Tremolo circuit of the Tweed tremolux

[Victor Denance]

Hello!

Seems like the Tweed Tremo is quite the hype these days...
This time my questions are about the trem circuit.
I plan to build something like a Tweed Deluxe w/ trem (or a 5E9 w/ cathodyne PI, depends how you look at things...). I love the bias vary trem of the Brown Vibrolux / Vibroverb, and I'd like to install this type of trem on a cathode biased amp.

I'm wondering if I could connect the oscillator to the 2 220k PI resistors normally connected to ground in a 5E3, instead of connecting it to the negative bias line (like in the fixed biased Brown Vibrolux/verb, 5G9 Tremolux...).

Would that configuration work?
Would I have to mod the trem circuit a bit to accomodate?
Would I need the full 12AX7 trem (w/ cathode follower triode) or the single triode circuit (Princeton like)?

Thanks again!!

Victor

[tubeswell]

I was asking a similar question a few months ago. If you want to have a bias vary trem on a cathode biased amp, you have to connect the trem circuit like the Gibson GA17RVT. The resistor (R25) between the ground and the grid load resistors is necessary to give the oscillator a reference point, otherwise it doesn't work as well. (FWIW my question arose when I added a switchable cathode bias/fixed bias switch to a 5G9 I built, but didn't put this resistor in (yet) and found that the trem isn't as effective on cathode bias as it is on fixed bias. However I haven't been bothered to go back and add it (yet), because I have decided for now that I prefer the sound of the 5G9 in fixed bias mode, so I have left it switched there)

The 5E9A and 5G9 use a whole tube for the trem because one triode is like a cathode follower and the oscillating is still done by one triode. So if you have a one triode trem (like a Brown Princeton 6G2), it will still work, but it is not as hypnotic/intense.

Another trem I find intriguiging is the Vibrasonic 5G13 (or for that matter the 6G13A). I am so curious, I simply must build one

[Victor Denance]

Tubeswell, thanks so much for your answer!

I looked at several Gibson amp schemos (featuring a trem AND cathode biased power tubes). And it seems like the GA17RVT is the only one using that "R25" resistor (350k connected between the oscillator / 2x220k PI resistors and the ground). All the other designs lack this resistor (GA8T, GA16T, GA18T...).
Pretty strange they would not install it if it makes the trem more effective.

If you happen to add this guy in your own Tremolux in cathode biased mode, please let me know the results!
BTW, I don't have any 350k resistors handy, may I use a 220k or 470k or 1M for the same effect?

Thanks again!

Victor

[tubeswell]

I looked at several Gibson amp schemos (featuring a trem AND cathode biased power tubes). And it seems like the GA17RVT is the only one using that "R25" resistor (350k connected between the oscillator / 2x220k PI resistors and the ground). All the other designs lack this resistor (GA8T, GA16T, GA18T...).
Pretty strange they would not install it if it makes the trem more effective.

If you happen to add this guy in your own Tremolux in cathode biased mode, please let me know the results!
BTW, I don't have any 350k resistors handy, may I use a 220k or 470k or 1M for the same effect?

Thanks again!

Victor

My understanding (limited as it is) is that in the GA17RVT, R25 is part of voltage divider (in conjunction with the 1M at R34 - or is that R"39"? - can't read the schematic clearly enough) that 'regulates' the charge on C13 (C13 couples the oscillator to the output tube bias, and the waveform of the tremolo passes through this, raising and lowering the negative voltage to the grid load resistors). At the same time it prevents the voltage swing from the trem oscillator dissappearing straight to ground (where if would go if the grid load resistors were grounded, like they otherwise would be in a 'normal' output stage of a cathode biased amp - i.e. 'normal' being a stage without a bias trem). So (as an afterthought), R25 also would 'elevate' the trem waveform above ground, in effect, biaising it. So having R25 (instead to going direct to ground) helps keep the trem waveform from being clipped/cut off (or at worst, switched off). (Does that sound like too much guesswork?)

I'm theorising (sounds like a more scientific word than guessing) that if you make the value of R25 too low, then the voltage swing from the trem oscillator might get cut-off or clipped (depending on how low R25 is). However if you make the value of R25 too high, then (i'm also guessing) that the surplus output tube grid current might not be able to escape 'down' the grid load resistors to ground fast enough to prevent the output tubes going into thermal runaway.

On the other hand (now I'm guessing like there's no tomorrow), if you take R25 away altogether (without having the grid load resistors grounded), then you aren't getting a voltage divider at all, and the 1M just becomes a resistor in series with the oscillator that saps up the wavering current (and voltage) from the oscillator, and the quality of the resulting trem waveform would depend on the quality of the actual individual resistor (which I imagine would not be perfectly constant under different parts of the trem waveform cycle). So having R25 there would be a more reliable way of regularising the trem waveform.

So if you don't have 330k, then try 300k or 270k. You could try 470k and see whether your output tubes red-plate. I think it would be okay, but the only way I would proceed to find out, in the absence of better advice, is to suck it and see.

In fact, looking at the GA18T schematic just now, the 250K RA Depth pot has the same function as the R25 on the GA17RVT (which doesn't have a depth pot). When the pot on the GA18T is on max depth, there is 250K between the bottom on the grid load resistors and ground, and when it is wound back there is no depth (so I'm feeling more confident every second about what I just said :-)

In fact, you could replace R25 on the GA17RVT with a 250k-330k pot and bingo, you would have a depth control.

You may almost certainly get a better explanation/advice from others on this board. I'm only attempting to offer an explanation in order to be refuted (and thence to learn).