Onkyo TX-4500: distortion driving me nuts

[Bert Beukema]

Hi all,

I was gifted this TX-4500. It had an apocalypse in the power amplifier, but is otherwise pristine. I rebuilt the power amp stage (done several 4500's before). Full rebuild with quality components. Power supplu was done at the same time, including a new relay. Problem is, there is major distortion on the right channel. And for the love of me: I have absolutey no clue what is causing it. I must have changed all components at least twice and have even reinstalled old semiconductors. All from Mouser, no Fleabay junk. Nothing helps.

1. Not the pre-amp.
I have separated the power amplifier and driven it with a brand new pre-amp.
Problem persists.

2. Distortion never changes channels
I have routed the speaker out "right" to the "left" and vice versa.
No effect.
Which is bloody odd because I am dropping the signal into a different channel.
I would exect it to swap channels.

3. DIstortion --> trimpot R604
Moving this trimpot (it is brand new multiturn) DOES lower the distortion.
Opening / closing it has a large audible effect on the distortion.
But ALL components in the R604 circuit are just fine.
I believe this is the idle current trimpot.

4. Bypassed the relay / power supply
Also no effect, distortion remains.

5. Distortion not on headphones
The distortion isn't present on headphones (or not as noticable).
It only appears when I activate the speakers. Then the headphone audio also distorts.
My audio tracer confirms this.
I would normally suspect the outputs but these were disconnected.
And also brand new high power transistors.
Connecting the power transistors changes nothing.
I even built a new output stage (4 power transistors), no effect.
The driver stage is MJE150030/31 and has enough power to drive the audio.

So I just don't know what else to do.
I feel that I am having tunnelvision here.
Voltages are all way off from the service manual, but this could be related to 230v and a state where a load is or isn't present. Makes measuring the voltages for the right values rather difficult.

I could really use some fresh ideas, help is much appreciated.
Too nice a machine to give up on.

Thanks!

Bert

 

[hamrules]

R604 is DC offset for the right channel.

 

[gadget73]

whats it look like on the scope ? Does it bias normally ?

 

[hamrules]

Item 2: Are you saying you took the actual speaker out and connected the right to your left speaker and the left to your right speaker ?

 

[Bert Beukema]

R604 is DC offset for the right channel.

R604 is DC offset for the right channel.

Yes it is, question is why is it influencing the amount of distortion?
At the moment I can't change any of these trimmers (504 / 604) using the correct method.
Values just don't change. Doesn't matter what trimmer I solder in.
Nor is there a change when i reinsert the old semiconductors.

BB

 

[Bert Beukema]

whats it look like on the scope ? Does it bias normally ?

When it rains it pours, my trusty Philips scope died last week. It is being serviced as we speak.
And no, it doesn't bias normally no matter what I try.
Not even with the original components reinstalled in that channel.
Neither channel will allow me to change any value of any trimpot.
I've tried many different approaches but perhaps tunnelvision is at play here.

Might be something upstream causing a problem.

BB

 

[Bert Beukema]

Item 2: Are you saying you took the actual speaker out and connected the right to your left speaker and the left to your right speaker ?

No. I simply soldered the wire from the left out channel to the right out channel terminal and vice versa.
Which did not change anything, distortion is still on the right channel. It should have moved to the left channel if the amplifier was the cause.
Then I bypassed the relay circuit by soldering a speaker (wire) directly to the right output terminal (and ground).
This also did not do anything.

BB

 

[Bert Beukema]

for the record, these are the changes made in semiconductors:

2SC1681 --> 2SC2240
2SA706 --> 2SA1220 (tried the old A706 but no change anywhere)
2SC1625 --> MJE15031
2SA815 --> MJE15030

These should all work normally, i've used them often.

BB

 

[Bert Beukema]

Oh yeah, there are also the mystery trimpots in the middle of the PCB.
These are labelled R537 and R537.
Unclear what these are doing nor whether or not they could impact the circuit.

BB

 

[Bert Beukema]

Final observation, Q506 is getting very hot.
This is an MJE15030. The other one in the left channel is just lukewarm.
It appears a load of current is drawn here. Question is why.
That transistor was changed already (just to be safe), same issue.

BB

 

[gadget73]

ok, so no bias and audible distortion sounds like crossover distortion issues. Have to figure out why it won't bias, and I'd wager that will sort it out. Check Q504 and the stuff around it, if something is funny there it will likely affect the bias and since it is connected to Q506 it may be causing both issues

wild guess but possibly the offset adjustment is adjusting things far enough that its providing some bias for at least one side of the output pair. If so that should cut down on the distortion a bit at low levels at least. Use that to zero the offset and see what you get from that point.

 

[sregor]

5. Distortion not on headphones

n, Q506 is getting very hot.
This is an MJE15030

Both of these symptoms suggest power or signal is not getting to one of the output transistors (Q509)
Could be bad resistor to base of ouptut,. Could be bad connections - seen it with TO-3 sockets. Any one, although collector connection is first concern.

 

[Hajidub]

No. I simply soldered the wire from the left out channel to the right out channel terminal and vice versa.
Which did not change anything, distortion is still on the right channel. It should have moved to the left channel if the amplifier was the cause.
Then I bypassed the relay circuit by soldering a speaker (wire) directly to the right output terminal (and ground).
This also did not do anything.

BB

If the problem didn't follow the internal rewiring the only thing it could be would be a bad internal ground connection. The swap of positive's routes to S814a and 814b, nothing in common (pun intended).

 

[hamrules]

So the distortion stays in the right speaker I take it from what you said. If so I’d double check that speaker or wiring to it is OK. Hook the left speaker to the right side and see if it happens or not.

 

[krautmaster]

check your solder joints - when I recapped my M-508 I made a mistake and bridged an output to the positive rail, which caused no sound and no bias (I also had one leg of a transistor not making full contact with the socket.)

 

[Bert Beukema]

So the distortion stays in the right speaker I take it from what you said. If so I’d double check that speaker or wiring to it is OK. Hook the left speaker to the right side and see if it happens or not.

Speakers are fine and i've tried several of my bench speakers. All the same results.
:-(

 

[Bert Beukema]

check your solder joints - when I recapped my M-508 I made a mistake and bridged an output to the positive rail, which caused no sound and no bias (I also had one leg of a transistor not making full contact with the socket.)

Good advice and i've resoldered and reflowed the entire board.
I've checked the board soldering under my bench microscope and it appears to be good.
Of course i'm not eagle eye and even with a microscope it's easy to miss a tiny sliver.

 

[Bert Beukema]

If the problem didn't follow the internal rewiring the only thing it could be would be a bad internal ground connection. The swap of positive's routes to S814a and 814b, nothing in common (pun intended).

Hrmmmm. There are only two ground wires and they are soldered on their terminals.
They lead to a common chassis ground.
Can't see anything wrong there :-(

 

[Bert Beukema]

Both of these symptoms suggest power or signal is not getting to one of the output transistors (Q509)
Could be bad resistor to base of ouptut,. Could be bad connections - seen it with TO-3 sockets. Any one, although collector connection is first concern.

I' ve disconnected the output transistors for the moment. Working with a lump of aluminium is crap.
But you might be right, not sure.
I built a secondary output transistor set to test the situation. Still the same problem.
And the 'original' TO3 output transistor set of course shows the same issue.

Thing is, with or without the output transistors connected the problem persists,
But Q605 driver getting so damn hot is not a good sign at all.

 

[Mico]

Are you referring to these two mysterious pots? I used to know what they did, will look into my notes.