Author Topic: Any amp geniuses here willing to give advice?  (Read 976 times)

Offline Rick

Any amp geniuses here willing to give advice?
« on: April 15, 2018, 08:38:23 PM »
I got hold of a Technics SU-8088K that I'm trying to repair.

I really like it and would love to get it working but its becoming a bigger challenge than I could imagine. I have repaired many amps and generally have a decent understanding but the circuitry in this thing is confusing me me they are using things like 2 diodes in series in single packages all over the place as well. Also, made in 1979 most of the transistors used are now obsolete, I have had to substitute best I can spending many hours immersed in data sheets.

I have the preamp section working perfectly now, its just the power amp, biasing seems to be an issue, after shorting out and wasting 8 X MJ15003/4 I have learned one lesson at least to power up though a 100W light bulb.

Below is one channel, bias is way too high, all damaged components have been replaced I switch on and it goes almost straight into thermal runaway, could be oscillation but I cant see anything on the scope.

I'm looking at how to modify and lower the bias.



Offline Rick

Re: Any amp geniuses here willing to give advice?
« Reply #1 on: April 15, 2018, 09:02:57 PM »
It might help if I post some specific questions about the schematic.

1) Q409 serves what purpose?

2) Q427 and Q429 are not a complimentary pair in fact they are even different packages, were the designers trying to compensate for something?

3) If I remove Z401 the output resistor pair (in a package) so it doesn't short, I can power up the amp feed it a sine wave and get clean signals on the base of Q423 and Q424. Put it back, and thermal runaway and boom the smoke leaks out, well used to, now the 100W light bulb goes on full.

4) Right channel totally distorted left channel plays, nothing is blown I even measured every single capacitor and resistor. After initial repairs I keep replacing only Q423 and Q425. Working backwards from Q419 and Q421 voltages are virtually identical on both left and right channels, I even swapped components between left and right on the SUP 15251 A board just in case (took ages) Right channel still distorted, its something on the SUP 15251 A board after Q401/3/5 and before the SUP 15251 B board I "suspect" .
« Last Edit: April 15, 2018, 09:06:42 PM by Rick »

Offline marantz123

  • Trade Count: (+1)
  • AVForums Veteran
  • ***
  • Posts: 283
  • Total likes: 3
  • I learnt more from The Internet when it was slower
Re: Any amp geniuses here willing to give advice?
« Reply #2 on: April 16, 2018, 02:38:52 AM »
Hi Rick

No expert or genius, but just a little bit of experience and some thought about this. Had a Technics Z200 that had similar symptoms wrt the distortion. Low and behold, I also tested everything, including transistor and diode junction resistances (very important), gain and junction type tests in diode mode..all fine. In 25 years, I have never had to change a small value ceramic cap...was told that they rarely go faulty..never had a bad one....until this  Z200 appeared.  I had a 10pF measuring 20nF....that was causing the distortion...i think it was even presenting itself as a short at Hz where it shouldnt be. I changed a few others out (of the same type)  as didnt trust them. Even the channel that was working ok sounded so much better with new ceramics...I was surprised.

If you havnt tested them, or even if you have, under the current circumstances where you have checked everything else, & still killing semiconductors, I would  consider replacing the ceramics to rule them out. They test ok, look ok, but when operating under Hz, or if they took a hit a some point, can be unreliable. (The ceramics in the AC power section should be ok though). Theres only a few on there, with C405 in a possibly important position, in the path of your 'output resistor pair'.  Also check that zobel capacitor, even though a larger value. NPO/COG or X7R /Z5U dielectrics are decent for the small values. Its a shot in the dark for me as I havnt worked on your model of amp as yet, but my circumstances were similar to an extent.

Good luck
Marantz123.


Offline ludo

Re: Any amp geniuses here willing to give advice?
« Reply #3 on: April 16, 2018, 02:48:33 AM »
Sub-genius input:

Q409 cascodes Q407. By having Q409 collector fed from the Voltage amp stage it must have a sort of feedback effect as well. This will be a minor effect though, as current changes in Q407 will be small. Could be worth simulating to see what exactly is going on there, especially at clipping.

Doesn't solve the fire though.

