Interpreting data from Oscope

[oldman55]

I know next to nothing about interpreting and repairing irregularities found. Here are some traces on generated sine and square waves from a heavily modified Eico ST70 (all tronola mods plus loudness mod).
Sine wave, low volume and to distortion:


Square wave, low volume and to distortion:


oldman

 

[cpt_paranoia]

It would help if you explain what each of the traces are, and where they are measured, and into what loads.

Is yellow your input signal, and magenta your output? If so, your generator is poor at high output.

 

[gadget73]

Output levels would also be good to know. If you're actually at 20v output, thats too high for square wave testing. I'd keep that to about 2v p-p to avoid saturating the transformer and getting misleading results.

 

[FileFixer]

Is it possible to change Coupling to AC and set division to Fine instead of Coarse?

 

[oldman55]

Each color is a different channel. Input signal introduced at line level input and measured at the 8 ohm speaker terminals with some wire wound 8 ohm 100W resistors on a heatsink for dummy load.
Will change coupling to AC and to 2V.

 

[oldman55]

sine and square:


Strangely when I put in "reverse mode", got:


I can adjust balance to bring to same output value but this is what happens to just switching the mode. Even stranger things happen when I switch "speaker phase"

 

[gadget73]

I see signal imbalance, and if you're getting square out of one and sine from the other with the same input, something is definitely wrong. If you're inputting square and getting sine out, thats also telling me something is wrong.

Speaker phase might cause odd results if it switches it at the speaker output. With a common ground you'll likely be shorting the output when its in one position so the output will be very low and/or distorted. I'd probably run the scope - lead to chassis and use the + leads on the 8 ohm taps, bypassing the phase reverse stuff entirely just to rule out issues from that circuit.

 

[triode17]

Watch out for speaker "phase" switches. They are in the high current, speaker output lines. If they get worn or dirty, they can cause all sorts of abnormalities. There's no need for them anymore because we all know how to set the proper speaker polarities. So I eliminate them from the circuit.

 

[oldman55]

OK, operator error. With both generated signals now the same on both channels, the following are the traces I am getting at the dummy load for sine and square;

square with loudness switch on:

At distortion:

Tolerable? Sounds good once balanced (which is a whole other issue). Also trying to calc power at distortion.

oldman trying to catch up

 

[petehall347]

am trying to follow this ... but finding it hard ..
sine wave looks rather squarish .

 

[gadget73]

Channel 1 looks to have significant high frequency roll-off based on the square wave output with loud off. Channel 2 has a bit of overshoot, but its probably typical.

Loud on looks opposite.

Since the imbalance gets a lot worse with the loudness on, I think I'd be looking at the parts in that circuit and the tone circuit in general.


The traces at distortion show some crossover distortion (that "step" where it crosses zero) which usually indicates the bias level isn't quite right. Signal level is also not equal. Might be worth swapping tubes channel to channel to see if that changes.

You can also check output right at the volume control to see if this is being caused in the preamp section or the power amp section. I'd expect the power amp when fed directly, no tone controls and such in the circuit, to produce nice sharp square waves at 1khz with equal output levels.

 

[cpt_paranoia]

I'd expect the power amp when fed directly, no tone controls and such in the circuit, to produce nice sharp square waves at 1khz with equal output levels.

OP appears to be feeding a test signal of 10kHz, which is rather high for a square wave audio test. 1kHz would probably be better.

I agree about the crossover distortion. The sine wave (at least the yellow one) looks to have some asymmetry; quadrants 1 & 3 are steeper than quadrants 2&4, so the trace looks 'pushed over'.

L/R channel imbalance is clear (assuming yellow is L, and magenta is R (or vice versa)).

I'm still not entirely sure what each of the traces is...

 

[oldman55]

Thanks for input.
The output of the power tubes are all set to 34ma. The tone section has been modified so I wonder if that would do it.
Will take above advice and retest.
oldman