. So I wondered...what could I do to improve its performance?

Well, the first thing to do: check to see if someone else has already been down this path. Unfortunately, a google search revealed no internet articles for improving the Cheyenne. But, because of the similarities between the Cheyenne and the DX-60, I wondered if I could apply any of the DX-60 modification articles to the Cheyenne...

First Steps...Make the Cheyenne more like a DX-60...

Electric Radio magazine has several very interesting articles by Bill Breshears, WC3K, on improving the DX-60 audio. I thought they might be a good starting point for modifying my Cheyenne, but to do so, I'd first need to correct those few differences between the Cheyenne's audio/modulator stages and the DX-60's stages.

Upon comparing the Cheyenne schematics with those for the DX-60B, the significant differences in the modulator section seemed to be:
  1. The DX-60B's 6DE7 Cathode Follower has a 33K ohm resistor from its cathode (pin 9) to ground. This resistor was lacking in the Cheyenne (and indeed, its lack prevents the AC signal on the Cheyenne's cathode-follower cathode from going below about 50 VDC).
  2. The grid of the first stage of the DX-60B's 6DE7 modulator (pin 7) has a 22 Meg ohm resistor to ground, compared to the Cheyenne's 10 Meg, and this grid is driven by the previous 12AX7 stage via a 5 nF cap, instead of a 510 pF cap in the Cheyenne.
  3. The DX-60B's PA screen voltage is driven by the 6DE7 Cathode Follower through a 47K ohm resistor paralled with a 0.1 uF cap. The Cheyenne uses a 10K ohm resistor and a 0.25 uF cap.
I incorporated the changes in first two items above (although I used a 6.8 nF cap in lieu of 5 nF in step two, because that's what I had in the junk box). I left the third item for later, until I could incorporate the Carrier-Level adjustment pot that WC3K described in his articles.

One of my goals was to drive my AL-811 linear amplifier with the Cheyenne. I don't feel comfortable running this amplifier in AM mode at more than about 100 to 120 watts carrier output power. For this level of power output from the linear, I needed the Cheyenne's idle-carrier power output to be to be in the range of about 9 watts or so.

I initially incorporated the 47K ohm resistor in step 3 above (keeping the cap at 0.25 uF) and added a 25K ohm pot in series with it to the PA Screen Grid (the 0.25uF cap paralleling both) -- similar to the carrier control pot described by by WC3K in his DX-60 mods. Unfortunately, I felt I had a bit too much drop in power, and I instead replaced the 47K with the original 10K. (This would later change again. See below...)

The new pot (mounted conveniently in the rear-panel's Key Jack hole (after all, who needs a key for AM operation?) allowed me to easily adjust carrier level. But during testing I wasn't satisfied with audio performance -- there was a still a bit of "fuzziness" on the audio (pointing to distortion) that bugged me.

Examining the audio chain, it became quickly apparent that part of the problem was with the 6DE7 "controlled-carrier" modulator itself. This modulator adjusts the carrier level such that the carrier level is low for low-level signals and higher for high-level signals. To accomplish this "dynamic" carrier-level adjustment, the first stage of the 6DE7 modulator, in addition to being an AC amplifier, "clamps" the input AC signal on its positive peaks, thus causing an additional DC voltage to be impressed across the coupling cap (that couples the signal from the second 12AX7 stage to the input grid of the modulator). This DC voltage is proportional to signal level and thus drives the modulator grid (6DE7 pin 7) more negative with higher audio levels.

As this grid is biased more negative (relative to the grounded cathode), the tube conducts less, reducing the DC plate current. The plate voltage goes up, thus raising the grid voltage on the next stage (cathode-follower) and consequently, of course, its cathode voltage.

As this cathode goes up, the PA Screen Grid voltage goes up, and more carrier appears at the output.

BUT -- the key here, and the source of the distortion, is the clamping action at the grid of the first 6DE7 section. This clamping action essentially flattens the positive peaks of the audio signal at this grid, which in turn results in flattening of the "troughs" of the modulation on the output RF signal.

This flattening of the audio signal is easily observable at the 6DE7 with a scope (monitor the plate of the first stage, for example), and is quite obvious on a 1 KHz test signal. Not good. My rule of thumb is...if you can see the distortion, you can hear it.

I also had a problem in which, as I tried to adjust the Cheyenne's audio level towards 100 percent modulation, I'd get compression (i.e. distortion) on modulated RF envelope "peaks". Again, this was readily apparent by comparing the modulation on the RF signal with the audio signal driving the modulator.

Here's photo showing both of these two distortion mechanisms (exaggerated to make it clearer) . The top trace is the modulator output. You can see the clipping on the largest negative peak due to clamping by the modulator input grid. The bottom trace shows peak compression on the output RF envelope, which you can see by comparing the different levels of the positive peaks of the modulator output with the peaks of the RF envelope -- they're all the same level!

