I put the “troublesome” TS-930S back on the bench again yesterday to investigate its PA instability some more.  A few points are worth mentioning:

1. The instability onset power gets lower with increasing frequency.  That is, at 7 and 14 MHz, the instability onset is at about 60 watts output.  At 28 MHz, that drops to 40 watts.
2. I was able to increase the onset power slightly by increasing the resting bias of the MRF-422/2SC2510A pair from 1.1 A to 1.5 A.  I don’t know if that was a real effect or just a coincidence, though.
3. I tried adding 0.001 uF, 0.1 uF, and 4.7 uF capacitors shunting the 28B line to ground to remove RF that I observed on that line during transmit.  I don’t have a good measurement of the amount of RF, yet.  But, I know that it’s the same frequency as the transmitter because it varies as I change bands.

I still need to find that Helge Granberg article from RF Design to locate the references.  Still counting my lucky stars that I haven’t destroyed anything in the course of these experiments.

I’m not sure that I’ve shared this on the blog, but I’ve long known that the reason that my “troublesome” 930 burns up PA drivers is a low-frequency oscillation.  This mechanism was also mentioned by Leeson in his comprehensive list of 930 upgrades and repairs.  At first, I thought that the 28-volt power supply’s filter capacitors might be the culprit—insufficient ripple-suppression.  After all, the oscillation sounds like it has a 60-Hz component (I’ve never tried to measure this because I’m concerned about blowing up more transistors).  I finally completed the capacitor replacement last night, realigned the power supply voltage and PA bias levels…and the oscillation is still there, just as I increase the drive past the point where I get 50 watts out.

There is a good Helge Granberg article in the September/October 1980 issue of RF Design titled “Good RF Construction Practices and Techniques.”  I would like to get a copy of the original article because Motorola, in their infinite wisdom, reprinted the article as AR164 but neglected to include the list of references.  Granberg devotes an entire section of the article to instabilities, a full of half of that section to low-frequency instabilities.  This is the area I will focus on first:

Causes for the low frequency instability are usually inadequate collector DC feed bypassing or an extremely poor ground in that area.  Two or three RF chokes together with various values of bypass capacitors from 1000 pF to several uF may be required in the DC line to stabilize the circuit.  (See examples in Reference 1.)

Ah, Reference 1, where (who) art thou?

If I ever get to the bottom of this problem, there should be a useful body of work and notes to prevent this from happening repeatedly to others.

This is just a quick note summarizing my repair of the 5-million S/N TS-930S.  Both the fragile MRF-485 drivers and the hardy MRF-422 finals were toast.  The power supply pass transistors (2N5886s) were also dead.  Two electrolytic capacitors in the PA bias network over-voltaged and exploded.  Original cause unknown, but certainly exacerbated by a slipped probe when setting the driver bias previously.

I almost wrote this one off and parted-out the goodies from this radio to finance a new radio.  But, when I saw the price of a used Orion or properly-configured K3/100, I realized I could repair this one a lot of times and save my money for my newly-acquired VHF+ habit and the eventual, mythical, first tower.

To begin with, I got the power supply working again with my last set of spare 2N5886s.  Should buy a half-dozen more of these to keep on-hand.  With the PA disconnected, I set the 28B line at 28.5 volts.

Now for the big job:  I tore-down the entire PA board, cleaned it, and cleaned the heatsink.  Replaced all of the bias resistors with new carbon composition types from Mouser.  Replaced the MC1723 regulator IC in the PA bias with LM723.  Replaced all electrolytic capacitors in PA.  Replaced drivers with MRF-485MP-HB (matched-pair, high-beta) from RF Parts.  Used all-new mounting hardware and mica insulators.  Replaced finals with 2SC2510A-MP, also from RF Parts.  This part is essentially the Toshiba version of the Motorola MRF-422, but at about 1/2 of the price and a little higher beta.  Yes, I know the higher beta may cause instability if run at full output or poorer IM if run at lower power.  But, I was in a thrifty and experimental state of mind.  Performed the PA bias alignment according to the Service Manual.

I fired the radio up into a dummy load on my bench (which is in the basement, two floors away from the shack—note to self:  this was a dumb idea).  No output on CW.  Check ALC—pegged.  Check Ic meter (PA collector current)—pegged.  The radio wasn’t groaning (or popping/stinking) like a circuit drawing too much current.  So, I figured it was in the current transducer circuit.  Righto!  This circuit contains a beefy 0.5-ohm cement resistor bolted to the power supply heatsink—basically they use the voltage drop across this resistor to estimate the current.  One of the sensor leads from this resistor read 0 volts.  After a good deal of tracing and troubleshooting, including inductors on the power supply board, I tracked it down to a bad trace near the connector on the power supply board, which was easily shorted with a small piece of wire.

That fixed it right up.  The radio made slightly more than 100 watts on CW and about 55 in TUNE mode, just as it ought to.  Next, I noted that the PA and P/S fans did not come on while I was transmitting, even though the heatsinks had become appreciably warm to the touch.  So, I hard-wired the P/S fan to the 8-volt line in the fan case (see the W6NL mods for more details).  In the case of the PA fan, I hard-wired a 7812 regulator (with an appropriate capacitor–0.68 uF/50 volts) to the 28B line on the power supply and brought 12 volts out to the fan.  So, both of the fans run all of the time now, which is a better arrangement.  Plus, I now have a regulated 12-volt source easily accessible inside the top portion of the radio.

I put the radio back together and took it up to the shack.  Made a few QSOs in the Michigan QSO Party and it seemed to be working great.  However, I did hear a little popping on the second radio.  I don’t know if this is a symptom of poor isolation in my station (need to follow the K9YC directives for SO2R bonding) or a power spike or IM problem on the radio.  Probably should make some measurements on those things.

One of my TS-930s (2-million S/N) has always been a little touchy when it comes to the synthesizer PLLs. A while back, I put new electrolytic capacitors in the power supply because the old ones were drying out, reducing the regulation. After setting the supply voltage back to 28.5 volts, I noticed the PLL wouldn’t lock (indicated by the letters PL.L on the display—this radio has the PIEXX digital board upgrade) on any band.

Thanks to KA5IPF, whose expertise has been invaluable to me as a Kenwood owner, I learned that this was probably due to the master heterodyne oscillator not oscillating since it affected all bands. I tweaked L77 on the signal unit (front corner of the radio behind the PITCH/AF TUNE knob) until the oscillator dropped out, then backed up until it came back and dropped again. Then, I split the difference, just as Clif recommended. I had the radio on 40 CW at the time. And, it was so touchy that when I changed bands, it would unlock again. So, I got it dialed in on 40 CW and left it. This was fine since I was operating domestic contests at the time and having a radio on 40 all the time from the Midwest is usually a good idea.

But, now that I’ve moved to the East Coast where DX contests are the name of the game, I’ve decided that it would be a good idea if the radio were more frequency-agile. I pulled it apart last night and got to the PLL board, which is notorious for cold solder joints. (Is it a bad thing when you “just know” how to get to any circuit in your radio in 10 minutes or less? Probably.) I didn’t see anything that looked visibly cold, but I went around and reflowed a bunch of joints and, fresh off my success with the TS-700, I cleaned all the connectors. Then, I repeated the L77 master heterodyne oscillator adjustment.

It seems to be working now, although I need to give it a good run in a contest to be sure. This leaves only the radio that eats expensive transistors like popcorn to be repaired…