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

[Mark]

I have a Tremolux in for work and the primary problem is the tremolo is thumping. I tried the Sluckey’s diode across the intensity control mod and found it to be quite effective but the amp is quieter volume wise.



The issues I refer to is the change in volume and the tremolo sounding different (better in my opinion).

In a perfect world, the volume drop wouldn’t be noticeable. I do intend to get the CRO onto it for greater clarity, but I was wondering if someone here has tried this mod previously and knows all the tricks in implementing this mod?

I will post photos of waveforms on the grids where possible.

Thanks for your assistance.

[Stevem]

Off the top of my head I would say that you lost the amount of audio drive sign equal to the turn on voltage of the diode.

[pdf64]

I suspect it's all to be expected.
The stock modulation turns the current up and down about the nominal level.
With the diode, it can only turn it down. So the average current goes down as per your meter reading, and the power amp gain goes down, because lower average anode current results in lower average gm results in lower average stage gain.

Maybe remove the diode and try a different pair of 6L6.
The ones fitted idle close enough, but go rather different when the modulation is applied. Hence the thumps can't cancel out, because they're not balanced.

[Mark]

As a picture is worth a thousand words, I thought a few of them might help the situation.

At the time of measuring with the CRO these are the voltages around the output stage.

The main voltage is 240 V

Rail voltage is 427 VDC

The screen grid voltage is 426 V DC

The bias voltage is -43.3 V DC

The left valve is drawing 28.3mA.
The right 6L6 is drawing 30.6mA.

IMG_5057.jpeg

The photo shows the positive waveform exceeding 0V. The CRO is set to divide by 10.

IMG_5058.jpeg

The positive waveform doesn’t exceed 0v and the distortion at the bottom of the waveform is still distorted but smoother. The CRO is set to divide by 10.
The picture shows why the output stage runs hotter when the tremolo is being used.

IMG_5059.jpeg

IMG_5060.jpeg

The diode is in circuit in this picture. The positive part of the waveform is clipped at -43VDC. It shows why the valves run so much cooler when the diode it across the intensity pot.

IMG_5061.jpeg

The output is a lot cleaner, but there is a drop in volume.

IMG_5079.jpeg

The CRO is set to divide by 10 and waveform is smoother in appearance.

IMG_5080.jpeg

For some reason the positive peaks of the tremolo waveform are more negative than previously. The waveform is smoother too.

IMG_5081.jpeg

I noticed that the waveforms of each grid is a bit different. I’m not too sure what to make of that.

IMG_5082.jpeg

One thing that I noticed is that the thumping is louder when the P.I. valve is in circuit. I noticed that the thumping was significantly quieter when I measured the grid of the Phase Inverter with the CRO on X 1. I did try putting a 1M resistor from the P. I. input to ground, I found the output significantly decreased.
I suppose a good question is why there is signal on the input of the phase inverter?

[EDIT: Gemini later answered my question that it’s coming via the negative feedback. I later tried removing the negative feedback and found the amplitude of the thumping decreased slightly. ]

IMG_5083.jpeg

As expected the shape of the waveform is somewhat different.

I will look at same points with the Phase Inverter out of circuit and see how the waveforms differ.

Gemini had this to say about the P.I. stage making the thumping pounder: “In the Fender 6G9-B Tremolux, the tremolo uses bias modulation where a Low-Frequency Oscillator swings the DC grid bias up and down, creating a low-frequency AC pulse that manifests as a thump. The thumping is louder with the phase inverter tube in place because the tube acts as a high-gain driver stage that heavily amplifies this sub-audible LFO wobble while sitting directly in the injection path right before the power section. Removing the phase inverter tube breaks that driver path and eliminates the high-gain amplification of the modulation signal, causing the amplitude of the bias shift to collapse and making the thump drastically quieter.”

Perhaps I ought to consider lifting the feedback resistor?

Thanks for your interest in reading my thread.

[Roe]

I removed the diode on a bias trem twin/showman, but keept it on an ac30

[John G]

Hi, I usually put a 330 k resistor (thereabouts) in series with the diode to lessen the effect. My 2c worth
John

[Mark]

I removed the diode on a bias trem twin/showman, but keept it on an ac30

Thanks for your reply. I don’t think I know of a Twin/Showman that uses bias tremolo, was this custom build?

Where did you use it on an AC-30?

They have a speed control and vibrato tremolo switch.

[Mark]

Hi, I usually put a 330 k resistor (thereabouts) in series with the diode to lessen the effect. My 2c worth
John

Thanks for your reply John, perhaps I don’t understand what you are saying, are saying you use a resistor to reduce the bias negative voltage?

