The AM Forum

Working on one of these with the notorious PS problems. It saw a prior fail that took out the RF driver transistors and someone went with 12V driver transistors (2SC1969's) and a 1A 12V regulator, but the regulator blew out.

I know about the new PS available, but am curious the problem with the original PS design to try and keep it
in there due to costs. Another ham in the UK that retained his old PS said he placed 28v muffin fans on the
28V feed to help supress the notorious spikes on the line, which eventually cause the original driver transistors to blow, and likely caused the 12V regulator I saw fry in there as well.

I wanted to know how large the spike is, so I set my Fluke DMM to min/max and plugged in the AC to the rigg.
(with the KW power sw off) It immediately put out a loud BLEEEP ! and remained bleeping , displaying: OL on the readout. So whatever its seeing, its quite nasty. I placed a 48V MOV on the line, tried again, and no BLEEP...max said 31.5V , which is where it stays till you unplug AC power and after a min or so it gradually falls. If you keep AC plugged in, and hit the power sw you get 28.5V exactly which is on the mark.

Another ham who works on these alot said "The reason you continue to see voltage on the pass transistors is due to the legacy AC wiring.  Japanese manufactured products from the 1970s and 1980s all followed a convention to switch the Neutral AC line across the front panel switch.  This leaves the HOT AC line active on the transformer all the time when the radio is plugged in.  When you measure the pass transistors, you are measuring the output of the high side of the transformer in relation to chassis ground, which is tied to neutral at the source in 3-wires systems that are NEC compliant in the U.S.  This is not only dangerous since the HOT AC is always present and is contrary violation of NEC code and should be corrected.   If you use the radio in a GFCI circuit, it will result in intermittent breaker trips.  "

I have studied the schematic and cant see how the neutral is on the panel switch. Perhaps someone here can explain whats going on? I have a decent scan located here : 
https://www.mediafire.com/file/m4kvelb3w5fwwf8/TS-930_schematic.pdf/file

The 28.5 volt overshoot spikes is caused by the regulator circuit unloading at power down to correct this put a 22k resistor across D-7,  goes from Q-1 collector to ground that will stop the 28.5 volt spike on power down.  
1/8 or 1/4 watt is fine.

Rebuild the power supply with good parts and use 2N5886 for the pass transistor if you can find them. Another good pass transistor are MJ802G  mouser has them. Change the caps too.

I don't know what 930 version you have. There were many changes over the years. On the AC side of the power supply early models had a relay to switch the AC to transformer primary and latest 310 serial number the front panel switch did the switching of AC and 21 volt DC sub supply. Some schematics are not accurate in that regard.

No matter what, when you shut the radio off the transformer AC input turns off like any normal power supply so check the AC relay if you have that version, check the front panel switch if you have the later version. The later version many times has the contact welded in the AC side front panel switch.  Powering off the front panel switch, means no power anywhere !!

Ok interesting. This is an early SN, it has the relay for the primary of the power xfmr. The AVR board
is the older version so it does not have the 5 pin jack on it (j7) as shown in the PDF.

The power switch has a DC side and an AC which powers the relay for the main xfmr.

I noticed when you hit the power switch the spike happens,  and you get the 28V till you turn the switch
off, then it jumps to 31 and rests there a while. Will try the 22K resistor.

After more study, I thought Q1 was being kept biased by C6 still loaded up with 28V
and C4 will still be charged as will the bridge caps, so that resistor on D7 may be able to help.
D7 is a RD33FBD-B1, and with a search found this doc explaining the entire PS circuit, says its a 33V zener.

https://docplayer.net/7037156-Circuit-analysis-and-improvements-of-the-ts930s-power-supply-unit-switching-off-your-rig-can-kill-the-driver-transistors.html

Looks like its also talking about the 22k resistor you mentioned.


 

Ok ... You got the old model with the super simple power supply. It has no 21 volt sub supply and should have a bank of power resistors behind the Antenna tuner in an aluminum enclosure. The DC side of the power switch send 28.5 volts to the bank of resistors that powers the whole radio.

The 22k resistor mod still applies to that old AVR board and will correct the spike issue.  D-7 is a 33 volt one watt zener.

