KA-9100 output board resistor value wrong/contradiction

[repairmaster]

Contradictions for values of R23/R24 as 100 ohms or 1K ohms:

I am repairing a KA-9100 and procured a salvaged TA-100W.
Also replacing Q5 Q7 R23 R25. Q1 & Q3 check good.
During this repair I found schematics online. Checking for other suspect parts, I noticed R23/R24 on two prints shows 1K. Other resistor in good channel is 100 ohms.

I reviewed a pic I had seen earlier, it appeared to have a 100 ohm. I searched out another pic, and it definitely has a 1K, but is in German, and appears to say, repaired/refurbished and reinstalled board.

[comparing to KR-9600, with similar circuit, also has 1K ohm indicated]

My calculations predict a 100 ohm would allow 365mA, and 21W across R25-Q5-R23, and 13W at R23; in non-linear operation (eg. V/I limiting).
With a 1K ohm for R23, values would be 54mA, 3W total, 2.9W at R23.


So I have several examples indicating 100 ohms (including the unit in front of me), and several sources indicating 1K ohm.

Should I use the original 100 ohm? or 1K ohm?
Has there been a production line change, a schematic error? and/or repairs done using wrong values from incorrect schematics?

I am accustom to seeing emitter-coupled pairs either having a constant current source, a high value emitter resistor, or high values in the collector load resistors.

Any engineer types out there to comment?
Anyone with an unaltered KA-9100 care to see which value they have in their amps? (and serial number of unit).

If this IS a production line change, then changing your own R23/R24 to 1K ohm may prevent burnout of Q5/Q6 during current limiting in the output modules, and subsequent burnout of the output module(s).

With these modules being 'unobtainium', and the value of these vintage units, it would be good to find a definitive answer to this question.

Thanks for any insights!
Mr K L
[I can provide links to the internet pics and schematics, if needed]

 

[pdm4606]

res. values

I rebuilt an old Dynaco PAT-4.
The resistors originally installed were all over the board. As much as several hundred ohms above and below schem. spec.
I of course bought new better quality resistors that were right on.
The improvement in sound after the rebuild was very noticeable. I believe it approaches high end.
Along with new trans. at q4 the unit has a new life.

Paul

 

[repairmaster]

Thanks pdm4606,
I have a tendency to be long-winded, so I left out some details:
I was talking about the marked values, not measured.
The origin values on some KA-9100s for R23/R24 is 100 ohms.
But the schematic indicates the values at 1K ohm.

I have seen pics of both.
What is unknown, is the pics with 1K are units that have been repaired, and is the value 1K because someone used the value indicated on the schematic, and not the original value installed in these amps.

Schematic errors or production line deviations are not unheard of.
I am trying to sort this out for this repair.
100 ohms seems too low for this circuit, but I am not a designer, only a tech.

Thanks for the quick reply!
Do you think this post should be on 'DIY' or 'Vintage Solid State'?
I asked chat room and they agreed I should start here...

Mr K L

 

[srinath]

Hey repairmaster - nice name, must prompt a lot of silly questions like this one about to follow - he he.

OK My KA9100 has a phono stage that loses 1 channel. I have great stereo sound with aux or tape or tuner. Just not phono. I suspect a transistor or 2 in the phono board. Can you tell me the best way to remove the phono board. I am very reluctant to remove those wire twists cos I invariably break them then have to solder them etc etc and dont want to do that. I had a thought about removing them from the rear, but maybe there is a better way - Like removing the power amp section and heat sink and getting at it that way.

Cool.
Srinath.

 

[repairmaster]

Nope, not been here long enough (2 days). Yours is the first.
(I DID see another post about the preamp while researching, just like yours, was it you?)

I had the same though while working the amp board, which I just removed with wires still wrapped, like you mention; as I did not want to unwrap them either.
They loose their integrity, IMHO, when unwrapped and then rewrapped.
You could just unsolder the wrap pins...

Did you eliminate the switch as the cause?
Try jumpering the phono position contacts while switch is in 'phono, and see if the channel comes back...

Chances are someone may have put line-level, or even speaker level into the preamp phono input. You can compare voltages at the transistors from channel to channel, if you don't know how to do junction tests.
I see a 'sticky' post, in DIY IIRC, about testing transistors, but have not viewed it...

Thanks for you question anyway,
Do you think it is OK to PM one of the senior members I have seen posting about KA-9100 outputs, with my question?

