New Tremolux 5G9 build --> input welcome!

[tubeswell]

I think slightly more headroom. The JJ6V6S is a different tube (from a 6V6GT) - plates are more like a 6P3S, and it (going by the schematic on the JJ dadsheet) it appears to be a pentode (whereas the 6V6GT is a tetrode), - but I've never opened one up to see. Might have a go on an old one this w/e maybe if I get around to it.

[klooon]

Tremolo Speed Range --> pot wiring

Log taper pots are best for vol and tone controls. Use a linear taper for the depth control and RA taper for the speed control. 3MRA will work fine. When I did mine used a NOS 2M5RA and tweeked the tapering resistor to 120k. With 3MRA, you may want to try a 150k tapering resistor (instead of 100k) to get it working closer to the original. But when you get it working we can talk more about adjusting the trem if you want.

I have the idea that my tremolo is not working properly…

The speed knob starts to pulsate only at position 4 or higher (before 4 it doesn’t pulsate at all!). I tried to slow down the tremolo with putting a 0.022uF instead of the 2nd 0.01uF and 0.047uF instead of the 0.033uF caps. It slows down, but still the tremolo starts to work at position 4 or higher. I have the feeling that this is not right (and that the tremolo should be even somewhat slower…).

I thought it had to do with the (tapering?) resistor on the pot. I have a 3M Reverse Audio speed pot and a 160k resistor between the left and middle lug, see my layouts. But I also tried to put the resistor between middle lug and ground, a bit like in the BF amps. But there seems to be no difference, there is still the start of the pulse at position '4' of the speed knob.

Further, what happens exactly to the speed if you change the value of the resistor? What values did you use to experiment? (I see some posts in which you state that you have experimented...)

[tubeswell]

Are you using a 12AX7 triode? What are the idle voltages?

[klooon]

Are you using a 12AX7 triode? What are the idle voltages?

I am using a 12AX7. The B+ is about 390VDC. At the trem. part I measure 260 VDC at the input of the 0.033uF (now 0.047uF) cap. (and of course fluctuating if the trem is engaged). To the 25uF cap I measure 1.95VDC... Any other values you need for checking this??

[klooon]

V1: (1) 167VDC; (2) 0 VDC; (3) 2.7 VDC; (6) 160VDC (7) 0 VDC (8) 2.7 VDC
V2: (1) 190VDC; (2) 21 VDC; (3) 31 VDC; (6) 185VDC (7) 22 VDC (8) 31 VDC
V3: (1) 260VDC; (2) 0 VDC; (3) 1.96 VDC; (6) 390VDC (7) 260 VDC (8) 262 VDC (trem not engaged)
Further all heaters at (4&5&9) 3.3VAC

Are all the 0 Volts okay??


Speed pot dial:

Resistance Between right and middle lug ----- Between right lug and ground (so including the 160k in series)

position 1: 3.2MOhm ------- 3.39MOhm
position 2: 2.9MOhm ------- 3.1MOhm
position 3: 2.1MOhm ------- 2.3MOhm
position 4: 1.2MOhm ------- 1.4MOhm
position 5: 0.5MOhm ------- 0.6MOhm
position 6: 0.3MOhm ------- 0.5MOhm
position 7: 0.2MOhm ------- 0.4MOhm
and so on....

I have the feeling that this is not like it should... Who knows??

[klooon]

OK, I found what is happening. I have calculated the total resistance and plotted it in the following graph:

You can see that at the speed pot at dial position 4 gives the start of pulsation. The total resistance is ca. 1.4MOhm. In the original design the pot was 2MOhm. You can see now that it takes only a dial from 2 to 1.4MOhm to start working and this is within the original pot position 1 to 2. There is just an offset using the 3 MOhm of 1MOhm and this takes a few pot positions... Further the 160k is kicking in at the higher speed to prevent that the speed doesn't go mental, so it just limits the maximum speed!

[klooon]

Further, adding for instance a 5 MOhm resistor parallel to the right and middle lugs of the pot will put you in the right direction...



Please tell me if this makes sense. What do you think.... So is there a limit to the minimum speed where it kicks in??

[tubeswell]

Those voltages look in the ballpark.

Maybe try a 120k tapering resistor between the pot wiper and the ground. Are the other resistors in the R-C part of the LFO stage 1M? Do you have .03uF for the first cap and .01 uF for the next two?

[loudthud]

I've often wondered about these tremolo oscillators that use a 25uF cap for cathode bypass. It seems to me that it would be unaffective at the lowest frequencies. I suggest you try something like 220uF. You can just tack it in place in parallel with the 25uF. Any cap 10V or higher would work.

[klooon]

Those voltages look in the ballpark.