Be sure to replace Q-1 with the better part like whats in the newer 930's  I believe it was a 2SB861 or its equivalent

The pass transistors are very important to replace too with better parts like 2N5886, MJ802G, MJ15003G

Also change the three caps with 2200uf 50 volt snap in types. That power supply is very reliable when upgraded with newer better modern parts

Also when you buy parts do NOT buy from Ebay !!  Most all parts are counterfeit, buy only from Mouser or Digikey or known reputable vendors

If you ever are inclined to change out the entire power supply, take some time to go through this site: https://www.w3afc.com/

My early 930 had multiple  common pass/final transistor failures.  Scorches on the AVR pcb made repairs more difficult.  I was an early adopter of the Compudigital supply for my early 930 and it has been running beautifully.  The hardest part was commiting to the swap. This is a major architecture change and not for everyone.  My supply was a smaller version rather than the 20/26 that people use today in the kits. 

W3AFC has also done nice work testing alternative final transistors should you ever need to replace them.

Yes, have heard about the power resistor bank, they say its very inefficient,
and I think the newer 930's dont have it, and the 21V from the AVR replaces it?

For sure lots of counterfeit semi's on ebay, I try to avoid unless its an American
seller and something simple like a resistor or something.

The pass transistors are Onsemi 2N5886G's, they look new, checked on on the teter.
They were feeding a mere 7812/1A regulator on the PA, which probably eventually shorted
with those spikes. The AVR fuse didnt blow as it was drawing less than an amp with the short
so the pass transistors held up ok.

Roger that, I know of that new PS and many hams said to give it a try,
but I am the type who tries to not throw out useful parts like the transformer
if it can be helped. Apparently its a torroidal design, according to KW's tech
reports so it will be low noise / RFI compliant.

I sent W3AFC an email, never got an answer, but yea lots of good stuff on that site.

You can change over to the newer switching supply if you want but it's not necessary. The old power supply is super simple and reliable when parts are replaced with newer robust typed. I have restored about (35) 930's and about (50) 940's over the years and have never had a power supply fail after rebuilding them with better parts.

Good you have the 2N5886 pass transistors those are the best type for the job and what I used on all my rebuilds. They can handle anything. Unfortunately they are no longer available and obsolete. Use the MJ802G or MJ15003G. The main reason for the power supply blowing up is Kenwood use the 2N5885 as the pass transistor. That part is only rated at 60 volts and the input to it is around 43 volts depending on your AC line. Way to close and a few AC s spikes and it fails and takes out other parts on the AVR board. Also if your radio still is using the thermo pads on the pass transistor change over to mica and thermal compound, much better.

For cooling, you can use the stock fan setup and it's fine. But to improve, use a 24 volts computer fan and put a dropping resistor inline to slow it down. Power the fan from the 28V line. Also I find blowing air into the heatsink fins cools better then drawing it out like Kenwood did.

On the driver replacement I use Mitsubishi 2SC1969 and ran them on 18 volts with 1.5 amp regulator. Trying to find original 2SC1969 is about impossible now all are counterfeits. The transistors are good for 25 volts collector to emitter. The IMD distortion is lower and the regulator has to dissipate less wattage. Set the idling current at 125 milliamps. Never had a failure with that setup and radio has full output and low distortion , actually better that the original Motorola MRF485

Be sure to run a wire in the mechanical feed thru's on the Final board, the have a habit of going intermittent. Plenty of info on net for that. Plus it would be a good idea to replace all the carbon resistors in the final unit amp circuits.

Use this regulator for the drivers  L7818CV-DG

https://www.mouser.com/ProductDetail/STMicroelectronics/L7818CV-DG?qs=WHlX%252B%252B9%2FRwDN5eTUn%252BQLtg%3D%3D

Wow, great stuff.

Yes, I put in a pair of 2SC1969's but there are reports lots of chinese fakes out there
that will go into self oscillation, so I am wary.

Also put in Mica insulators and compound, not a big fan od the pads....tho seems KW used them in
alot of stuff.

I used an SI-3122V 12V/2A (TO-3PL) regulator with 0.01uF decoupling caps on the in
and out leeds. The spike killed it though before I had a chance to test the PA properly
and yes I heard 120mA is a good idle for them. Didnt know you can go up to 18V, will
keep that in mind.

Not sure what the "mechanical" feed thru's are. I did see they placed the round lug
under the original driver transistors, so there was kinda a gap when thats done
and I put the lug on top of the tabs and used compound on all the surfaces.

The SI-3122V has the metal tab as the positive input so I had to use an insulator so it didnt ground.
I have a couple of original SANKEN brand spares , so I am glad I thought of keeping some in stock.
Was told the drivers could see up to 2A on voice peaks, so the regulator with the good heat sinking ought to
be ample. Only trouble is that spike, which killed it, and the resting 31.5V after u kill power probably helped
blow it as its only rated for 30V max on the input side. I have some resistors coming in to drop it 10V for 2A,
but at Idle with 120mA, the resistor is only gonna drop about 2v, so am concerned about that a little.....


