Help With Accuphase E-203

RWood

Slowly Speeding
Subscriber
Hoping to repair this for a friend. The problem seems to be in one channel of the power amplifier, as it presents when the "separate" switch is engaged and a signal goes directly into the power amps.

Dim bulb is fine, main relay clicks "on", there is good audio through one channel and bad audio through the other. By bad I mean broken up, intermittent signal that pushes the cone of my bench speaker, irregularly. Looks like DC on the output, but not in a continuous way.

There are two sets of six pins easily accessible from the top (which will look like the bottom - this amp layout is upside down) , so I have included my voltage readings there, on the pics below. One pin's fluctuating DC looks suspiciously related to the bad audio in one channel.

I have a good schematic thanks to HiFi Engine. It's too big to upload here but it can be viewed here. I have not been able to secure a service manual so I don't have the DC offset or the bias instructions. If anyone can divine them from the schematic I would appreciate it. Two trim pots are easy to access but I don't see the TPs or the target voltages.

Any suggestions on where to look next to track down this problem, I would appreciate it. It looks like the power amp uses three ICs instead of discrete circuits.

Here are the pics:








.


 
The output stage uses an STK IC for the input and voltage gain stage. There is an OP amp for the DC servo as well as regulators Q1 and Q2 (2SD669A and 2SB649A) for the IC. So not a whole lot to deal with.

I'd double check the solder joints. Assuming those are good, check the servo op-amp and caps. After that you are left with the STK IC being on the fritz or possibly a wonky output mosfet.
 
AFAIK this voltage amplifiers are the same as used in the marantz pm80 and the sansui B1000 and other amps. If not cooled properly they all fail over time, they are not reliable.
If you do have cold spray, spray it on the metal back of the one of the failed channel and if the bad sound "changes" right away then your module is dead.
Not to worry they are cheap (although all those sold as originals are fakes, too).
 
AFAIK this voltage amplifiers are the same as used in the marantz pm80 and the sansui B1000 and other amps. If not cooled properly they all fail over time, they are not reliable.
If you do have cold spray, spray it on the metal back of the one of the failed channel and if the bad sound "changes" right away then your module is dead.
Not to worry they are cheap (although all those sold as originals are fakes, too).

Thanks. Is the the notorious STK3082? It looks like this amp uses one of them, half for each channel. I can certainly spray it and see if there is a change.

FootFungus, thanks for the advice. The opamps you are referring to, are they the NJM4559D? There do appear to be two of those on the board.
 
The servo op amps are the NJM4559D. If the integrator caps for the servo have issues, then that can cause problems too.

Like gslikker said, it is the notorious STK3082.
 
... If not cooled properly they all fail over time, they are not reliable.

+1
The STK voltage amplifier ICs were often installed without a heatsink as per the Sanyo application note, but they really do need an external heatsink. Without one, its not a question of "if" they will fail, but rather "when". The widely fluctuating DC offset is a common failure symptom.
 
...and indeed this one is just flying solo:

95EB6524-02EC-4F2E-A448-F6F1261CE528_zps9wjqdwxy.jpg


3E1399D1-5FF6-4BD2-8135-97267700D616_zpseirze9kk.jpg
 
Which ones are they? I see a bunch of caps around the NJM4559Ds.

I'm guessing the orange ones which are probably low leakage and/or bi-polar (schematic says bipolar). The other black ones are probably 10uf/50v which provide some power filtering for the op-amps.

My unit had a heatsink (aftermarket?) on the STK3082.

What does the back of the board look like?
 
Actually R43 in your picture looks like it might be toast which would indicate an issue with the drivers/outputs.
 
It looks just like this. STK has been removed:




Actually R43 in your picture looks like it might be toast which would indicate an issue with the drivers/outputs.

In-circuit, both R43 and R44 read about 5x what the schem says their values are. Instead of 100Ω they read 381 and 386Ω respectively.

Thanks for noticing that.
 
Last edited:
In-circuit, both R43 and R44 read about 5x what the schem says their values are. Instead of 100Ω they read 381 and 386Ω respectively.