[Mark]

Off the top of my head I would say that you lost the amount of audio drive sign equal to the turn on voltage of the diode.

Thanks for your reply, I thought the pictures were a good response as to what I’m seeing. I will post up more info tomorrow if I have time.

[Mark]

I suspect it's all to be expected.
The stock modulation turns the current up and down about the nominal level.
With the diode, it can only turn it down. So the average current goes down as per your meter reading, and the power amp gain goes down, because lower average anode current results in lower average gm results in lower average stage gain.

Maybe remove the diode and try a different pair of 6L6.
The ones fitted idle close enough, but go rather different when the modulation is applied. Hence the thumps can't cancel out, because they're not balanced.

Thanks for your reply and advice. I haven’t tried another pair of 6L6’s yet, for no scientific reason, but I don’t have a lot of faith in common mode noise reduction coming to the rescue.

I have found removing the P.I. 12AX7 drastically reduces the thumping noise. I will take photos of my findings hopefully tomorrow.

[Roe]

I removed the diode on a bias trem twin/showman, but keept it on an ac30

Thanks for your reply. I don’t think I know of a Twin/Showman that uses bias tremolo, was this custom build?

Where did you use it on an AC-30?

They have a speed control and vibrato tremolo switch.

yes I did a custom twin with bias trem. it sounds nice but the trem got weak with the diode.

the (custom) ac30 also got bias trem on the output tubes (with larmar master volume). in this case the diode worked better for sound and tubes. before the diode was installed, the el84s struggled somewhat.

Both amps use fender blonde style bias trem on the output stage but adds larmar master volume

[John G]

Mark.Sorry for the confusion. The resistor is in series with the diode you put across the intensity pot. It lessens the clamping effect; some were between just the diode and no diode.
John

[Mark]

As I mentioned earlier the thumping coming from the output differs if the phase inverter valve is in circuit or not.

The mains voltage was varying somewhat today but largely is was 234vac (ideal voltage is 240vac.)

The B+ rail measured 417VDC with 3.6vac ripple.
The screen rail measured 417VDC with 100mV ripple.
The bias voltage measured -42.6VDC with 210mV ripple.

The bias current was 27.3mA through the left 6L6 and 29.7mA through the right 6L6.

I will start with the difference in output between having the P.l. valve in and out.

IMG_5107.jpeg

IMG_5097.jpeg

The output of the oscillator is similar regardless of whether the P.l. valve is in circuit or not. (The divide by 10 setting on the CRO probe was used.)

IMG_5105.jpeg

IMG_5095.jpeg

Waveforms with Intensity control set to maximum.

IMG_5106.jpeg

IMG_5096.jpeg

When the intensity pot is set to maximum the current through the two 6L6’s differs.

With the P.I. valve in circuit the left valve measures 54mA while the right valve measures 43mA.

With the P.I. valve out of circuit the left valve measures 45mA and the right valve measures 47mA.

The latter measurement is what I would expect from the circuit.

The bias voltage at the node where the two 220K resistors (grid leak/bias) meet the bias voltage is slightly different. (The divide by 10 setting on the CRO probe was used.)

IMG_5110.jpeg

IMG_5098.jpeg

The voltage on the left 6L6 grid reflects the current drawn by the left 6L6. (The divide by 10 setting on the CRO probe was used.)

IMG_5111.jpeg

IMG_5099.jpeg

[Mark]

The continuation.

The voltage on the right 6L6 grid again reflects current drawn by the right 6L6. (The divide by 10 setting on the CRO probe was used.)

IMG_5112.jpeg

IMG_5100.jpeg

As the P.I. stage is somehow affecting the thumping of the tremolo, I thought I would measure the output of both sides of the P.I. though it may have limited benefit.

IMG_5113.jpeg

The amplitude is a third smaller with the 12AX7 removed.

IMG_5102.jpeg

The output on the right side of the P.I. stage plate is quite unusual. (The divide by 10 setting on the CRO probe was used.)

IMG_5114.jpeg

IMG_5103.jpeg

The last thing I measured was the input to the P.I. stage to see if signal was present there too.

With the 12AX7 in circuit I had to use the CRO probe on /10 as it loaded down the input and there was no signal present. I did try putting a 1M resistor in circuit to replicate the CRO probe but it loaded down the signal too so there was very little output.

IMG_5115.jpeg

There wasn’t any signal present on the input of the P.I. stage with the 12AX7 removed.

The earths for the oscillator and driver are on the earth bar for the P.I. stage and preceding stages. I think the next move is to separate the ground connections and have the oscillator and driver earth on the screen grid earth where it derives voltage.

However, I’m open to suggestions, thanks for reading my post.

[pdf64]

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