Oh and did u also reduce the values of the 220 ohm and 33 ohm resistors when you put the new drivers in?

That SI-3122V regulator you used has way to low input specs. 30 volt is disaster that has happened. The L7818CV-DG input spec is 35 volts so it's fine and there's really is no other ones that are higher.

The 220 ohm feedback resistors and 33 across the input transformer I left at stock values. But I did change them because of the values being off from age and heat. Also change the R-10 and R-9 to new ones. Same 22 ohm value.

The feed thru's on the final board are those rivets that connect one side of the board to other side. If those go open or intermittent the drivers are not connected and blow up. So put Z bend wire thru them and all the others on the board.

The one thing I did do is modify the bias circuit per kenwood TS-940 service bulletin SB-988 it amounts to changing R-16 from 1.2k to 2.2k  which lets the tracking diode's have better control on the bias's.

Ok very good.

2SB861 is the original Q1 in there.

Yes, I will reverse the connections on the PS fan, I guess the PA
fan would benefit from that as well?

I read if you lower those feedback resistors, its an improvement
of sorts so I lowered them.

I am gonna try and pull the board out to put the 22k resistor in.
There was what I think is a redundant gnd wire, I left hanging.
I suspect it went to the bridge or the large caps, but I read chassis ground
to the AVR board, the bridge and the caps, so I guess I can remove it?

Update:

I got the 22K in there but left Q1 alone. The big lug with the thick wires
to the 28B side of the fuse cold worked loose as I finagled the board.
After fixing that, inspected the pass transistor connections and one wire
I didnt solder good enough broke off the pin, so had to redo that.

When I powered up, PS was dead, then I thought what did I do with that
extra ground wire...I left it in and it was shorting something out down there.
I removed it and luckily, there was no damage and I now have a nice 28V
with no jump to 31V on power off like I did before.

My Fluke is still freaking out tho, when checking min/max like it did before.
So I guess the spike is still there.

You can lower the feedback resistors if you have 2SC1969 drivers that have a high Hfe. All my 1969 were of low Hfe so I did not have to do any of that and were completely stable. That thin black wire is a dedicated AVR ground that goes to chassis.

The AVR board gets ground multi ways, from the two top screws on the board to the mounting bracket, to the heavy black negative wire to the final unit. But for some reason kenwood felt it needed that thin wire to ground too.

The only reason I can think of why they did it is, if you were troubleshooting the AVR board and had the final unit power leads (the negative specifically) disconnected and had the board physically out of the chassis floating while testing it would then have no ground in that scenario.

If I remember right ( I have not worked on a 930 or 940 in many years ) there is a lug the down low on the chassis next to the AVR board, I think a heavy black negative wire goes there too. That's were the thin black wire goes.

Leave the final unit fan alone, as it came from the factory. It works best as kenwood has it blowing air into the fins.

What spike are you talking about measured with your fluke meter ??

The old 930 brought back memory's so I found my old picture files of when used to work/restore on them. I use to redress all the sloppy ass wiring that kenwood did and detail all that to perfection along with the proper size cable ties.   I even found a picture of a AVR board and setup that I rebuilt for a guy in Australia too.

More pic's. If you notice the last picture is the power supply that's in your old serial number radio

Its on the 28V line, I set the Fluke DDM to DC min-max record
and upon power up, the display reads: OL and it beeps constantly.

I have a big 40KV DC probe that goes into any DMM for reading really high
volts, and the DMM does not go beserk with that, it says the max is 32VDC.

I remember jiggling the Fluke leeds with one hand on the pos, and it seemed to freak out,
so maybe its a false positive.

I dare not use the SI-3122V with that overshoot, and it seems the datasheet on the 3122V is wrong.
They claim input is center pin but when I tested on a bench supply the input is pin1, output is 2 and gnd is 3.
With no load it puts out 13V with about 16V on the input. However, when I placed a 330 ohm resistor as the
test load, which is only about 60mA current, the output went from 13 to 5. So this is a terrible regulator
and surprising to see Sanken create a dud like this.

I will go with the L7818CV-DG you mentioned.

Roger on the photos and the ground wire, I think you are right on the chassis ground post, I saw it
near the screw to the AVR frame down there.