Maybe try a 120k tapering resistor between the pot wiper and the ground. Are the other resistors in the R-C part of the LFO stage 1M? Do you have .03uF for the first cap and .01 uF for the next two?

please have a look at my previous two posts with the graphs. please correct me if I am wrong, but the at slow/zero speed the 3M pot is at max with 160k in series. Turning the dial up lowers the 3M and the remainder part is parallel wih the 160k. then I got the figures in the graphs and I notice that the 160k is just the limiting value at highest speed. at low speed it doesn't do much compared with the 3. in series in total.
Apparently the 3M is in this design way too much and the tremolo kicks in at about 1.5M with a certain minimum speed. This is set by amongst others the three caps you mentioned. I changed the 2nd 0.01uF into 0.022uF and the neighbouring 0.03uF into 0.047uF. This actually slows the minimum speed down a bit, BUT at a similar pot dial position (about 4).
I still have the two 1M's. Is this related to the intensity or also the speed (offset)?

What about my suggestion of putting a 5M in parallel to the right and middle wiper as suggested in my previous post?

[klooon]

I've often wondered about these tremolo oscillators that use a 25uF cap for cathode bypass. It seems to me that it would be unaffective at the lowest frequencies. I suggest you try something like 220uF. You can just tack it in place in parallel with the 25uF. Any cap 10V or higher would work.

thanks. how does it work? the 25uF is attenuating when dialing in a (too) slow speed (i.e. low frequency)?

[loudthud]

Bypassing the cathode of a triode increases the gain about 6dB, but only down to a certain frequency. Increasing the capacitance lowers the frequency before the gain drops off.

The phase shift oscillator works by feeding the signal from the plate back to the grid through an RC network that shifts the phase 180 degrees. The RC network must not attenuate the signal amplitude too much or the tube will not oscillate. The tube needs to have enough gain so that together with the amount of attenuation of the RC network, enough signal gets back to the grid to cause oscillation. Note that the RC network only has 180 degrees of phase shift at one frequency. Above that frequency, there is less phase shift, below that frequency, there is more phase shift but also less signal that gets back to the grid.

So two things can cause the oscillator to stop oscillating. The gain of the tube being too low and too much attenuation in the RC network at the frequency where the phase shift is 180 degrees. The attenuation of the RC network changes as you adjust the 3 meg pot. The best way to evaluate the RC network is with a circuit modeling program because every capacitor (3) raises the order of the equasions. Generally try to keep all the capacitors about the same value.

One other thing that should be mentioned, that is the B+. In a tremolo amp such as this, there is quite a bit of the tremolo signal that gets on the B+ because the current of the output stage is changing. This acts against the oscillator like negative feedback. A symptom would be that oscillation stops or changes frequency when you turn up the Intensity control.

Tubes manufactured these days don't have quite as much gain as in previous decades. The one you have just might be low gain, try another brand. A low heater voltage will also cause the gain to be low. Check the heater voltage at the socket and check the socket for tight contact to the tube pins.

[tubeswell]

please have a look at my previous two posts with the graphs. please correct me if I am wrong, but the at slow/zero speed the 3M pot is at max with 160k in series.

The slow speed should be at the 'zero end' of the pot rotation, not the other way around. This position is where the resistance between the pot input terminal and the pot wiper terminal is the greatest. You're using 3MRA right? ( and not 3M?). If your pot is kicking in differently, you may have the input and output terminals the wrong way around on the pot. Have you tried swapping them around?

Turning the dial up lowers the 3M and the remainder part is parallel wih the 160k. then I got the figures in the graphs and I notice that the 160k is just the limiting value at highest speed. at low speed it doesn't do much compared with the 3. in series in total.

The 160k is just altering the taper of the pot. With the pot at maximum rotation you have 160k and 3M in parallel, and as you decrease the pot rotation you get increasing resistance between the pot wiper and the pot input (but the resistance between the pot ground and the pot wiper remains in parallel with the 160k - you can chart this on a spreadsheet if you measure the pot's actual taper* at various rotations first)

* With a linear taper pot, you can assume even splits of resistance and rotation, but with an audio taper, the resistance at different rotations will be logarithmic, and different types audio pots can have different audio tapers, depending on how the taper was specified in the pot design.

Apparently the 3M is in this design way too much and the tremolo kicks in at about 1.5M with a certain minimum speed. This is set by amongst others the three caps you mentioned. I changed the 2nd 0.01uF into 0.022uF and the neighbouring 0.03uF into 0.047uF. This actually slows the minimum speed down a bit, BUT at a similar pot dial position (about 4).
I still have the two 1M's. Is this related to the intensity or also the speed (offset)?

