Dialing in the trem?

[Chuck H]

On a new build I used a trem circuit normally employed for 6L6 amps with an EL34 amp. The trem specs B+ @ 303V. I thought my calculations were on but I have a B+ on that node of 320V. The oscillator swing is supposed to be 55V. With the higher B+ it's probably a little more. And the bias voltage on the EL34's is probably less than what 6L6's would use. So there is some odd distortions at the extreme ends of the effect at higher intensity settings. Also, even at lower intensity the trem sort of goes WOMP WOMP WOMP instead of WA WA WA if you get my meaning. Now...

I've never built a trem amp before. I'll figure out how to reduce the oscillator swing and hopefully adjust the waveform. But are there any tips for a trem newb about how the oscillator swing should relate to the bias voltage or what sort of waveform is considered good. As in, we all know that the more staccato on/off effect of the BF roach system isn't revered. So what am I chasing for a bias vary trem? I'll be putting it on the bench some time tonight. I haven't even put a scope to it yet.

Other than these two glaring issues there may be other parameters that I need to idealize. Any experiences and advice would be appreciated.

Chuck

[Enzo]

So if you back off on the intensity pot will it come into reasonable tremming? The oscillator is just an oscillator, and if the waveform is clean, then all you need worry abpout is amplitude. If for some reason the LFO waveform is clipping or something, that might be involved. I don't thin 17 extra B+ volts is going to clip the LFO. Just idly thinking I'd even doubt the extra 17v is causing the over-trem. I would believe the different between 6L6 and EL34 bias levels in involed.

Working on the assumption the LFO is OK, then we need to reduce the trem drive. If turning the intensity down to say 8 instead of 10 makes the difference, then just add a resistor above the pot to voltage-divide the reduction you need. Enough resistance so the pot all the way up can't take you into womp land. I'd say try that experiment and then you tell us what sort of relationship they ought to have. I know when I encounter womp womp problems, I just take steps to reduce drive, I never thought of it in terms of percents of bias voltage or anything.

Just had a Vox AC30CC2 in here, and it had the nicest sexiest trem I have heard in a long time.

I like a smooth throbbing trem, and I just hate ping pong trem.

[Chuck H]

Thanks Enzo. The series resistor with the intensity pot was part of my plan. But... On the scope tonight, this is what I found...

The trem does indeed "womp". It sweeps smooth on the bottom (cold biasing) and then spikes sharper at the top. In fact there are five more volts of swing on the hot end than the cold. Driving the amp straight into clipping. Now...

This is a reissue Fender design. I used it because it doesn't have the silly huge 470k Rk cathode follower. I didn't want to die someday and have a moron plug some weakling, new issue Rusky tube into the trem socket and burn it up. I wanted the amp to work with any tube. So I used this circuit and elevated the filaments 50V.

The reissue trem does use some NFB from the CF stage back to the input as part of the oscillator. It's very different from other oscillators I've seen (and understand better). The forum won't let me upload the image but the circuit is used in this amp (it's not a very good image. Sorry) :

http://www.webphix.com/schematic%20h...verb_schem.pdf

I suppose I'll just need to dick with it until it works I've heard this circuit work very well. But in my amp it doesn't yet.

[Chuck H]

OK... This get's a little goofy... The amp switches between a Vox AC/TB type circuit and a BF type circuit. And it seems the zero NFB handles the trem more gracefully in all capacities. I've managed to smooth out the oscillator and the trem is downright magical when the amp is under 2/3 power, BUT...

Even approaching higher drive levels, the trem contributes distortion that just sounds garbled and messy. I actually predicted this. But is it common??? It's much worse on the BF settings (where NFB to the PI is employed). And if I clip the power tubes when the NFB is employed the trem not only adds garbled distortions, but also fails to trem on heavy signals and then fades in as the signal decays. The no NFB setting doesn't do this.

C'mon!!! Am I the first guy that's ever dealt with this??? Is the garbled distortion normal? Is there a way to minimize it? Is bias vary trem stricktly a clean tone thing?

Some preliminaries... Foreseeing this outcome I biased the EL34's at 20W each. In hopes that it would minimize distortions inherent in bias shift. That's with 450Vp and 430 screens. I don't think biasing hotter is a good idea, though it seems to help.