You might want to increase the power rating of the replacement resistors to 0.5W or higher.
 
Thanks Ben; will do.

I am about to pull the trigger on what looks like an original STK3082. US supplier with a return policy so I'll give it a shot.

RWood
 
The STK might not be at fault. I'd hold off until you figure out what made those resistors go toast. I'd pull those outputs for that channel and check them with a multimeter.
 
Thanks for the suggestion. I made a first pass at testing the outputs and got some strange results. I need to go back and take another pass. I have pulled the devices and am using the diode check function on my meter.

This amp uses two P-channel and two N-channel mosfets . What was concerning was that on two similar units ( I forget which it was) I got differing results. Is there anything unusual about testing these kind of transistors?

Adding to the fun, the resistor that had overheated, and its mate from the other channel, are actually 390Ω 5% resistors, not 100Ω as in the schem. The one from the good side clearly showed the color bands Orange-White Brown-Gold, and there was enough visible on the burned one to confirm this. So they actually measure well within tolerance, although were underspec'd for an amp that was probably not biased right.

On that count, I have emailed several Accuphase resources in the US, GB and DE and cannot secure a service manual. As a company they seen overly protective of their support materials. Maybe they want their busted gear sent to Authorized Service Centers.......not good for those of us DIY'ers and people like me who view this as a learning opportunity. Oh well. Maybe with some help from you guys I can figure out how to adjust VR1 - 4 based on the few voltages that are listed on the schematic.
 
Each channel uses 1 N and 1 P channel device. The source and drain pins are swapped on the Hitachi laterals. So the case is actually the source instead of the drain. So long as you aren't getting short on the diode check between different pins you probably are ok.
 
It's been a while but I got back into this amp over the weekend.

I replaced all four outputs with some original Hitachi units. I have not yet replaced the op-amps or the caps on the power amp board, but I have put an order together and will pull the triigger soon. Before I do, I want to give an update on how it is behaving.

Here are the symptoms now. The amp comes out of protection after about 15-18 seconds. DBT with a 60W bulb shows moderate current draw at first but it calms down. Preamp does not pass any audio signal, not only to the internal power amp but also when run in "separate" mode and pre-out is taken to a different amp. This is a new phenomenon and I am a bit alarmed by this.

Power amp does work when an external signal goes directly into the Power Amp input but only a well-attenuated signal (I used a tape deck with an output level control) will pass. Otherwise, a moderate or high level signal will light up the bulb and the protection relay quickly kicks in to shut things down. With a quiet signal, though, the amp sounds good and it works on both channels.

Do these symptoms, particularly the power amp, sound like things that will fall in line once the power amp board is updated? Any ideas about the preamp?

Thanks (again)!
 
Hello..... another update and a couple questions, if I may.

First, the amp is working and sounding very nice. After recapping the electrolytics on the power amp board and replacing the two op-amps it all came together. Many thanks to those who replied, especially FootFungus - I appreciate your guidance! :thmbsp:

Now that it's working, I want to make sure that I didn't just fix the symptom and leave the cause. What I notice is that of the four output devices, Q9 is running noticeably hotter than the other three. My meter shows it at 60°C where the others are running around 44°C.

Perhaps not coincidentally, it was R43 that FootFungus noticed had been overheated - - that is connected directly to the gate of the hot Q9. I have replaced R43 with a 1/2W metal film. (Btw, the value of the unit in the amp was 390Ω rather than 100Ω as drawn on the schem, so that is what I used. I replaced R44 as well with a matched 390Ω).

I adjusted VR1 and VR2 to get 22mA as indicated on the schematic:

203-ePowerAmp_zpsb6bee5a3.jpg


I also see VR3 and VR4, coming off pins 5 and 3, respectively, of each op-amp, but don't know how they should be adjusted (and they are not accessible when the board is installed and connected to the power supply.) Any help with this?

Ideas on the hot Q9? I see a max channel temp listed on the Data Sheet of 150˚C so maybe I shouldn't be concerned.

Thanks,

RWood
 
Back
Top Bottom