The Next Step...Improving the Cheyenne's Audio...

The fuzziness on the audio was just enough to make me want to keep working on the transmitter. One source of this distortion, as discussed above, was from the clamping action of the "controlled carrier" modulator in order to dynamically adjust carrier level.

Hmmm...suppose I eliminated this clamping action and thus the distortion that it created? Would audio be improved?

Why not? I only needed a carrier to be somewhere in the range of 8 to 12 watts to drive my linear. There's no reason why the 6146 PA in the Cheyenne cannot handle this. In other words, why not remove the "controlled-carrier" feature of the modulator (and thus the distortion that it introduces) and keep the carrier at a fixed level?

And I wanted the carrier level to be adjustable so that I could adjust the level to give me my 100 watt "sweet-spot" output from my linear.

One way to get around the "control carrier" feature is to bias the first stage of the 6DE7 modulator so that there is a fixed negative grid-to-cathode voltage that is large enough to prevent clipping on the positive peaks of the incoming audio, yet provide sufficient carrier to drive the linear. I already had a 25K pot mounted on the back panel of the chassis that I had intended to use to adjust carrier level (and had been wired in series with the original 10K power resistor to the PA screen grid -- see discussion above). I wired it instead into the cathode of the first stage of the 6DE7 so that I could adjust its operating point.

Through a process of iteration, I adjusted the PA screen grid resistor value and the position of the 25K pot so that, for the carrier output power that I wanted (8-10 watts), the cathode of the first 6DE7 was at a high enough voltage that full-modulation audio wouldn't be clamped by the grid. Thus, the PA screen grid resistor (from V6 pin 9) was changed from the original 10K ohms to 50K ohms. Although I used a robust power resistor (it was in the junk box), there's no reason why, say, a 2-watt resistor couldn't be used. And you can play around with the value of this resistor -- lower values of resistance will increase the maximum carrier power, while higher values will lower the maximum carrier power (maximum carrier power occurs when the 25K pot is set to its maximum resistance).

[Important Note: there's a trade-off when selecting the value of the PA screen resistor: for a given carrier output power, lowering the PA screen grid resistor value means that the resistance of the 25K pot in the cathode of the first 6DE7 section must also be lowered to maintain the same carrier output level. This in turn will bring the cathode voltage closer to the grid voltage (which is essentially at 0 volts), which means that it's more likely there will be audio distortion introduced at this 6DE7 grid due to grid "clamping" the positive peaks of the audio signal. I found that a 50K ohm PA screen grid resistor worked well for my application.]

I found that the value of the pot is about 8K-9K ohms for about 9 watts carrier (no modulation) RF output from the Cheyenne. Given this value of resistance, I added a 1 uF cap in parallel across the pot to bypass it for audio frequencies. (Note: This cap can be made larger, if it's desirable to run the Cheyenne at lower power (and thus a lower potentiometer resistance, which means you need a larger cap to maintain the low-frequency cutoff), but increasing its value will also increase the amount of time that it takes for this stage to reach its bias point each time PTT is pressed.)

By the way, with these mods made and the Cheyenne set for about 9 watts carrier output (no modulation), I measure the following DC voltages during Transmit:
  • V6.5: 6 volts (6DE7, first cathode)
  • V6.2: 110 volts (6DE7, second grid)
  • V6.9: 200 volts (6DE7, second cathode)
  • V4.3: 52 volts (6146, screen grid)
These changes gave me the ability to control the output carrier power from about 4 watts to about 12 watts. Note -- at low powers there may still be some peak flattening at the grid to the first 6DE7 stage (this occurs when the cathode-grid bias voltage is less than the audio peak voltage at the grid), but I've found that there's no limiting when the 25K pot is set to give me 8-12 watt carrier power output.

OK! Now that I had the distortion reduced, I next tackled the frequency response, which was a bit too restricted in the stock Cheyenne.

This was accomplished (in addition to the changes above) simply by :
  1. Changing the 0.001 uF cap feeding the grid of the first 12AX7 stage to 0.01 uF.
  2. Changing the 510 pF cap feeding the Audio Level pot (from the plate of the first 12AX7 stage) to 0.01 uF.
This gave me an audio passband with -3dB break-points at 100 Hz and 5 KHz.

The new schematic:

(Modulator Modifications -- Click on schematic to enlarge...)
[10 July 12 -- Please note that there is an error in this schematic!  The 1uF cap paralleled with the 25K pot should connect between pin 8 of the 6DE7 and ground, not pin 5.]

Other problems:

1. 6.3 VAC reading low on DVM at the terminal strip: only about 5.5 VAC (causing the relay to chatter):
  • Bypassed the fuse in the filament line with a wire soldered to the fuse-holder's terminals (this fuse is not needed, after all, the power supply is fused, and this fuse added a few additional tenths of a volt of voltage drop).
  • Cleaned the Function switch contacts.
2. The 0.02 uF cap attached to V6 pin 1 (600v bypass) "popped." Replaced.