Maybe the question is, is there a way to reduce the swing of the oscillator and still have it affect the actual bias? Attempts at reducing oscillator gain have proved unsuccessful so far.

Am I alone here? Hello (TAP! TAP! TAP!)... Is this thing on?..

[Tom Phillips]

Chuck,

Do you find that the garbled distortion goes away if you back off the tremolo intensity control?

I have found in general that all amps that use power tube bias vary tremolo will get nasty sounding if driven hard while using the tremolo IF the tremolo intensity is set too high. It is often that a customer asks for more tremolo intensity. The % modulation can only be increased if the amp is always played well below clipping. I don't try to change the amplitude of the LFO, I just adjust the range of the intensity control and make sure that it can't be set too high for the maximum desired paying level.

Tom

[Enzo]

Tom has just explained in different words what I said in post #2, so that makes two votes for the same approach.

[Chuck H]

It's probably my unfamiliarity with using this effect that has me spinning. Everything said so far seems accurate to what I have. And consistent with what I would expect considering what the circuit is doing to the bias. So, thanks you for playing along. I guess I thought there was some magic way to make it work without the drawbacks?!? Considering how many people proclaim the superiority of bias shift trem, and that I never read about the ugly distortions until after building one for myself, I didn't think it would be so finicky to use.

I will say that, when used strictly clean, that the bias increasing and decreasing does have more shimmer and vibe than the bug trems. I'll find a sweet spot in the intensity range and static bias setting and call it good.

Thanks.

[Tom Phillips]

...Considering how many people proclaim the superiority of bias shift trem, and that I never read about the ugly distortions until after building one for myself...

Some of those people may be mesmerized by the pulsing blue glow seen in some 6L6 amps that use bias shift trem. As with many other things each different type circuit approach to tremolo has its pros & cons.

[jimboyogi]

Chuck H, are you finding the trem oscillator voltage varying with frequency?

With the various values of Ra, and RC s that I have tried, I inevitably get about 1/2 the voltage swing at ~3 Hz compared to at ~9 Hz. Ie at 3 Hz, "speed" pot all the way down, the trem voltage output is say 20V p-p. But if I turn the "speed" pot up full, the voltage climbs to say 45V p-p.

This is a problem, because as discussed you want to set the maximum voltage swing to the output tube grids via a voltage divider. But as the voltage swing varies so much with frequency of the LFO, you are forced to set the maximum swing for the highest frequency. This leaves the swing at the lowest frequency substantially less than desired.

Anyway, that's the problem I've been having

[loudthud]

With the various values of Ra, and RC s that I have tried, I inevitably get about 1/2 the voltage swing at ~3 Hz compared to at ~9 Hz. Ie at 3 Hz, "speed" pot all the way down, the trem voltage output is say 20V p-p. But if I turn the "speed" pot up full, the voltage climbs to say 45V p-p.

Could be the bypass cap on the LFO is too small. Try 100 or 220uF.

Also, doesn't the LFO modulate the B+ and get to every stage in the preamp?

[jimboyogi]

Thanks Loudthud. With the circuit that I have been trying, I am using a red LED as Rk/Ck on the LFO 12AX7 triode. According to Merlin's website, this should provide de-coupling down virtually to DC.

As far as modulating B+, with no input, all volumes maxed, and trem "depth" all the way down, you can hear the LFO modulating the noise signal at the output. Not loud, but it is coupling in there somewhere

I am also using a (relatively large) 0.47 uF coupling cap on the LFO output, to minimise lower freq losses through this cap.

[Alex R]

bias vary trem can be tricky in bigger amps. I had a similar problem with a Pro that someone had put in one of those general bf new boards, and it had bias trem instead of the optoisolator type. Tried all kinds of alterations but nothing worked - until I tried running it with a much cooler bias, which got rid of the whomping etc. So try cooler bias.

[Chuck H]

Thanks gents. I have an embarrassing confession. The ugly distortions during trem bias shift were due to a failing power tube. The net current for the two tubes did line up with the plate volts and all appeared normal. The tube finally failed. The filament would simply fade in and out. I did swap sockets and the problem followed the tube. With a new set of power tubes the odd distortions are virtually non existent. Even when clipping. Shame on me for not checking the bias of each tube individually.

[mort]

I didn't quite understand it all, but I learned several things in this thread. Thanks