Certainly one heck of a project, but very interesting. Thanks for your expertise. I will post the
next point of progress here once the regulator arrives.

 :)

On the regulator you may want to be on the safe side to use the 15 volt L7815CV-DG version. I don't know if you have original Mitsubishi 2SC1969 and what Hfe they have. I did not have any problems running mine at 18 volts but they were very low hfe. So 15 volts would be safer. Same idling current at 125 millamps.

Not sure on the spike the fluke is reading. I did all my testing with an older fluke 77 before all that min/max stuff came out. Like everyone noted, the only issue with the circuit was the power down 28 voltage rising.

Buy the way, kenwood on very late 930's and new 940's added that 22k resistor update. But they never had a service bulletin on it and or updated the schematic. The AVR board was never update on the 940 and they just soldered that resistor on the foil side thru out the whole production run of the 940.

Then the excellent TS-950S came out which had ZERO power supply problems. Interesting, in all my years of servicing radios I never had to repair a TS-950 power supply and the circuitry was very similar to the 930 & 940's

I would say the bad reputation the power supply got was for two reasons, Not having that 22k resistor & voltage rising on power down blowing voltage fragile MRF485 and using way to low voltage 2N5885 pass transistors and those shorting collector to emitter putting 43 volts on everything destroying parts on the AVR board and blowing the drivers. Also I'll never understand why kenwood stayed with the 2N5885 which were not right for the voltage involved when Motorola had much higher voltage better parts available in their TO-3 case transistor line.

Just replace all parts as we talked about and that's really it. It's super simple primitive power supply and very reliable with newer better parts. My 930 and 940 have been going strong for over 20 plus years since I restored them in 1998



https://www.mouser.com/ProductDetail/STMicroelectronics/L7815CV-DG?qs=T%2FOtf55vL7fQ8ZCL4uIjcA%3D%3D

You got me thinking on voltage regulators and I found a higher current models from ST Micro. Both from mouser and rated at 2 amps in both the 15 & 18 volt versions. They are the thinner single gauge models meaning the mounting tab is thinner metal but from what I read the dissipation is the same. The DG version ( Dual Gauge ) is the thicker tab which seems not to be available.


https://www.mouser.com/ProductDetail/STMicroelectronics/L78S15CV?qs=v7fRTcZn1IeUYNcrkUDSWg%3D%3D


https://www.mouser.com/ProductDetail/STMicroelectronics/L78S18CV?qs=GGwmtXQPF%2FnQmfjz7QpSVQ%3D%3D

I ordered some of the L7818CV-DG, 35Vmax/(1.5A), and some L78S12CV
35V max/(2A). Also discovered Rohm makes a 12V/2A with a plastic tab
BAJ2DD0T so ordered some of those too.

Datasheet says typical input is 4-25V but absolute max is 35v...
and the peak is 50V so I may try the Rohm one first. I have some 6A
silicon diodes from another project that I suppose I could place in series
to drop it a little if its a problem, or with the 18V regulator, do the same to
tweak the output dwn to 15 as you said. 6A a little overkill, probably ought to have
researched some 2A ones before I completed the order.

I think the Rohm BAJ2DD0T is not going to handle the voltage. The L7818CV-DG is a safe deal with it's 35 volt rating. Anyways try them out and let me know on the results.

Got the regulators, and I put the ROHM one in first for a
try. I also got a bunch of SMT 600V/2A diodes, and I have (6)
in series to get about a 3V drop to the regulator as a buffer.

Powered up, have a good 28V to the PA, the diodes bring
the volts to the reg to about 24.7v, the regulator puts out 12v
exactly.

Tried to key tx in tune on CW, dont see any power out,
then I realized I left the Tx-on connector off the PA.


I let her sit there in Rx a while, the PS heat sink gor kind of hot
after maybe 7 mins, waited for the fan to come on, it eventually did,
and cooled the sink blowing air into it, then turned ott after a couple mins....
and the sink was much cooler, so I guess thats normal? I do have the resistor changes in
there for the fan to come on sooner, and run a little faster. The 12K thermistor had broke on me,
and I found one on Digi Key, a 12K enclosed in a ring terminal, so I guess its working.

Also I saw the schematic took the Vcc for the drivers from the main 28V line, but the
stuff out there had you lifting the L17 connection....which I thought went to Q7. I have my feed
on the cathode of D1 which is how the drawing is.

With the Tx-on connector hooked up, I do see idle current on the feed to the drivers,
and was able to adjust the pot for 120mA and the final current for 1.3A on USB.

