nfb issue

[catalin gramada]

Hello.I need a little help to debug a nfb issue into a marshall jtm clone. I put some traces of my problem, The sine is pretty neat without nfb conected.I really don.t know wher to start.What do you think ,please?Thanks.Catalin





There are only 9 db feedback into circuit 27k series with 5k pot shunt with 100nF between wiper and ground. The last pic shows how the trace looks like with presence on "0"
Thanks

[nickb]

Hello.I need a little help to debug a nfb issue into a marshall jtm clone. I put some traces of my problem, The sine is pretty neat without nfb conected.I really don.t know wher to start.What do you think ,please?Thanks.Catalin

[ATTACH=CONFIG]37395[/ATTACH]
[ATTACH=CONFIG]37396[/ATTACH]
[ATTACH=CONFIG]37397[/ATTACH]

There are only 9 db feedback into circuit 27k series with 5k pot shunt with 100nF between wiper and ground. The last pic shows how the trace looks like with shunt resistor in parallel with 100n (presence at 0) first and second with presence on 10
Thanks

What is the frequency and voltage scale?

I wonder if it might be marginally stable. Can you restest with a 1kHz square wave and post those images?

[J M Fahey]

Hello.I need a little help to debug a nfb issue into a marshall jtm clone. I put some traces of my problem, The sine is pretty neat without nfb conected.I really don.t know wher to start.What do you think ,please?Thanks.Catalin

[ATTACH=CONFIG]37395[/ATTACH]
[ATTACH=CONFIG]37396[/ATTACH]
[ATTACH=CONFIG]37397[/ATTACH]

There are only 9 db feedback into circuit 27k series with 5k pot shunt with 100nF between wiper and ground. The last pic shows how the trace looks like with shunt resistor in parallel with 100n (presence at 0) first and second with presence on 10
Thanks

I have some issues with your post, please clarify
1) when the cap is in parallel with the full pot track, or from ground lug to hot lug, that is "Presence on 10" , the maximum treble boost because of the least NFB , you call it "Presence on 0"
2) when the wiper is touching the ground lug = capac itor shorted = zero influence on NFB = zero treble boost, that is called "Presence on 0".

I have some comments on the waveform, but we need to agree first on scale numbers or wiper position.

Only advance possible now is that you have an incredibly clean sinewave in any case (for a Marshall) and you definitely do NOT have "an NFB problem" ... at all.

Another doubt is: how do you know you have "9dB NFB"?
I'm perplexed.

[catalin gramada]

Thank you guys
All right. the first test was with 400hz, i did short one at 1 khz in next pic. Looks more better at 1 khz but not perfect. I did also a sketch how my layout looks like. Thanks

[catalin gramada]

I have some issues with your post, please clarify
1) when the cap is in parallel with the full pot track, or from ground lug to hot lug, that is "Presence on 10" , the maximum treble boost because of the least NFB , you call it "Presence on 0"
2) when the wiper is touching the ground lug = capac itor shorted = zero influence on NFB = zero treble boost, that is called "Presence on 0".

I have some comments on the waveform, but we need to agree first on scale numbers or wiper position.

Only advance possible now is that you have an incredibly clean sinewave in any case (for a Marshall) and you definitely do NOT have "an NFB problem" ... at all.

Another doubt is: how do you know you have "9dB NFB"?
I'm perplexed.

You should not be. I just measured considered the voltage where it start to clip.with and without it. at those moment I remembered was ca.960 mv without nfb in pi input just the moment start to clip and ca 2,6v with nfb conected. I think I did it right
Of course you're right cap shorted means "0"presence, full counterclockwise position on my panel at least.Allready edited

[nickb]

I see some blips - are they always in exactly the same place or do they move along the waveform? Try changing the frequency by say 5Hz. Do they move now?

[catalin gramada]

You're right was not in the same place and also is moving if I turn the presence pot or the controls in tone stack, I haven't this issue without nfb conected everything is allright

[nickb]

You're right was not in the same place and also is moving if I turn the presence pot or the controls in tone stack, I haven't this issue without nfb conected everything is allright

Let's try that again.


Set the controls so you see the blips and then leave them alone.
Do the blips move along the waveform?
Change the signal generator from 400Hz to 405Hz - what happens - do the blips move faster/slower?

[catalin gramada]

Oh, now it is clear what you asked for, sorry for confusion...there are not blips on the screen and definely nothing moving on the trace on duty...sorry, was confused. It is just those bug which is frequency dependable as time it react to presence and tone control. The voltage level did not change it

[nickb]

Oh, now it is clear what you asked for, sorry for confusion...there are not blips on the screen and definely nothing moving on the trace on duty...sorry, was confused

OK - Well I guess that rules out any mains hum issue.

Now try the square wave input. This will give us some idea of how stable it is by looking at the ringing. Post the images.

[catalin gramada]

presence on "10"

and with presence on "0"

400hz

[nickb]

It is safe for power stage , please ?
I'm just begginer so have to ask if you don't mind.
[ATTACH=CONFIG]37400[/ATTACH]

400hz square

It should not normally be a problem. Don't drive it too hard other wise the power stage will clip and it will hide what we are looking for.

There's some pretty weird stuff going on there (and not quite what I was expecting) but highly suggestive of instability. Edit: -see post #14.

Are you testing using a dummy load or a speaker?

I think that in post #5 you said that for the same voltage level out you input ~960mV without feedback and 2.6V without it. If that is true than I think that eliminates the possibility that the phase is wrong on the transformer which would give you positive instead of negative feedback. But, just the be sure, and also because I don't have a better idea right now , can you swap the wires from the output transformer to the plates and see that happens?

[catalin gramada]

It is a power resistor 7.5 ohm. I swaped OT leads...just for a fraction of sec cause some started to boiled. I did not realize what but I think the noise was comming from OT transformer.
Yes I talked about the voltages aplied at the input of PI in the border of clipping

[nickb]

It is a power resistor 7.5 ohm. I swaped OT leads...just for a fraction of sec cause some started to boiled. I did not realize what but I think the noise was comming from OT transformer.
Yes I talked about the voltages aplied at the input of PI

Well at least we can be 100% certain that you had it the right way round to begin with AND you now know what can happen when it's wrong

Edit:
For some reason the second of the two square wave shots didn't come up the first time I saw the post #11. Now that I have seen it looks like you might be driving the square wave into the input jack. I thought you were going into PI input from post #5. I think all we are seeing is the HF boost from the preamp. I really don't see any sign of instability.

Go back to your 400Hz sine wave and set things up as in post #1. Now scope the input to the PI - is there any sign of the blips? I'm testing the idea that the blips are present there but only when you get to 2.6V

[catalin gramada]

Yes all tests was done with signal aplied at the input of the amp. The remark about PI voltage was relative to the amount of nfb

the input of PI with 2.7V

and 2V