Pioneer VSX-407RDS amplifier repair

[npnlamp]

Hello, I got this amp from a friend, and tried to repair it. The problem is that there's about -8V DC on the output. Here's the manual : https://www.manualslib.com/manual/1088125/Pioneer-Vsx-607rds.html#manual

I applied 1kHz signal to the input and tracked it. It's all fine until it reaches IC401 (STK4211II). Here the problems start. Input signal measured from AGND is fine, but when measured from GND(RET) which is main IC ground, it is biased to -8V. There is 8V potential difference between AGND and GND(RET).

I built a circuit to test both STK4211II(PAC009a) IC's, they are both working, so the issue is somewhere else. All power rails are OK. Again with an exception that there's 8V difference between grounds.

One thing that I don't understand is, why are signal ground and power supply grounds not connected in any way?

The problem exists on both channels, so I think there must be a mutual cause. I checked the transistors and diodes around the STK IC's, all of them seems good. Where else could the problem lie? Any advice on what else I should check?

Thank you for your help!

[mzacharias]

There is an "IC Protector" fuse in the ground path, given as "IC403". Have you checked that?

[npnlamp]

I checked it, it's good.

[mzacharias]

I do think the grounds should be at one potential. I have seen a faulty eyelet at the power transformer circuit board cause similar issues. (Lack of a center-tap connection)

[4thDoctorWhoFan]

I do think the grounds should be at one potential. I have seen a faulty eyelet at the power transformer circuit board cause similar issues. (Lack of a center-tap connection)

I agree.
Short the two grounds together and see what happens.

[npnlamp]

Thank you for helping guys, I really appreciate that

I do think the grounds should be at one potential. I have seen a faulty eyelet at the power transformer circuit board cause similar issues. (Lack of a center-tap connection)

Checked that, all connections and transformer are perfect.

Short the two grounds together and see what happens.

I tried shorting the grounds, amplifier restarts for a brief moment but then everything works fine. The current between grounds is 2.5mA when shorted. Although I wouldn't trust on this method as a solution.

I also noticed that if I disconnect DC detect resistor R435 (page 12), the voltage difference jumps from -8V to -35V. It must be that parts R508, D505, D506 (page 16) somehow equalize voltage difference a little bit.

What else could I check?

I attached few oscilloscope measurements of the output, when grounds are shorted, and when there's a voltage difference.

[4thDoctorWhoFan]

I tried shorting the grounds, amplifier restarts for a brief moment but then everything works fine.

Then what's the problem? 
As you said, Amp restarts and work fine.
Seems like problem solved.  You never know what the amp has been through before you got it.  Somebody could have forgotten to re-attach a jumper or wire which connected the two grounds together.

[npnlamp]

Well in the schematics grounds are not shorted so I don't think this is a nice solution, who knows what kind of weird ground loop this is making, and what issues it can cause, noise etc...

[4thDoctorWhoFan]

Well in the schematics grounds are not shorted so I don't think this is a nice solution, who knows what kind of weird ground loop this is making, and what issues it can cause, noise etc...

That's your prerogative but schematics are not always correct/up-to-date/latest revision etc.  Sometimes, they don't even show the connection on the schematics.

I had a very similar problem with an old JVC amplifier and after shorting the grounds together, it has been working for years and sounds great.  I still use it to this day.

[npnlamp]

Thanks for the tip, I will probably end up doing this if I can't figure out a different way

[npnlamp]

Well it turns out it doesn't work with grounds shorted. At first I made measurements without any load on the channel so I thought everything was working. But if I put 12ohm load on the output, without shorting grounds, biased sine wave turns to zero, and with shorted grounds amplitude decreases greatly and negative half of the waveform disappears.

I attached pictures of output without load with grounds connected, and with 12ohm load with grounds connected.

EDIT: Turns out by playing around I killed the main IC. Installed a new one, problem persists, although it works the same with load now (Outputs nice sine wave, but DC biased.). Also if I try to short grounds now, display shows "power off" blinking, and amp doesn't work.

[mzacharias]

I wish I had this thing on my bench. What concerns me is that the unit is not in protect mode, even with an 8 volt offset or when the 1/2 waveform at the output - this should also trigger the protect.

This suggests to me that the ground which is referenced for the protect circuit (and so the larger microprocessor area itself) is no longer being referenced to the power supply (and amplifier) ground.

SOMETHING is floating.

So you have been on the right track with the ground issue. I would be tearing it apart looking for a possible burned foil run, maybe a mechanical connection to the rear panel (like at the input jacks) or something like that. And although you said the ground circuit fuse I mentioned was OK, do not fail to check every fuse and pico-fuse you can find.