Went to Tune, and CW into the dummy load, and there is a brief level of good power, then
it falls immediately to about 6w. I tested modulation on SSB, seemed to sound ok, then I hit the proc
switch, now it wont modulate at all.

ALC meter setting is pegged all the time. Seems if I put the meter switch to compression setting,
the power level returns up to 100W or so.

I set it down to TUNE, and let it run there a long time into the DL. PA gradually got pretty hot,
as soon as I thought there was gonna be an issue, the fan finally did come on, so that looks like
its working properly.

First thing is to check the negative voltage on the bottom main board. R-400 runs hot & deteriorates from heat and soldering on board too. One side of R-400 should be around -40 volts the other side will be -12 volts  D-210
is a 12 volt zener that sets the voltage.

If all that is fine, then make sure of all the pots in the ALC circuit,  VR-13 set the base ALC voltage at 3.2V,  VR-9 set SWR protection and VR-10 sets current limiting

VR-11 and VR-12 set the ALC meter. Both ALC meter zeroing and ALC meter scale. Maybe those were screwdrivered

All this is in the service manual  Let me say also Soldering issues and intermittent plugs / switches by oxidation is major problem in fixing those radios. Actually in any old radio regardless of make.

Sound like you got one of those project 930's that has been hacked & butchered by hams over the years. I have worked on those before and have spent countless hours bringing them back to life.

90 percent of hams should not own a soldering iron and should never even look inside of a radio. Don't get me started on newer SMD component radio. The butchery I have seen on those over the years is unreal.

Ok will check that stuff you mentioned.

I work the KW Hybrids alot (TS530/830) and they have a mod where you place
negative feedback on the ALC pin to reduce power out to go down to QRP,
or use a linear easier. I wonder if there is something with the ALC circuit
doing that, or could it be the 2SC1969's I used are not ok with those decreased
resistor changes I did?

I dont get why selecting compression with the meter switch gives you full power.
I did notice the mike jack is a bit loose, like someone was working on it.

I looked at the schematic and see no reason why just putting the meter switch in comp position would affect the power output.

Normally when the proc switch is Off & you put the meter switch in comp position it will read nothing until you turn the proc switch on, then you set the compression level from the processor " In " front panel pot and the meter will respond as you talk.

Can you see if anyone did any modification to the radio ??

It's possible the 1969 are oscillating, but that normally would make the ALC go nuts. The comp switch affecting power is a new one on me.

Nothing immediately obvious like mod work at a casual inspection. All the connectors seem seated ok
on the main board.

I know with that ALC mod on the KW Hybrids I mentioned, if you have it applied,
and you set the meter sw to ALC, its usually pegged. Because I guess of the negative feedback
from the 9v battery in reverse on the ALC line. So perhaps thats applicable to these radios as well.

Does the driver section connect to the ALC circuit somehow in the schematic? ALC
meter is pegged whenever I am in tx, so its def goin nuts.

I was told there are cheap chinese 1969's out there, prone to self oscillation
if used in this kind of config. The ones in there now were given to me by another ham,
not sure where he got them. They were not perfectly matched beta, but close. The ones
that were in there were perfectly matched, cant remember what they measured for gain on
my tester tho. I remember reading that lowering those "feedback" resistors helps stability
but you said only if the gain is low. Perhaps if its high, it de-stabilizes?

Months ago I was gonna bench test the PA, someone said put 50 ohm load resistors on the
output RF and the input for "balance" to do an idle current test. When I applied 28V, the
resistor on the output jack fried - it was only 0.5W and with the input jack basicly tied thru
gnd, in / out were connected, so I figured it was natural self osc of some sorts. But now I think
it could be these 1969's. Since I have the tabs above the pcb now, they will be simpler to swap out,
maybe I will try the originals which are a perfect matched pair.

Also this site says there are subs for the 1969 (2SC1969 �> 2SC1944 2SC2050 2SC2098)
https://masterelectronicsrepair.blogspot.com/2020/08/table-of-2sc-series-transistor.html
Dunno how accurate it is.

What the radios current ( IC ) meter reading when the ALC is pegged ??

If it's oscillating it will be pulling current above idling as a sign.

Simple test, Disconnect the 28 volt power from the final unit to disable it. If the ALC meter goes back to Zero in transmit then the final unit is oscillating.

If it stays peg out in TX with the final unit disabled, then the problem in the bottom main board.