My suspicion is that the fire originates in D409 or D413. Both see only 0.7V less than the supply voltages for the output stage at idle. So if they leak into the bases of Q419/421 you can have over-bias in the output stage. Adjusting things in the bias string (D407 + R461/R459/R463) may not be enough to compensate for a leaky D409/D413

D409/D413 look like clamp diodes to me, so they should only switch on when the output approaches clipping against the supply rails, to keep Q419/Q421 & output devices out of saturation. R467/R473 provide a path to dump the current from the VAS to, during clipping. These resistors must also slow the VAS down and stabilise it a lot when D409/413 conduct during clipping, as they load the VAS quite aggressively to ground then. Very clever this, unless I'm missing the point ;D

If you temporarily unsolder and lift one leg of each of these clamp diodes you might have no further runaway. Then the culprit will have been found. I hope so...



When things misbehave, I always like to disable the over-current protection until the basics are sorted. (By removing D421/D423 here) This seems counter-intuitive but it serves a purpose when there is danger of oscillation. If there is oscillation (after swapping in "equivalent" devices) the over-current protection may also switch on and contribute to the madness. Without it one has more of a sporting chance of debugging the basic setup.

A rule of thumb for me is 100 Ohm base stoppers on the drivers (Q419/Q421) and 10 Ohm base stoppers on the output stage (Q423/Q425) The On-Semi's likely benefit from this more than the original Japanese parts did. (Seat of the pants opinion and subject to testing in situ...)

If it 's going vrot, put it on the buffet ! - Fats

Offline Rick

Re: Any amp geniuses here willing to give advice?
« Reply #4 on: April 16, 2018, 08:06:45 PM »
Thanks both for valuable advice much appreciated, noted, I will play detective now, and feed back with any success or discovery.

The amount of failed components I have already replaced, especially diodes in this thing defy belief, but have not done a few that measured "OK".

Cap story very possible, also I have seen transistors that pass the diode test but the gain curve is totally messed up. This amp is a beautifully built classic to me, but I am surprised at how many components in it were shot, then again I do not know its history and 1979 was a while ago (Just short of 40 years old it is). It does remind me of working on the two Marantz PM94 repeatedly exploding-for-nothing but everything measures ok with something that never fails has failed amps with separate boards all over the place.

There's a lesson in this for me to just go buy a variac now and stop avoiding. To those that know there is nothing worse than that sub one second loud transformer hum, a click and then smoke - flying at the off button anyway even though its too late and you know that short sound and strange smell is at least another R240 again, it adds up fast.

« Last Edit: April 16, 2018, 08:15:50 PM by Rick »

Offline JANNAS

Re: Any amp geniuses here willing to give advice?
« Reply #5 on: April 17, 2018, 07:36:53 AM »
Build yourself one of these, it current limits the load and will protect the circuit under test. I have build one on a variac and it works very good. I use it when testing transformers or any electrical thing that is suspect. No more blitzen sparken !

Offline Rick

Re: Any amp geniuses here willing to give advice?
« Reply #6 on: April 19, 2018, 06:12:27 PM »
Just some feedback and another little help call.

I have the amp up and running now finally! The effort has been worth it, using the MJE15032/15033 predrivers and MJ15003/4 outputs the thing packs much more punch than I ever expected thanks to the beefy power supply and the +/- 55V rails and nice power buss bars all over. I know some people don't like Onsemi but I love them, seldom can I break them they are rugged devices However not quite sorted with this amp yet.

Story - short version because the long one is just way too long but I measured diodes all checked out, I then, not trusting went and measured the forward voltages powered up (Some double diodes in the design or at least back then you could select your diodes with forward voltage drops from 0.4 to 1.8V)). All of them were well out of spec with voltage and 100 milliamp across them most around 1.6V and you just cant buy such a range anymore. I figured replacing the entire lot with 1N4148 may work better than what was in there already and indeed it did. Posting the LHS channel below with voltages from the service manual. Note that D408 is screwed onto the heatsink as a sense diode and its a 1.2V (Double?) Diode, there is not data sheet for these but both were tested well and seem just fine so keeping those two because they are in a special package that attaches to the heatsink with a spring clip. To anybody else experiencing a similar vintage problem, note that Matsushita electronics in the late 70's made some very bizarre and NOT lasting diodes, I have replaced every diode in this amp except 2 that were not made by them. They might measure ok in circuit but power them up and they are totally off spec, not something one would expect.