Mr K L

 

[srinath]

Plenty of people with switch and similar problems in this amp.
I started out suspecting something else, then zeroed in on the phono pre amp. So maybe it was mine. I think, If I get that board out, I'd fix it, but I also am not too keen on desoldering the pins, they are rather lets say, a square peg in a round hole. They're in there and hard to get out.

Let me see, its now in storage, I would have to get it out before I start fiddling wiht it.

Cool.
Srinath.

 

[repairmaster]

Update, but still desire feedback...

I did some further searching and determined that 90% or more of the units in the field have the 100 ohm resistors installed.
I studied the circuit some more and noted other limits on maximum current in that stage.
365mA would only flow if/after Q5 or Q6 shorted.
While intact, only about 100mA could flow due to negative bias at R25.
This is still over double the ratings for that transistor.

Changing R23-R24 to 1K ohms would limit the max current to about 50mA.
This is the max rating of those transistors...
And wattage would depend one duty cycle and waveform, and power transfer function; and I have not tried to estimate that.
But my 'guestimate' is about 1/2W to 1W during clipping/limiting with 1K ohm resistors, close to the capabilities of those transistors to handle for short periods.

I saw or got feedback on 10 or more units (4 loose boards I counted as two units). All but one had exclusively 100 ohm resistors...
The only unit I saw with a 1K ohm resistor was a lone example, and was on the German site that retrofits the modules; and the caption labeled it as a repaired or rebuilt board, implying image was taken post-replacement of parts.
I suspect they use the schematic indicated value, but I have no information if they checked on this, or if they have seen 1K in any other stock units before they have been altered in any way...

I contacted another member who verified that all his units and extra boards all had only 100 ohms resistors on them.

After considerable indecision, I chose to install 1K ohm, due to the estimated conditions I calculated. I changed the good channel to 1K also...
I ran it overnight, and also for periods of moderate output; enough to warm the heatsinks good. No apparent problems.

Nevertheless, I think most or all KA-9100's were built with 100 ohms, and the 1K ohm is a schematic error, or a later version print with an improved value. Either during, or post-production?...
I would still like a definitive answer though: did I re-engineer the circuit, or am I merely installing the 'correct' values?

Please let me know if you find a different version of the schematic, or see an original amp with 1K ohm resistors installed at R23/R24.
I am very interested to see if there are any other instances of KA-9100's with 1K ohm for R23/R24!

Grateful for this excellent organization,
Mr K L

 

[avionic]

The simple rule of thumb. Go with whatever is installed--unless it is grossly obvious that it's not OEM. Schematics and service manuals are notorious for being incomplete and incorrect as well as circuit board silkscreening.

 

[srinath]

R23-24

Where is R 23 and 24 - power amp section - I should check mine.
Cool.
Srinath.

 

[repairmaster]

Thanks Avionic;

I usually tend to agree with that.
That is why I am so torn deciding in this instance.
If it was reversed, I would be positive it was a misprint.

I have worked for years in consumer repair and warranty work so have seen my share of errors.
If it was 1K in the unit, and 100 ohms on the print, it would be obvious.
Trouble is, 100 ohms is too low for that location, and those transistors are burning out in the field, which corroborates that.
So it could be an intended change or production change (or an assembly line print error).

Every time I decided one way, I found evidence to the other way.
First unit had 100, then schematic had 1K (so vote 100); then did calculations to confirm, (changed vote to 1K).
Then another two pics with 100 (vote back to 100).
Then found another picture on the internet with 1K (from a rebuilder whom reversed engineered the next stage, so they have talent); (vote 1K).
Then more pics, all with 100 ohm (vote solid 100).
Then learned these are overloaded and burning out in the field...
Redid/reviewed my calculations, and considered evidence that this circuit leg burns out prematurely, cited 3 cases out of ten or less examples; (vote back to 1K). Installed 1K and tested it.

I was hoping to establish a pattern, and eliminate/establish one or the other...
That leg operates in SOA for that transistor during normal drive; but in I/V limiting, it will exceed the ratings of that transistor; and then the resistor, if the transistor avalanches.

I am attempting to contact the German rebuilder, that produces their own output package (not this board, the stage it drives), and ask them if they have confirmation for the 1K value they are using on this board.
Anyone here speak German?? (seriously, their website has no English option, and have not replied to my email in English).


srinath:
Those resistors are on the collectors of Q5/Q6 to ground (left/right driver board, mounted on/below the output chips. They are cross-wise, laying away from the module mount, near the center of the board. They are visible on internet pictures. The stage drives the output modules.
I will try to attach pics, and a schematic...

