Early (1957) Supro tremolo... anyone?

[EFK]

One of my Supros has a hacked trem circuit and while the amp works perfectly, I don;t know how the trem was originally wired and can;t get it to work. What's left of the original circuit is a 250K switched speed pot and a 500K intensity pot (actual measured values are off a little, but I think that's what they were intended to be tolerances aside). The circuit ran off 1/2 of a 12ax7, *not* a full triode, and the amp is cathode bias. The trem does not act on the power section - it seems to have been working on the preamp section only, probably altering the bias on the cathode of one of the other triodes (I'm guessing, not positive). I have schematics for Gretsch/Valco tremolos that work off of a whole 12ax7, but I can;t find a schematic for one that works off only 1/2 a triode. The voltage going into that 1/2 triode was pretty hot too, I think it was coming right off the B+ rail at about 350 volts with no plate resistor that I can find (it's a rat's nest in there). HELP!!! I'd like to keep this as original as possible - does anyone have a schematic for one of these old Valco trems with such pot values/half triode arrangement? Anyone? Thanks!

[EFK]

ACtually I've found some other schematics - ampeg, gibson 16t, epi devon - that look like they might be the same kind of thing, and these act on the PI so maybe this trem acted in that manner. This is a paraphase PI, though, so I'm not sure where the trem signal would tie in...

[tubeswell]

there's some supro schematics over at the Hoffman forum (Don't know if what you're looking for may be there)

[EFK]

Thanks for the link - they sure have a lot of schematics there, but as far as I can determine all of them use both halves of a triode. I think by the early 60s Valco/Supro/Gretsch/Airline etc. had settled on a fairly standard trem that used both halves. The earlier amps from the 50s must have had a more primitive arrangement - still looking!

[madkatb]

Every Supro schematic I have shows both halves of a 12ax7 being used for tremolo but several show the circuit tied to a preamp cathode. The only single triode tremolo I could find is a Kay 500. Jack Darr's amplifier book shows a "typical" tremolo circuit on p. 23 using only a 6sq7 (single triode) as the oscillator.
edit. Found one. Wards Airline (so it's a Supro) GIM-9131A. Single triode tremolo modulating the cathode of the second preamp triode.

[EFK]

Thanks for the info! Been scouring schematics left and right. The Airline looks possible but I think it's different that what was in there to some degree. I may have to settle for something like that.

Here's one that is even simpler - a later 60s Gibson Skylark trem - this would be PERFECT and easy to implement, and I think it's very similar to what was in the Supro as it just ties the cathodes together, however, I'm not sure where I could insert a depth/intensity control. It seems to have a speed control only, 500K. I'd like to use this original 250K speed pot, which probably could be adjusted, but while I can follow the schematic easily enough, I'm not really familiar with trems so I'm not sure where an intensity control could be added in. Anyone?

[EFK]

Alternately, is there any way to insert a trem that typically ties into the PI "bias splitters" into something like this Supro unbalanced paraphase PI? This trem is from a Sano 160r, the PI sketch is what I have to work with in this Supro amp.

Could I *possibly* tie two 270K (value picked at random, I've seen 220, 330 as per the Sano, etc.) resistors after the .05 output caps, one off each leg going to the output tubes, and insert the line off the depth pot (here a 500K) to the junction of those two resistors?

[tubeswell]

I'm not sure about inserting intensity into that skylark trem. My hunch would be to have some more padding between the cathode resistor for the LFO and the (other) cathode of that final pre-amp stage/driver. Possibly by putting that driver stage onto its own Rk but having a pot and some blocking caps in between that and the LFO stage's cathode?

[tubeswell]

On second thoughts have a look at the marshal 18Wer trem intensity control. The LFO stage's grid is boot-strapped to a split 820R-1k Rk, and the 1k is a pot.

[EFK]

Hey tubeswell I appreciate the help very much. I didn't want you to think I'm brushing you off, the problem here is that I understand the oscillator action but that's about it. I'm fairly bewildered by the various means of using that oscillator to effect a change in sound - how the oscillator is tied into the rest of the circuit, in other words. So I'm going through schematics and finding about 100 different trem designs, which isn't helping things. The Airline trem mentioned above looks promising but I know it's a bit more involved than what was in there. I think the skylark trem was really close, if I could figure out a way to essentially use the skylark trem and work a 500 K intensity pot into the cathode somehow as per the Marshall 18W type, I think it would be pretty darn close. I know the original trem acted on the V2 cathode somehow, which is 1.5K/35uf. 500K seems too big to have been inline like the little marshall 1K pot, I wonder if the 500K was paralleled somehow? The airline circuit mentioned above seems to be putting the oscillator cathode in parallel w/ a signal cathode, or maybe I'm completely misreading it.

[tubeswell]

This is my take on LFOs (and someone will no-doubt chime in if I'm in error) - a triode in a normal gain stage inverts the signal between the grid and the plate. This inverted signal (at the plate) is 1/2-a-cycle 'out-of-phase' with the signal at the grid. If you connect the plate to the grid (- to avoid DC being dumped onto the grid you do this with a coupling cap -) it would start feeding back on itself (and oscillating). (At this point, if you connected an audio listening device to the stage's output, the oscillation would sound like a ticking or chopping effect I think). If you connect the plate and the grid with a coupling cap and a resistor to ground (- the grid should have a resistor-to-ground anyway -), a voltage (that varies with the fed-back oscillation) develops across the resistor that has a certain frequency (because of the Resistor-Capacitor network relationship). The frequency of voltage swing (from the RC network) combines with the frequency of the 1/2-cycle phase-inversion that normally occurs between the grid and the plate, in such a way as to shift the phase of the ('naturally-occurring') inversion past the 1/2-way point in the cycle. However, one such cap and resistor doesn't provide enough voltage swing at the right frequency to result in a complete whole-cycle of difference in the phase between the grid and the plate - that is why a conventional LFO has several such caps and resistors-to-ground between the grid and the plate.

You vary the time/rate/speed of the inversion of the phase, by either changing the cap values, or changing the resistor-to ground values. Changing the resistor value (with a pot) is more straightforward. You only need to change one of the resistors in the RC network to effect a change in the rate of the oscillation.

You vary the depth of the inversion in a number of ways that involve attenuating the gain of the 'voltage signal'. One way is to put the output of the oscillator through a voltage divider. Another way is to change the bias of the LFO stage (like the 18Wer does) to make the cathode resistor of the LFO stage variable (to a certain extent).

Is that any use?

[EFK]

That's a very clear explanation - I appreciate it! I think I'm going to just get my hands dirty and come up with an oscillator that seems to suit the circuit, then experiment with different ways to tie the signal in to either another tube cathode or to the output section. I'm still not clear on what will happen if I tie two 270K + resistors to each leg of the PI after the coupling caps, decouple the oscillator signal off the oscillator tube plate with a .1 or .2 cap, run it through a 250K pot as depth to vary how much of that oscillating signal is either used or run to ground, and tie into the junction of those 270K resistors. Many amps tie the trem into the signal chain via this manner, I'm just not clear as to whether it would work with this particular paraphase PI (being as it is a very early "unbalanced" paraphase). Or maybe I can rig up a depth pot between tied cathodes, LFO and signal triode, and use it to vary how much of the oscillator is permitted to effect the cathode of the signal tube. I'll figure it out sooner or later!