3. AC Hum on AM signal which gets louder as mic gain is increased. The PTT signal of the Cheyenne's mic jack directly keys the Cheyenne's relay, which is powered by 6.3V AC. A mic with wired to a 4-pin plug to mate with the Heathkit mic jack shouldn't have an issue with this, because, assuming the mic and its cable have separate grounds for the PTT return and the audio return.

Unfortunately, a number of my mics have common PTT and audio grounds (and are terminated with PJ-068 plugs), which means that the 6.3VAC on the PTT line runs on the same ground line as the audio return and thus contaminates the audio signal with AC hum.

I really didn't want to rewire a mic with a Heathkit-compatible plug -- I preferred to keep them terminated with PJ-068 plugs so that they're interchangeable among a number of my transmitters. Instead, I made an adapter with a PJ-068 compatible jack and a 4-pin Cheyenne compatible plug. Because the common ground would create a hum problem, I decided to have the PTT switch control a DC, not AC, signal, thus removing AC crosstalk from the common return line.

To do this I added a second, 5VDC relay, and rectified the 6.3VAC filament voltage to provide the voltage to drive this relay (see the schematic above). The mic's PTT button now switches this DC voltage. The new relay then switches the 6.3VAC signal to the original Cheyenne relay.

The new relay also has an additional benefit -- I wanted some way to mute an external receiver during transmit as well as key an external amplifier, and the extra contacts on the relay now provides these functions. (A previous owner of the Cheyenne had rewired the 6-pin connector that connects to a receiver (such as the Commanche) and several of these pins were left unused, so I brought these two new signals (Receiver Mute and Amplifier Key) to these unused pins on this connector).

(Additional note: the 6.3 VAC relay has a coil resistance of only about 8 ohms. This means that it draws about 0.8 A when ON. The only 5VDC relay I could find in my junk box has contact ratings of 1A at 30 VDC. OK, 1A is greater than 0.8A, but personally, I'd prefer a bit more margin. It seems to be working well so far, though.)

(Mounting of the 5V relay)

4. The SPOT switch did not work in STBY mode. Incorrectly rewired by someone in the past, I connected it to pin 5 of the relay (300V when not transmitting, although it really should go to pin 4 of the relay (per the schematic), but the remaining wire wasn't long enough, and pin 5 is a good compromise).

5. VFO tracking way off. Re-adjusted, but during this readjustment I discovered that the bottom-end of 80 meters would quickly diverge despite the rest of the band tracking well. Because I intend to use the transmitter for AM only, I decided to leave well-enough alone and I adjusted the bandspread to track the VFO dial over the range of 3.7 - 4 MHz.

Still to be Resolved:

1. Oscillator: the 1.8 MHz fundamental at the plate of the oscillator 6AU6 (that's later doubled to provide the 80 meter signal) looks terrible, as can be seen in the top trace below (the bottom trace is the RF output (sampled via an attenuator)):

I'm surmising that the signal on the oscillator plate looks this way because the 8.5 uH inductor in the 6AU6 plate circuit is differentiating a 1.8 MHz plate-current pulse-train from the 6AU6 (after all, v = Ldi/dt). But is this the way it really ought to look? I've no idea.

Never the less, the transmitter seems to perform OK and I cannot find anything obviously wrong in the oscillator circuit, so I'm going to reserve judgment...

(Additional note: The waveform at the grid of the oscillator looks great -- a nice sine wave. Just the plate signal looks weird. Ought to have a resonant circuit to make it look nice, I think.)

2. There is a bit of audio roll-off from about 1.5 KHz (down 1 dB from 1 KHz) to 5 KHz (down 3 dB from 1 KHz). The first place this roll-off appears is at the plate of the second 12AX7 stage (yet it looks fine at this stage's input grid). I've yet to identify the cause -- I suspect it might be roll-off caused by the RC network formed by the AUDIO pot and the 12AX7 grid capacitance at pin 2, but...

3. I'm not sure if this is a problem or not, but carrier power does increase a bit as I approach full modulation, even though I've disabled the "control-carrier" feature of the modulator. I'm not sure what the cause is. Perhaps a non-linear PA Screen Grid transfer function? Or...?

4. VFO drifts.

Tuning up the Transmitter:

It's important that the transmitter's loading be properly adjusted. If the loading is too light you'll get peak compression on the RF envelope. I find that adjusting loading for peak power in CW mode actually puts it about where it needs to be for AM.