Q1: I am now sitting with it working, however to keep the bias in check VR459 is shorted and VR460 has a 33 Ohm resistor across it, this is not enough, so the bias string needs to be changed, question is how where? I can short the sense diode and the bias drops so its working just needs modification

Q2: As much as I stare at this thing I have hit a blank..... what resistances must go up or down to lower the bias?

Q3 Q428/430 are still not a complimentary pair by design, this bothers me I don't know why they did that, it sounds fine with a sine wave swinging equally on the scope, so will probably just leave that bit alone then?



A little trick I tried after I stood on my 100W light bulb late at night was to remove the 5Watt output emitter resistors and throw in 1/4 watt 1 ohm metal film. If there is a problem with the output transistors or bias it will smoke them happily without anything else getting damaged and if its working you can even play music up to about 7 watt and listen. if they burn who cares, they 3 cents each, chuck in another set and start over.
« Last Edit: April 19, 2018, 06:19:44 PM by Rick »

Online Steerpike

Re: Any amp geniuses here willing to give advice?
« Reply #7 on: April 19, 2018, 07:35:16 PM »
Q428 and Q430 are only there to protect the amp from output short circuits (or very severe loads). During normal operation they don't even conduct, so their usual characteristics are very unimportant, and they don't need to be matched in the conventional sense.

(I'm thinking this clamping topology was the invention of Harold Black, but I might be attributing it to the wrong designer)

Offline Rick

Re: Any amp geniuses here willing to give advice?
« Reply #8 on: April 19, 2018, 09:48:17 PM »
Q428 and Q430 are only there to protect the amp from output short circuits (or very severe loads). During normal operation they don't even conduct, so their usual characteristics are very unimportant, and they don't need to be matched in the conventional sense.

(I'm thinking this clamping topology was the invention of Harold Black, but I might be attributing it to the wrong designer)

Thanks I see now, wasn't looking properly and making assumptions they were another two. In the previous posts there was a lot of curiosity about the clipping performance. When I play it even winding the volume to full it exhibits very well behaved clipping very little harshness, it enters very gently into it and it doesn't get worse it just stays at a sort of "stable max" if you try push it further, so that is interesting, will take some scope trace pictures approaching and at at clip and post here once the bias is fully sorted.

Also thank you all for the input, I'm used to fixing more conventional modern amps, you have helped this one back to life and Im really chuffed about that it was really a mess. At least the  Faceplate is very clean and I'm re-doing the cool looking explanatory graphics Technics put on the covers of these products of the same era. I am having them scanned any missing stuff from rust in the scan fixed and then silkscreened back after powder coating. Will be part of a nice collection and matches other items I have.

Now just for that Bias string fiddling.
« Last Edit: April 19, 2018, 10:02:47 PM by Rick »

Offline ludo

Re: Any amp geniuses here willing to give advice?
« Reply #9 on: April 19, 2018, 11:21:10 PM »
...
Q1: I am now sitting with it working, however to keep the bias in check VR459 is shorted and VR460 has a 33 Ohm resistor across it, this is not enough, so the bias string needs to be changed, question is how where? I can short the sense diode and the bias drops so its working just needs modification

Q2: As much as I stare at this thing I have hit a blank..... what resistances must go up or down to lower the bias?
...



If using a pair of 1N4148s in place of D408 you could place a resistor in parallel to one of the pair. 100 Ohm or so could be a good start. You can still trim bias on R462 then.

1N4007 used at low current (~9mA here) might already  give lower Vf than 1N4148 and be easier to insulate and mount?

You might consider removing that whole bias string and R464, making an adjustable single transistor bias generator on a small PCB (strip-board even) to replace it, and mounting that where D408 was on the heat sink. (Ideally it might go right on the mounting tab of one of the output devices.) Won't look original though.


If it 's going vrot, put it on the buffet ! - Fats

Offline Rick

Re: Any amp geniuses here willing to give advice?
« Reply #10 on: August 11, 2018, 09:37:20 AM »
I got this beast going, I used the iN4148 and it worked I can adjust the bias. Have to the old diodes were terrible quality and had drifted way way out of spec.

Offline El Sid

Re: Any amp geniuses here willing to give advice?
« Reply #11 on: August 11, 2018, 11:03:58 AM »
Well done Rick! Just looking at that circuit gave me the heebie-jeebies.  :clap: