Sluckey’s diode across the intensity pot mod on a 6G9-B Tremolux

[Mark]

The waveforms inside a negative feedback loop often look weird, because they're a composite of the actual signal and the error correction signal.
In this case the amplitude modulation and any thumping are part of the error correction.
It's best to open the feedback loop before trying to analyse what's going on.
When that's done, I suggest you remove the PI valve and assess how evenly the modulation affects each push pull side.
Then refit the PI valve and verify it hasn't changed anything.

https://www.thetubestore.com/lib/thetub ... ematic.pdf

Thanks for your help. I have measured a couple of waveforms but I don’t know if they are what you meant.
If not what waveforms do you suggest that I look at?

I did remove the negative feedback and took a few measurements. Here is the speaker output where the voltage is 4.71Vpp, the voltage with the negative feedback in circuit is 5.79Vpp.

IMG_5127.jpeg

IMG_5107.jpeg

The voltages around the oscillator were the same regardless of the negative feedback.

The shape of the waveform on the input of the Phase Inverter looks cleaner and measures 479mVpp as opposed to 559mVpp.

IMG_5131.jpeg

IMG_5115.jpeg

The waveform on the left 6L6 without negative feedback measures 38Vpp whereas the left 6L6 with negative feedback measures 37.2Vpp. The waveforms are quite similar.

IMG_5129.jpeg

IMG_5111.jpeg

The input to the right 6L6 without negative feedback measures 31.4Vpp whereas the right 6L6 with Negative Feedback measures 31.6Vpp.

IMG_5130.jpeg

IMG_5112.jpeg

I did try moving the earth connections of the oscillator and driver to the main earth on the chassis and while I haven’t measured it, it doesn’t sound quieter than before.

[Helmholtz]

The shape of the waveform on the input of the Phase Inverter looks cleaner and measures 479mVpp as opposed to 559mVpp.

I don't understand how you can have such a large LFO signal at the PI input, particularly without NFB.
Looks wrong. A signal at the PI input will cause a differential PI output signal that appears at the speaker output.
The heavy distortion at the output might be due to OT saturation because of the very low frequency.
Ideally there should be no LFO signal at the PI input and the speaker output.

Is the tremolo wiring original?
How does the PI input look without the PI tube?

[Mark]

The shape of the waveform on the input of the Phase Inverter looks cleaner and measures 479mVpp as opposed to 559mVpp.

I don't understand how you can have such a large LFO signal at the PI input, particularly without NFB.
Looks wrong. A signal at the PI input will cause a differential PI output signal that appears at the speaker output.
The heavy distortion at the output is due to OT saturation because of the very low frequency.
Ideally there should be no LFO signal at the PI input and the speaker output.

Is the tremolo wiring original?
How does the PI input look without the PI tube?

I agree, I don’t know why there is a signal there at all. I did try lifting the earth connections of the oscillator circuit and move them to the main chassis earth. I haven’t taken any measurements but the thumping seems equally audible.

[pdf64]

The trem oscillator is immediately adjacent to the LTP input / channel mixer; I wonder whether the circuit board is conductive?

[Helmholtz]

The trem oscillator is immediately adjacent to the LTP input / channel mixer; I wonder whether the circuit board is conductive?

Good point!
Hoping for the scope pic of the PI input without PI tube.

[Mark]

I will be away for a couple of days, but I will provide the picture. Previously there was nothing there but I might have had the probe on X 1, when it needs to be on /10 as it attenuates signal.

[Mark]

The trem oscillator is immediately adjacent to the LTP input / channel mixer; I wonder whether the circuit board is conductive?

Good point!
Hoping for the scope pic of the PI input without PI tube.

As promised this is the picture of any signal at the input to the P.I. stage. I will work on the premise that the circuit card could be conductive.

I had to use the CRO probe on /10 so I didn’t attenuate the signal, so the amplitude of the signal is 78.4mVpp.

I found that the thumping was quieter when I attached the probe to the input of the phase inverter. So in reality the signal is probably greater in amplitude.

Thanks again for your assistance, I appreciate your help.

IMG_5143.jpeg

[Helmholtz]

Does the signal at the PI input disappear when you disconnect the PI input coupling cap (1nF)?

[sluckey]

Does this 10Hz waveform show up on any of the B+ filter caps?

[Mark]

Sorry, I missed the posts after my post. I will try the suggestions that are mentioned.

As far as the board being conductive I removed all the components from the cathodes and the grids of the P.I. and mounted them off the valve socket. I used the same earth point for the P.I. stage, but I will try earthing it elsewhere.

IMG_5165.jpeg

EDIT: I changed the waveform picture as I realised that I had initially measured the waveform with DC selected on the CRO and the lower amplitude waveform was measured in AC selected on the CRO. This resulted in a lower value waveform being shown.

The thumping noise is still there, the original output waveform was 5.79Vpp whereas the waveform is now 4.97Vpp.

IMG_5169.jpeg

The oscillator waveform is still present at the input of the P.I. stage. I think it’s much the same as previous measurements which were 559mVpp, 479mVpp, and this measurement of 428mVpp.

IMG_5167.jpeg

I will try removing the input to the P.I. stage and measuring the waveform on the rails. Unfortunately, I only have a regular CRO probe and not a high voltage CRO probe.

[Mark]

Does the signal at the PI input disappear when you disconnect the PI input coupling cap (1nF)?

No unfortunately, it gets a bit larger.

The first waveform at the speaker output is the initial picture I took where the waveform measured 5.79Vpp.

IMG_5107.jpeg

The second image is with the components attached to the valve socket. The waveform measures 4.97Vpp.

IMG_5169.jpeg

The last image is waveform is with the components mounted on the valve socket and the input to the P.I. stage removed. The 1000pF cap is hanging in the breeze.

IMG_5168.jpeg

[Mark]

Does this 10Hz waveform show up on any of the B+ filter caps?

Thanks for your reply and suggestions. Sorry it took me so long to reply to you.

I thought poorly made videos might be a better way to respond to your suggestion.

In the first video the P.I. stage still has grid and cathode components mounted off the valve socket to determine if board conductivity was an issue. Given the amplitude of the thumping waveform at the speaker output was 5.61Vpp with the amp stock and 5.49Vpp with the components mounted off the valve socket. I haven’t put pretty much discounted board conductivity being a factor for now.

In this video we look at the waveform at the output of the amp.

https://youtu.be/2T_idUvmc74?si=UGE61bLZTQOi5r7x

This shows the oscillator frequency being superimposed upon the preamp power supply node.

https://youtu.be/QQN0auKMOXk?si=R3I61DBA42zCB34t

This shows the oscillator frequency being superimposed upon the P.I. stage power supply node.

https://youtu.be/K8E5fCBW5pI?si=0pItKyqlUPOCw_-t

This shows the oscillator frequency being superimposed upon the screen grid/oscillator power supply node.

https://youtu.be/wMSClmQEcEY?si=Sjh6dRvGmn_lTRCI

I thought the power supply could use more capacitance to in order to regulate the rails better. If this was achievable then perhaps the thumping at the output could be minimised. I used two 40uF@450VDC caps. One was placed in parallel with the screen grid filter cap and the other one in parallel with the P.I. stage filter cap.

This is the waveform at the output of the amp after the extra filtration was added.

NOTE: I do make an error in the text where I say 4.49Vpp instead of the intended 5.49Vpp.

https://youtu.be/muSlJwXUMws?si=AcLDrGhLvbPOxZ00

This shows the oscillator frequency being superimposed upon the preamp power supply node after additional capacitance was added to the power supply.

https://youtu.be/vx263tCymtE?si=S1qo5JuUDToduHvG

This shows the oscillator frequency being superimposed upon the P.I. stage power supply node after additional capacitance was added to the power supply.

https://youtu.be/qxFh5XdtVPk?si=JFmy2Vz_OJr0DtPR

This shows the oscillator frequency being superimposed upon the screen grid/oscillator power supply node after additional capacitance was added to the power supply.

https://youtu.be/vJlo9YJn37Y?si=DC9hbt1sD58sopmc

I should point out that I haven’t tried the resistor in series with the diode yet, but I shall. I did try a 0.047uF cap across the intensity pot and it did attenuate some of the bass response of the thumping. The other part of the mod is to try a cap from the plate of the oscillator to ground or perhaps go to the screen grid power supply node.

Thanks again for your assistance, and I look forward to further discussion.

[sluckey]

Put the tremolo tube on a separate power supply node. Connect a 1K to the screen node and then a 40µF cap from the 1K to ground. Use this new node to supply the tremolo circuit. Does that help the thumping?

[martin manning]

IDK if this has been solved, but thumping seems to be either LFO signal leakage into the signal path, or by other reports badly miss-matched output tubes. I put a simulation together to see what the waveforms should look like and to investigate the effect of the GNFB. The traces below are with zero signal applied at the PI input and with NFB connected. Disconnecting he NFB has essentially no effect other than to reduce the small sinusoidal voltage on the PI. Traces shown are the following:

V lfoa - oscillator tube anode
V drivk - CF driver tube cathode (note clipping)
V bias - raw bias at 0 intensity pot
V bias_t - bias with LFO signal at wiper of intensity pot as applied to junction of 220k's
V a - power tube anode
V g2 - power tube screen
V load - voltage on 4 ohm load
V pig1 - PI input grid
I U3:Anode - power tube current

Speed and Intensity pots are at 50% rotation.

[Mark]

IDK if this has been solved, but thumping seems to be either LFO signal leakage into the signal path, or by other reports badly miss-matched output tubes. I put a simulation together to see what the waveforms should look like and to investigate the effect of the GNFB. The traces below are with zero signal applied at the PI input and with NFB connected. Disconnecting he NFB has essentially no effect other than to reduce the small sinusoidal voltage on the PI. Traces shown are the following:

V lfoa - oscillator tube anode
V drivk - CF driver tube cathode (note clipping)
V bias - raw bias at 0 intensity pot
V bias_t - bias with LFO signal at wiper of intensity pot as applied to junction of 220k's
V a - power tube anode
V g2 - power tube screen
V load - voltage on 4 ohm load
V pig1 - PI input grid
I U3:Anode - power tube current

Speed and Intensity pots are at 50% rotation.

Thanks for your input Martin, I don’t think it will ever be resolved. I find it so strange that the thump is quieter with the P.I. removed.

I have chased up nearly all of the possible causes and not found the root cause of the problem.

Given that this tremolo causes the output valves to run significantly hotter, I’m not a big fan of this type of tremolo now.