All this is providing that the ALC zero pot on bottom main board is adjusted correctly to zero out the meter

The ALC signal is from the directional coupler circuit in the low pass board. The forward ( D-15) diode provide rectification (VSF) to ALC circuit and ( D-16) reverse diode (VSR) for SWR protection.

Same basic circuit setup used even today on the newest radios.

With the 28V power disconnected from the PA, the ALC meter is still pegged,
so off to the big main pcb.

That resistor you mentioned, its on the foil side I guess?

Have you tried zeroing out the ALC Meter with VR-11 ?? someone may have fooled with it possibly.

The negative voltage resistor is  R-400 1.5k it's right next to plug 28 in the rear middle of the board.  The 12 volt zener is D-210 that is by plug 14 middle right on board between the two metal shields. All parts on on the component side of the board.

Just tried VR11, does not change the meter at all regarding ALC in tx mode.

Tried to locate that resistor, still looking. I found R399 over in the corner on the drawing
near Q71/72 , thinking R400 would be closeby but I dont see  it. 3070025 is the SN,
so its an early one.

As I inspected things more, I saw what I thought was a RED LED glowing.
Turns out D67 on the main board is now an LED. Red is shining thru the
glass. But I guess a false alarm, those diodes normally are suppose to do that...
very odd.

SM I have shows no schematic of the signal unit, also have a hires scan of the interconnect but
also leaves the signal unit empty.

After more research, found this PDF where it describes corrosion under the local unit as the cause
of the pegged meter in ALC mode. Theres also a video link on pg7, and the description in the video details
is exactly what I see, full power then a drop off with the meter pegged.

So I guess its time to remove the local board for a peek. I also will need to confirm all the plugs are in the proper
locaitons, as the doc sites its possible to place some where they dont belong and do some damage needlessly.

Corrosion is a possibility, who knows were it was stored. Although I never seen corrosion on the rigs I worked on, it could happen. Pull the board and inspect it. Here are the schematics for the bottom signal board and over all interconnect. Should make your life a bit easier.

Here's the Interconnect board schematic, plus this might be a better service manual than the one your using. It's the later one with both old and new serial numbers radios.


https://www.kg3m.com/sitebuildercontent/sitebuilderfiles/ts930_service_manual.pdf

Thanks for those diagrams.

In the PDF I posted, he I guess removed the DC side of the power switch
and jumpered it? Had read about that in my travels, but not sure if it really makes much difference?
They do switch the neutral side of the AC plug, and I hear thats not a desirable situation, so would it be
better to move it tot he line side? This one has the AC relay.

Was able to pull the local board, (attachment2)and the foil side looks like a modding madness episode. I guess the yellow jumpers are KW factory, as probably are some others. There is a black MOV near the long heat sink, plus a trace is cut. I see some foil burned near where you were saying that resistor is near,  and that was indeed R400. It measured 2.7K so it was shot. I had new 1/4w 1.5K's in stock so put them in series, then parallel to form 1.5K 1W.

1st photo shows some mods, a 2SC2458 tacked in with what look like may be shottky diodes?
Some others nearby look like they are using thermistors, or MOV's.

All those mods are factory changes. The radio had many, many updates and changes over the years. Then at 310 serial number they incorporated all mods into a new board.

I changed R-400 to a 1 watt small type resistor just so it looked correct on the board, what you did is fine.

Nothing wrong with the AC side of transformer. It has a fuse, relay to do high current switching and computer cord with chassis ground. All works fine, No safety issues  & no need to re event the wheel.

Ok then I will leave those things alone.
I read that R400 causes loss of AM power, but is it also
the reason for the ALC pegging the meter?

Since I messed with VR11, I will have to do the ALC calibration in the manual.
I believe you have to inject a strong signal into the mike connector?
At this point I may as well do a complete alignment and not assume
anything.

Signal board is back in, but now there are new problems.
VFO does not change freq, no audio.
Thought maybe the batteries I had in glitched the memory, but still the same after I removed the batteries.

Turned out the Dlock switch got hit, so now the VFO works.

Have a humm in the audio thru headphones I can hear with the vol all the way up and some sorta scratchy noise.
RF gain control only makes the meter read about half way when fully turned. Tuned to a local AM station, comes
in a bit lower than usual.


Dunno if this is due to me fiddling with the ALC potentiometer? Didnt try any tx yet.

You need to be sure you have negative 12 volts, that runs many circuits in the radio. One side of R-400 will have approximately -40 volts, the other side should have -12 volts. The -12v is regulated by D-210

Was the receiver working ok before the power supply problem or was the rig a Non working basket case ??

I will check that resistor for that voltage and look for D210.

Rx was fine before I removed the local board, the 12V regulator
on the PA was a mild short causing almost 1A draw on the PS, so it got warm
but the rx seemed to be quite good when I put it on my station antenna.

I moved the wire bundles around while listening, almost sounds like when
you have a mike with a loose ground someplace , that rustling and popping type noise.
I have also heard the same noise when an electrolytic cap goes high esr in audio circuits
or there is a poor ground. But I guess improper 12V could also produce the same effect.

You probably notice you have to use quite a bit of force to unplug / plug back in the wiring harness plugs. Many times you can create a new problem by cracking the solder on the board header pins, cause poor/no connections or intermittents. Kenwood soldering was not great in those early years. Also check the TMP connectors on the mini coax's and board plugs, verify you have all wiring plug back in correctly.

Con (30) pin 4 has 13.5V, Con(26) pin 5 has -44v. so that looks good.

I vibrated the board in lots of places, wiggled on the connectors while
listening, no difference.

There was one of those solder lug screw down terminals on my bench after working with the board,
looked like it was not ever soldered on the lug side. There is like a yellow wire with a lug down near the
wire bundle near the filters area, its secured with its proper terminal, otherwise have no idea where this one came from.

I can hear the marker signal, very very faint loudest on 7mc on the dot....and the humm in the audio is loudest there, most discernable on AM. Meter on ALC reads about half scale, same if you go to proc. RF gain moves the needle only a little bit, and the audio humm comes down a little but its not deaf.

Figured I would trace back from the RF gain pot, heads into Con (19). Have zero on pin1, 5.2V on pin 2, 18V on pin 3 but nothing on pin 4. Pin 5 is like 2.4V but if I move the RF gain control it varies a little.

Someone definately spliced the grey wire on the conenctor to pin 5 at one point. Feels sort of flimsy, so cracked the shrink tubong open and the wire came apart. I respliced it, and still only 2.4V on that pin. If I unplug that connector, hum gets crazy loud, so I guess its because the AGC is disconnected. I saw they said to have it in a sideband mode with the RF gain max, so when I moved it to side band and put the gain to max, I have 18V now on pin 4. Pin5 reads 18V with the AGC sw in slow, and -2.5V in Fast and off.
 

All this 930 talk got me to realize that my TS-930AT is sitting here, gathering dust. It works, but it's obvious that the filter caps need to be replaced. I'm not in for the job, so I think it's going up for sale. What is the going $$ for 930's? I know I could probably get more for it as parts, but I hate to kill a working radio. EBAY prices are no longer a reliable source of worth

Havent removed the board yet but I saw a Service Bulletin no. 875 from KW about C642 being in possibly backwards.
Hard to know what direction it belongs. https://www.qsl.net/bg4aaf/book/kenpic/ASB0875.JPG

The one I have has the + side of the cap facing the IC3 on the board. I think thats the proper way but wanna be sure.

Removed the local board again, went over it and made sure the ground points were clean, but still same problem.
Hear the marker signal best on slow agc, otherwise its very weak.

If I hit the DIM button, the S meter reads a much higher signal level, and the audio comes up almost normal.
I see the DC/DC converter on these are known issues, perhaps thats loading down a voltage line?

So on Con (1) of the Digital unit should have 5,12 and 24V on the pins. I have 9,17 and 24V.
Con(9) pin1 is good with 12V.

Not sure where the D6-D8 zeners are located.

Con(29) on the local board is suppose to have 23.7V on on that pin which seems ok, so I guess there
are problems with the Zeners, if I can find them.

So the zeners are not in this radio, they I guess are suppose to be in that big resistor module behind the ATU.
I opened it up and just the resistors are in there.


After recapping the digital board and resoldering it, still had the same problem.  Also bench tested the 5 and
12V regulators ok on the heat sink.

I found out I messed up my R400 and it was really 6K because I forgot to tie one end closed.
I put (2) 3.3K/0.5W resistors in there for about 1.6K which ought to be ok.

But upon powerup, R29 (330ohms) on the digital board smoked. I dont have a short on either side
of my new R400, and I read a diode in reverse polarity on one side, which is probably the 12V zener,
so its not shorted.

The DIM button was out when that resistor smoked.

