Get nipples.How can I get rid of ?

[catalin gramada]

Hello, I.m asking for a solution to get rid by horrible artefacts in my circuit. I remember I read somewhere there are due to overloading of power tubes and it was suggested to raise the value of grid stoppers as solution. It was raised progressively from 5.6k, 10k to 22k now. I have to stop over there to not raise the grid leak value over the specs. But let me to understand what can be done,please. I post a sketch and traces. In yellow is OT output at 4 ohm tap and blue the power grids with 47vpk signal applied. The bias set at 43v. Can You help, please? Thanks.





Edit : the voltages from sketch are idling voltages. The supply slide 50v around to full throtle

[catalin gramada]

Changing grid stopper with 47k, the same results. Grid leak changed to 100k for confidence.

[catalin gramada]

And this is without feedback applied:



Can someone explain what is going on ? How to apply some nfb without to get nasty artefacts, please ? Thanks

[Chuck H]

I remember I read somewhere there are due to overloading of power tubes and it was suggested to raise the value of grid stoppers as solution.

That's confusing. Are the grid stop resistors supposed to mitigate "overloading"? and how? What does overloading mean? The OT primary impedance is too low? Are we just talking about drive to the power tube grids? Nope...

The power tubes can easily be in full square wave shape when overdriven. You're not even close to that in those shots. I think what you have is a parasitic oscillation. Since it's manifesting more with the NFB connected I would suggest you route the NFB lead to the tail of the PI away from the preamp circuitry and ground the PI tail and it's supply capacitor separate from the preamp. The problem may go deeper, but this is a place to start.

[catalin gramada]

And get the same nipple with almost same amount of nfb direct into PI tail

[catalin gramada]

That's confusing. Are the grid stop resistors supposed to mitigate "overloading"? and how? What does overloading mean? The OT primary impedance is too low? Are we just talking about drive to the power tube grids? Nope...

The power tubes can easily be in full square wave shape when overdriven. You're not even close to that in those shots. I think what you have is a parasitic oscillation. Since it's manifesting more with the NFB connected I would suggest you route the NFB lead to the tail of the PI away from the preamp circuitry and ground the PI tail and it's supply capacitor separate from the preamp. The problem may go deeper, but this is a place to start.

Think is drive enough, the nipple is still there, but bigger. The feedback wire is proper routed, tight twisted 4ohm tap through nfb resistor and common directly over shunt resistor grounded directly into ground point stage, meant also decoupling cap, and yes, far away by sensitive circuit or leads.

[Chuck H]

And get the same nipple with almost same amount of nfb direct into PI tail

Where else were you applying NFB? And why do you have the same gain with and without NFB applied??? I think there must be an error in the circuit. Though I can't say what other than it's possible your NFB is PFB (positive feedback) and that is causing the instability.

[catalin gramada]

The trace looks very similar with what is happen there: http://www.valvewizard.co.uk/cathodyne.html.

[Chuck H]

The trace looks very similar with what is happen there: http://www.valvewizard.co.uk/cathodyne.html.

Does this amp have a cathodyne PI? I thought this might be the amp from your earlier thread.

[catalin gramada]

Where else were you applying NFB? And why do you have the same gain with and without NFB applied??? I think there must be an error in the circuit. Though I can't say what other than it's possible your NFB is PFB (positive feedback) and that is causing the instability.

If you take a look on my sketch the first PI s was applied into cathode of the first stage before inverter, which is DC coupled with that. I changed just to check with 100ohm resistor into pi tail and is the same

[catalin gramada]

No , it is a DC coupled ltp

[catalin gramada]

Those nipple appears exactly in the moment when voltage peak into power grids is identical with dc bias voltage. NFB can be applied either in PI tail or into cathode of previous stage equally as it is not any supplementary pole in between. I just checked in both for the same amount with the same results.And is clear a nfb consequence as time without nfb the nipples are not present. The frequency for test was 400 CPS.

[J M Fahey]

And why do you think those nipples are wrong?

All you can hear is what´s present at plates, which through transformer appears at speaker terminals, why do you care at all about what happens at power tube grids or PI plates?

That amplifier is working as expected.

[catalin gramada]

why do you care at all about what happens at power tube grids or PI plates?

Thank You.
No good ?

[Dave H]

To add to what Juan said -
The nipple on the grid is when the power tube is turned off so its plate voltage can't follow the input signal. The negative feedback loop tries to make the plate follow the input by driving the grid more negative. When you see the nipple on the grid that power tube is turned off so it is not contributing to the signal at the speaker.