Altering either the capacitance or the resistance (or both) will alter the rate of oscillation.

What about my suggestion of putting a 5M in parallel to the right and middle wiper as suggested in my previous post?

That will just be putting another resistance in parallel with the pot's input-to-ground resistance and would decrease the overall effectiveness of the pot in changing the speed.

[klooon]

The slow speed should be at the 'zero end' of the pot rotation, not the other way around. This position is where the resistance between the pot input terminal and the pot wiper terminal is the greatest. You're using 3MRA right? ( and not 3M?). If your pot is kicking in differently, you may have the input and output terminals the wrong way around on the pot. Have you tried swapping them around?

I think I wired it exactly how you describe. Zero end of pot rotation is the dial position "1" at the chassis display. So between the 'right'/input lug and the middle lug there is about 3M resistance. Together with the 160k in series (between middle lug/wiper and left lug) you get the total resistance then. My version is a 3M RA with a maximum reading of about 3.2M, so almost 3.4M in total at 'zero end'/lowest speed.

The 160k is just altering the taper of the pot. With the pot at maximum rotation you have 160k and 3M in parallel,

Right, so at maximum rotation the 160k parallel with 3M the total resistance is 160k. So at maximum rotation (i.e. maximum speed), the 160k is now limiting the maximum oscillation speed. And it might also change the shape of the pot..

and as you decrease the pot rotation you get increasing resistance between the pot wiper and the pot input (but the resistance between the pot ground and the pot wiper remains in parallel with the 160k - you can chart this on a spreadsheet if you measure the pot's actual taper* at various rotations first)

Here it is (I was too lazy to go up to the fender '12' position...I stopped at 7...):


I think it's fine...

Altering either the capacitance or the resistance (or both) will alter the rate of oscillation.

Yes, but it changes between 3M + added resistor value(100k or so) and that 100k (=3M||100k) or so value.

That will just be putting another resistance in parallel with the pot's input-to-ground resistance and would decrease the overall effectiveness of the pot in changing the speed.

So, all in all, I think we have the same wiring. Only you use a 2.5M RA. When do you notice the first oscillation (in terms of dial position 1 until 12); already between 1 and 2?

---> I have 'measured' my minimum speed around 3 to 4 Hz (about 35 ticks in 10 seconds) ---> is this 'normal'?

BTW
- Do you have the original pot wiring with the taper resistor between the left lug and the wiper OR in series like the black faces?
- Does the original 5G9 have a linear taper and the 100k is making it more reverse log?

[klooon]

Thanks so much for the lecture!

Bypassing the cathode of a triode increases the gain about 6dB, but only down to a certain frequency. Increasing the capacitance lowers the frequency before the gain drops off.

This means that a larger capacitance will probably also let go through the very low frequencies (i.e. tremolo oscillations??) ??

I will order a 220uF and see what will happen!

The phase shift oscillator works by feeding the signal from the plate back to the grid through an RC network that shifts the phase 180 degrees. The RC network must not attenuate the signal amplitude too much or the tube will not oscillate. The tube needs to have enough gain so that together with the amount of attenuation of the RC network, enough signal gets back to the grid to cause oscillation. Note that the RC network only has 180 degrees of phase shift at one frequency. Above that frequency, there is less phase shift, below that frequency, there is more phase shift but also less signal that gets back to the grid.

So two things can cause the oscillator to stop oscillating. The gain of the tube being too low and too much attenuation in the RC network at the frequency where the phase shift is 180 degrees.

The attenuation of the RC network changes as you adjust the 3 meg pot. The best way to evaluate the RC network is with a circuit modeling program because every capacitor (3) raises the order of the equasions. Generally try to keep all the capacitors about the same value.

I saw pictures of the Trem D'lux of Keedy amps, with 0.01uF -- 0.022uF -- 0.047uF. I have the same values now. Would it be better to have for instance 3 times 0.033uF?? (Keeping in mind your explanation of one single frequency and phase shift??).

One other thing that should be mentioned, that is the B+. In a tremolo amp such as this, there is quite a bit of the tremolo signal that gets on the B+ because the current of the output stage is changing. This acts against the oscillator like negative feedback. A symptom would be that oscillation stops or changes frequency when you turn up the Intensity control.

Tubes manufactured these days don't have quite as much gain as in previous decades. The one you have just might be low gain, try another brand. A low heater voltage will also cause the gain to be low. Check the heater voltage at the socket and check the socket for tight contact to the tube pins.

I have a NOS GE JAN 12AX7WA in v3. Should be okay - I hope. Would ordering a 12AX7 with a 'high gain' designation an option here?

I checked the heater voltages and they are at about 3.3 V