I would be very interested to hear if yours are 100 ohms also. (or any signs of overheating).

First pic from this thread has 100 ohm standard metal film:
http://www.audiokarma.org/forums/showthread.php?t=296250&highlight=KA-9100

It has a retrofitted Sony module on it, but the original 100 ohm resistors.

Second pic is from this website, and has 1K ohm flameproof:
http://www.audioklinik.de/beispiele.htm

It is labeled (Google translated) as rebuilt or refurbished.

The third pic is of the right channel from this website has 100 ohm:
http://www.oaktreevintage.com/web_photos/Stereo_Amps_Pre/Kenwood_KA-9100_Integrated_Amplifier_inside.jpg

It is overexposed but looks like an original metal film...

Will try to attach a schematic also.
Someone let me know if these worked right/are clear, never attached in a forum before... :^O

srinath; I saw your phono question in the thread for output mods too... (I trust you're keeping relevant?)

Thanks for you input, I am still undecided after doing those rough calculations...
Mr K L (Kim)

 

[srinath]

Brilliant my friend. I am gonna print this guy out, cos when I get my amp in my hand out of storage, I'll not remember this ...
Cool.
Srinath.

 

[SicMan]

well an easy solution is to put in replacement transistors with more robust power handling capability. Replacing the originals with original transistors will still give you problems later on, Kenwoods design runs those transistors and most of them to at least 75% of their spec. When they get old and have been heated for so long they start to fail. (most are obsolete anyway)
29 out of 30 KA9100's I worked on had the 100 ohm resistors in them. For me to explain to the "customer" about this problem is futile. I rebuild the amp boards with better transistors etc and all is well with or without the 100 ohm resistor.
I agree all this theory is fine for people in the know.

 

[repairmaster]

Thanks SicMan,

I upgraded the transistors also (Q5,Q7).
I used NTE32, 900mW, 1A; to replace the original A912's at 750mW, 50mA (IIRC).
Was in a rush to find a sub, as NTE distributor sent NTE288's which are only 625mW.
Interested if you can recommend a better one.

Did you experience many 100 ohm resistors burned out?, or was that infrequent? (I mean compared to module failure. Does module failure usually mean R23/R24 100 ohm burned too?)
Still I am happy with my solution, but a more complete history would be good for this thread for others encountering similar failures.
If it is infrequent, maybe it is a non-factor...

Thanks for you help. I was going to credit you above, but was hesitant about mentioning names from other threads. I though name-dropping could lead to uninvited inquiries or excessive questions to the people mentioned... (like putting someones email in a thread...)
I often reference the thread instead.
So thanks, SicMan was my source above for verifying these values on some of his boards, and helping me via some PM's I sent. )

I hesitate to go this deep, but, I left this background out of the discussion:
What is different (in my limited experience) about the Kenwood output chips is their means of internal I/V limiting.
During limiting, the internal protection transistors shunt the drive current directly from the input pins, to the output pin 5!
This shuts down the output transistor drives.
This means the drive current is directly driving the speaker load. (check it out).
If this chip was driven by a typical 'totem-pole' or bootstrap arrangement, they would burn out (AFAIK).
Instead, Kenwood drives their output module with a current limited differential pair; limited by current source Q11/Q12, set by D3/D4 and R31/R32.
This limits all dissipation's in the circuit; except, I believe they overlooked the idle-sink-leg's current parameters - in out of closed loop gain (Q5/Q7)...
Feedback is no longer closed, and the drive circuit will saturate. (due to the configuration of their internal limiting).
During the negative drive signal, the current at Q5,Q7 is no longer limited to the ~12mA current source; the positive saturated drive will still be limited to the Q11/Q12 current source, but the negative drive saturated current will only be limited by the saturated R23/R24-Q5/Q6-R25/R26; and the maximum saturated negative drive voltage from Q1/Q2 of ~51V at the base of Q5/Q6 (~6V below B+ and consequential negative feedback at R25/R26); limiting THIS stage to ~97mA during the negative drive, not the ~12mA during the positive drive. (about 10X nominal closed-loop operating values, but only on the negative cycles).