Turns out someone put a jumper on the foil side to make the trace that seems to be open
that ties one side of R29 to pin 5 of con(2) so I saw it was shorting to ground and redid it,
and replaced the resistor. Now the meter swings all the way to the left pegged, no display.
Wrong voltages now on R400. I will have to look at it again and confirm R29 is going to the
right pin.

Ok so looks like D210 the 12V zener was breaking down under load. Now the -12V line is stable
and no more s-meter abnormalities. Rx seems back to where it was and the 28V line is still reading good.

So I went back to the tx tests, and now there is no foldback to 10W, but its never hitting the 100W
level like it use to , no matter where the CAR control is set. Most output is about 55W cw into DL,
with either TUNE or CW set, it wont go past that level. The 28V line goes down to about 27.8V in tx,
not sure if thats acceptable?

Re-checked and set idle currents, output from the new 12V regulator is rock solid. 50W
is about 7A PA current draw. I have 2.31V as per step2 in the tx voltage setting for the ALC.
I saw I missed a nearby ground screw close to that spot on the board, so I put it back. Voltage still ok,
but now the most power out is 10W.

Well after alot of troubleshooting, some of it due to a bad coax
on the output of the PA, now have 100+W on the output finally.
12V reg is rock solid. Did some SSB tests and got great audio reports.

If I throw a long carrier into the DL, sometimes the s-meter scale for current
reading jumps around and seems not to settle in. As the PA warms up, current
goes up almost 1A from 8A, then the fan brings it down a bit. I notice when you put it on
the next day the power out is past 100W when you set it there prior, and you have to
back the CAR control down again, not sure if thats normal?

Have not tested FSK, I think thats a higher duty cycle than SSB? Never used that mode.

I will recheck the idle currents and monitor them during the fluxuating current to see if
its the drivers or the finals. I felt the part of the heat sink where the new 12V reglator is mounted
and it seems hotter than the rest of the heat sink, so maybe its working near its limit....I think
we could put some heat sink compound on the ears of the PA heat sink where it mounts to the
rest of the frame to give it a little more headroom.

Just my opinion:
Short answer is yes a higher duty cycle but can be easier on the finals.

TL;DR:
FSK is a 100% duty cycle. It's technically the same as FM voice, except the FSK frequency changes in a square-wave fashion from 'mark' frequency to 'space' frequency, instead of smoothly according to the modulation by an audio signal.

The beauty of FSK and FM voice is that the linear P.A. runs at its highest efficiency when set to full power.

The caveat is that if you reduce output power, the efficiency on FSK suffers for the same reason that it suffers at carrier-only level on AM. On many transceivers, only for SSB is the internal fan and heatsink close to adequate.

It may be worthwhile to aim a small 12 Volt centrifugal/squirrel cage blower directly at the heat sink no matter what you want for a mode.
If nothing else it is a quiet helper and should reduce heat, prolonging the life of the equipment. Something like this is what I use on a TS430-S:  https://www.amazon.com/12VDC-Squirrel-Cage-Brushless-Blower/dp/B00HWZ5XGC  but anything will help.

Thanks Patrick for the info on FSK.

Well the display developed a problem after running the rigg a couple of weeks.
The last digit was more brightly lit than the others in either dim or regular mode,
no matter what frequency, and the red pointer that follows the vfo to tell you what part of
the band you are on , had another pointer appearing 200kc apart from where you were,
lighting up gradually as you approach an even frequency like 14.000 with zeros in it,
then it would fade out as you went higher.

After shooting the digital board with heat and cold spray, no difference. I took the digital board out
again and redid more solder connections, no help. Vibrated the board and specific IC's and wiggled the
cables, no help. Even took down the display board with the display, cleaned, defluxed it and resoldered,
still no help.

That was last week and today when I fired it up, still a problem but after I went to different bands,
moved the VFO up and down the bands many times, the brightly lit segments slowly faded back to normal
intensity and its ben running like that about 1 hr. I shot it again with cold spray and vibrated the board
but no change.

Do VFD's GAS like vacuum tubes maybe?


 

Been a while and wanted to thank KI4NR again fo the support.

Turns out Q1 on the AVR board developed a fractured leg, so I replaced it
with a 2SB1136. That seems to have cured the problem with the VFD.

On of the dual gate 3SK72 fets in the xmit RF chain was degraded
causing a poor inout signal to the PA and also the PA had a bad output
coax to the LPF that had to be respliced and repaired.

All running good now. Just got in a newer SN 930 with no display lighting up
and the s meter pegged constantly. I plan to get it going as my own radio.