Here's the procedure I use. It seems to work.
  1. Rotate LOAD and AUDIO controls fully counter-clockwise.
  2. Place function switch in GRID position. Press PTT and adjust DRIVE for 3 mA (or for peak reading if 3 mA cannot be reached).
  3. Place the function switch in PHONE and the meter switch in the PLATE position. Press PTT and dip the plate using the FINAL control.
  4. Switch the meter switch back to GRID. Press PTT and ensure grid drive isn't exceeding 3 mA. (Important note: I've found that as I rotate the DRIVE control through 360 degrees, I hit the 3 mA level at four positions. And for two of these four locations, the output power is greater than for the other two locations, despite the equivalent 3 mA drive level. When performing the final DRIVE adjustment, be sure to select one of the two DRIVE positions that results in greatest output power (at 3 mA drive)).
  5. Switch the function switch to CW. Press PTT and advance LOAD to peak the power output. Dip plate current again, just to be sure.
  6. Switch the function switch back to PHONE. Press PTT and adjust the new CARRIER LEVEL pot (on the back panel of my Cheyenne) to the desired carrier power out (no modulation). Then, while talking into the microphone and monitoring the RF envelope on a scope, advance the AUDIO control until the "troughs" of the modulation envelope are just on the cusp of flat-lining. If you find that the peaks of the envelope reach their max level before the troughs reach their min level, you have too much carrier and you should back down the CARRIER LEVEL pot -- you're just wasting power.
You're done!

Other Notes:

1. One goal was to make my modifications without drilling new holes in the transmitter, so that if someone, at a later date, wished to return the transmitter to its original condition, they could. Fortunately, it was fairly easy to add additional terminal strips using existing screws, and the key-jack hole on the back of the chassis was an ideal place to mount the Carrier-Level pot.

Additional terminal strip for new components. No holes drilled!

2. For proper AM operation, the Loading control must be adjusted for peak power out (and perhaps even a bit beyond, to be on the safe side), otherwise the positive modulation peaks will be greatly compressed and your audio will sound lousy.

3. The 1 uF cap across the new carrier-level pot can be made larger if it's desirable to run the Cheyenne at lower power (and thus a lower potentiometer resistance) and to keep the low-frequency cut-off. But increasing its value will also increase the amount of time that it takes for this stage to reach its bias point when PTT is pressed.

4. At higher carrier levels the Cheyenne has somewhat less measurable distortion at close-to 100% modulation than it does at lower carrier levels (as long as you aren't exceeding the capabilities of the PA on voice peaks, of course). I attribute this to non-linearities in the PA's screen grid transfer function (hypothesized, but not proven).

5. Here's a transfer curve that I've made, using my Cheyenne transmitter, showing RF output voltage (attenuated by my RF "sampler") versus PA Screen Voltage (note: I'm using a 6293 tube in lieu of a 6146). Test conditions: PHONE mode, 80 meters, no modulation.

You can see the curve bending at the high voltages (this results in compression of RF envelope peaks). It looks pretty linear at lower voltages (the slight burbles are most likely due to measurement error (of either screen voltage or RF amplitude) on my part).

(Click on image to enlarge)
6. WC3K's articles in Electric Radio (regarding the DX-60, see below) describe a neat modulation monitor using an LED. There's no reason why this can't also be used with the Cheyenne. I didn't install it, because I didn't want to drill a hole in the front panel and, besides, I use a scope to monitor my modulation. But I recommend taking a look at it. (You can find similar circuits in some of the web sites discussing DX-60 mods, too.)
7. A reminder: Update (2 January 2010): New information on my Cheyenne can be found here.


Heathkit MT-1 "Cheyenne"Information HERE

Heathkit Schematics HERE

Unfortunately, I couldn't find any information regarding modifying the Cheyenne on the web. However, its design is similar to the DX-60, and there are articles that discuss improving AM performance of the DX-60...

Electric Radio articles on DX-60 improvements:
  1. "Fun with a DX-60," Bill Breshears, WC3K, Electric Radio, Issue 133, May, 2000
  2. "More Fun with a DX-60," Bill Breshears, WC3K, Electric Radio, Issue 138, November, 2000
Websites with DX-60 improvements or discussions:

Let's see...where did I put that screwdriver?

Standard Caveat...

I hope you find this information useful, but please, use these modifications at your own risk -- although they worked for me, I cannot guarantee that they'll work for you. (After all, I could have made a mistake in transposing them from my lab notebook to this post.)

If you do find any errors, or if you have any questions, please let me know. Thanks!

- Jeff, K6JCA


Anonymous said...

Nice work on the Cheyenne, and also on your site.

Ted, KX4OM

Anonymous said...

The 25K pot is shown in the ground return of the filaments on the 6DE7. Do you really want it there or should it be on pin 8 in the ground return for the cathode?

Mike N9MS

Jeff said...

Thanks for the note, Mike, and the catch of the error in the schematic. I've added a note at the top of the post notifying readers that there's an error in the schematic.

- Jeff, k6jca