I meant, during the negative drive, the chip negative drive is still limited to the constant current source/internal drive resistor, but the 'idle-leg' is not. (collector Q5/Q7).
Q6/Q8 is off and Q5/Q7 is on by saturated Q1/Q2.
I did not mean the current going to chip, the current in collectors Q5/Q7 is uncontrolled during negative drive in open-loop I/V limited operation.

I hate to ask, but if you have the interest to review this, am I correct?
Did I miss anything?
Do you know anyone else who might be interested in pursuing this???

In my opinion, I'm thinking R23/R24 should be changed on all units, and would make the driverboard bullet-proof to all failure modes of the output chip, or operational states.

PS, do you have an internal diagram for the TA-100WA???
I could only find one for the TW-200W posted by EchoWars (IIRC), and two substitute circuits for both chips.

I used those for reference of the I/V limiting arrangement.
I was assuming the TA-100WA substitute was a proper representation of the internal arrangement.

Thanks for everybody's input!
[my terminology above may be incorrect]
Kim L

 

[srinath]

well an easy solution is to put in replacement transistors with more robust power handling capability. Replacing the originals with original transistors will still give you problems later on, Kenwoods design runs those transistors and most of them to at least 75% of their spec. When they get old and have been heated for so long they start to fail. (most are obsolete anyway)
29 out of 30 KA9100's I worked on had the 100 ohm resistors in them. For me to explain to the "customer" about this problem is futile. I rebuild the amp boards with better transistors etc and all is well with or without the 100 ohm resistor.
I agree all this theory is fine for people in the know.

Oh yea ... But what about the 30th Kenwood ... just kidding, too tempting to pass up especially on april 1.

Anyway I will make a purchase and stash it when the kenny gets back to my hands.

Cool.
Srinath.

 

[gort69]

Old thread for sure, but my $.02 -

I got this KA-9100 a few weeks ago in working condition. The R channel amp board had been worked on in the past - obvious replacement transistors @ Qe6,8, and Re24 had obviously been replaced as well, and likely the R side TA-100, as it was bronze in color vs the black one on the L channel. The L channel appeared to be untouched.

At some point during my overhaul Qe5 on the L channel failed, taking Re23 and Re25 along with it.. Re23 burned to the point where the color bands were gone. Checking the S/M, found the value listed as 1k ohms, so I installed a 1k ohm resistor there. Since my plan had been to recap and replace all transistors on the amp boards I went ahead and did that. (mistake - fix first, then overhaul) Did the same on the R channel and in the process noted that the replacement at Re24 was "wrong" being 100 ohms, so replaced that with 1k ohm as well. Yeah - I thought to myself - some dumbass stuck a 100 ohm resistor in there. I'll show him. Ha! Anyway, with modules out, fired it up through DBT, 100W bulb. Bulb dims, a few seconds later, CLICK! Cool, I thought. Checked voltage on pins 7 & 8 of amp boards, roughly 25mv on each. I can live with that for now.

Soldered both TA-100s back to the boards. Powered up through DBT, 100W bulb. Bulb dims, no click. WTF? Fearing the worst, check for DC voltage at pins 7 & 8 on the amp boards. What should be a few milivolts to my horror was -55Vish DC on BOTH channels. Searched around here on AK and found that big DC on those pins means a "likely" toasted module. How on earth does a shorted transistor @ Qe5 take out both TA-100s?

Consulted AK member rjsalvi via PM - who had gone through his own version of hell with his 9100 - and he encouraged me not to give up on mine, that the TA-100s are tougher than one might think, and helped me troubleshoot. Big public thanks to him!

Last night stumbled across this thread (which is largely over my head) - and it reminded me of that pesky 100/1k ohm resistor deal at Re23,24. So I took a shot replaced those 1k ohm resistors I had installed with 100 ohm, powered up through DBT. Bulb dims, and a few seconds later - CLICK! Plugged it into the wall - the thing works perfectly.

The only WAG stab-in-the-dark guess I can make is that the particular mix of transistors I used to replace the originals do not play well with a 1k ohm resistor @ Re23,24. So for future reference, here is what I used:

Qe1-4 9&10 KSC1845FTA

Qe5-8 KSA1220AYS

Qe11,12 KSC2690AYS

Over and out.

 

[rossm]

Cross Reference my thread on factory mods in this area
http://audiokarma.org/forums/index.php?threads/kenwood-ka-9100-factory-mod.839913/