Continue testing with 1 kHz sine wave input and with loudspeaker connected. Put a 470 μF - 1000 μF capacitor in series with the loudspeaker to protect the loudspeaker. Keep the VOLUME control low to show moderate distortion.

Show the oscilloscope signals at the base and emitter of Q504 and Q506.
Test at resistors R510, R512, R514 if these are safer to reach.

Were Q504 and Q506 ever replaced?

Q504 2SC3944
Q506 2SA1535

With the power off,
check resistors R516, R518, R566, R568 in-circuit (all 2.2 Ω)
check resistors R548, R550, R570, R572 in-circuit (all 0.22 Ω)

 

No, Q504 and Q506 were not replaced. I removed them from the PCB and tested them with a component tester, and they passed. the tester is below,

However, since the test current on these multi-function component testers is extremely low, could be sometimes give a false positive under actual circuit conditions, so I will try to replace if I have spares that can be fit.
Update will come.
Thanks

 

Continue testing with 1 kHz sine wave input and with loudspeaker connected. Put a 470 μF - 1000 μF capacitor in series with the loudspeaker to protect the loudspeaker. Keep the VOLUME control low to show moderate distortion.

Show the oscilloscope signals at the base and emitter of Q504 and Q506.
Test at resistors R510, R512, R514 if these are safer to reach.

Were Q504 and Q506 ever replaced?

Q504 2SC3944
Q506 2SA1535

With the power off,
check resistors R516, R518, R566, R568 in-circuit (all 2.2 Ω)
check resistors R548, R550, R570, R572 in-circuit (all 0.22 Ω)

You hit the nail on the head!

The resistors themselves were fine, but after inspecting the PCB under a magnifying glass, I found both R514 cracked pads/traces in that section. I repaired the traces and reinstalled the board.

I powered the unit through the Dim Bulb first, and everything looked stable. I then switched to direct mains power and injected the 1 kHz sine wave into the input.

When checking the speaker output on the oscilloscope, I noticed a strong high-frequency oscillation superimposed on the signal, making it impossible to read the 1 kHz sine wave cleanly.

Could this high-frequency oscillation be caused by the replacement transistors for Q428 and Q430? The ones currently installed have a lower Transition Frequency ($f_T = 100\text{ MHz}$) compared to the original parts ($f_T = 200\text{ MHz}$).

Update on Audio Test:

I tried connecting a speaker and playing music/audio. To my surprise, the sound coming from the speaker is completely clean to the ear! I monitored the temperatures for a while: nothing is overheating (drivers and power transistors remain cool), and there is no sign of thermal instability or smoke.

However, the thick high-frequency trace is still present on the oscilloscope.

Thanks!

 

Try to use your oscilloscope to find where the oscillation originates.
Is it before Q502 or after Q502?
C506, 0.1 μF would have suppressed oscillation.

 

Try to use your oscilloscope to find where the oscillation originates.
Is it before Q502 or after Q502?
C506, 0.1 μF would have suppressed oscillation.

I checked Q502 with the scope, and the high-frequency oscillation is present on both the base and the emitter.

A key detail: the oscillation only triggers when the speaker switch is engaged (connecting the load). Once triggered, switching the speaker back OFF does not stop the oscillation—it remains active, though its peak-to-peak amplitude decreases slightly.

Looking at Q428 and Q430:

• Oscillation is present on both transistors, but the amplitude is significantly higher on their emitters.

On IC402:

• Pin 5 remains completely clean.
• Pin 6 and Pin 7 both show high-frequency oscillation.

All these tests were performed with no input signal and the volume set to minimum (0).

Thanks!

 

There is negative feedback from the power stage output to IC402 pin-6.

You need to check C508, C518 (0.068 μF) and R520 (I cannot read the value, 10 Ω 1/2W ?)

Try temporarily, 0.01-0.1 μF from R570-R572 node to GND.

 

Hi,

Sorry for the long silence—family comes first, and I prefer not to do any testing when I'm tired (too high a risk of letting something go up in smoke!).

I finally performed the test by adding a 68 nF capacitor from the R570/R572 node to GND.


A strange behavior started: high-frequency oscillation appeared on the left channel as well.



I also checked the components:

• R520: As you suggested, it is listed as a 10 Ω, 1/2 W resistor on the parts list and tests fine. However, while it is the bundled component, visually it looks more like a 1/4 W resistor rather than a 1/2 W one.
• Capacitors: In-circuit measurements match the left channel. C518 (and C517) reads around 128 nF (left channel is 132 nF), and C517/C518 show consistent values (about 2 nF)

Thanks!
Alberto

 

Hi,
A small update,
I tested C508, C518 (0.068 μF) and R520 after removing from board.
Replaced Q4228 and Q430 with original ones, no change, HF signal still there.

Thanks

 

As a temporary test, try putting a 0.1 μF from C508 (R570, R572, L502 node) to GND.
See if the frequency changes.

 

As a temporary test, try putting a 0.1 μF from C508 (R570, R572, L502 node) to GND.
See if the frequency changes.

I performed a test using two 68nF capacitors in parallel. I soldered one end to the jumper going to the node formed by R570 / R572 / L502 / C508, and connected the other end directly to power supply ground, not chassis, is that correct?
(Please let me know if this connection point is correct or if its exact physical location on the PCB could affect stability).

Interestingly, the oscillation behavior is strongly influenced by the speaker selector switch. Since the load is connected to Speaker Output "A", I expected no signal or load interaction when selecting "B" or "OFF". Instead, all four switch positions yield different results on the scope:

• Speaker OFF: The HF oscillation is present, but at a reduced amplitude
  
• Speaker A (Loaded): The oscillation triggers at full amplitude.
  
• Speaker A+B: Each position modifies the waveform/frequency of the oscillation differently, even with no speaker
  

Speaker B (No speaker connected)


Thanks!

 

What do you mean by Speaker A, Speaker B, and Speaker A+B?
Is that referring to the Left and Right channel speakers?

 

What do you mean by Speaker A, Speaker B, and Speaker A+B?
Is that referring to the Left and Right channel speakers?

No, all test have been done on R channel,
I mean moving this switch:




Could it be a waste of time checking these capacitors?


Thanks!

 

Yes. I think it would be a waste of time to test capacitors.
A better test would be to add about 0.1 μF from signal to GND along the signal path and see how it affects the frequency of the HF oscillation.
I think this is related to feedback. Sorry, I cannot put a finger on this at the moment (meaning, I don't have a clue).

 

Yes. I think it would be a waste of time to test capacitors.
A better test would be to add about 0.1 μF from signal to GND along the signal path and see how it affects the frequency of the HF oscillation.
I think this is related to feedback. Sorry, I cannot put a finger on this at the moment (meaning, I don't have a clue).

Thank you so much for all your time and help! I would never have gotten this far on my own.

Regarding the test with the 0.1 μF capacitor:

Does the lead length affect the test results? In other words, can I use test leads/alligator clips, or should I solder the capacitor as close as possible directly between the nodes and ground?

Also, is the amplifier safe to use in its current state, or could this high-frequency oscillation be dangerous to the amplifier itself or the connected speakers?

Thanks again for everything!

Alberto

 

Lead length can affect the test result. Do not use test leads/alligator clips. Keep the leads on the capacitor to less than 1-2 cm.

I cannot say with certainty but 1 Vrms at 400 kHz should not be a problem.
I would prefer to find the source of it and eliminate it. Something is oscillating where it should not.