Harmonic trem in Princeton reverb

[rootz]

So the last week I had the idea of combining a first gen harmonic trem (from amps like the Fender Pro 6G5) with an otherwise mostly blackface Princeton reverb. 5 preamp tubes, two power, tube recto. I've got a chassis that will do perfectly. Tried and tested circuits for most part.

Now that harmonic trem is some interesting circuit! Oscillator, phase inversion, low and high pass filters, summing amp. And I think they sound great too.

Some observations (from simulations):
- The headroom of the mixer is rather limited;
- The high and low pass filters add considerable mid scoop around 400Hz;
- The pesky 10meg RA pot.

So here is what I like to do, the goals:
- get rid of the 10meg pot. Those are not available in Europe AFAIK. Weber stocks them, but for how long? Hoffman stopped offering them, or at least I couldn't find them in the webshop. TAD had them a couple of years ago, but not anymore;
- get rid of the extra mid scoop if possible. This leaves the rest of the amps characteristics intact;
- some extra headroom would be nice.

Now, I'm not familiar with this trem circuit. Or any trem circuit actually. My knowledge about the harmonic trem is all from the last week, so really could use some experience.
I made some comparisons in ltspice. Always a handy tool to learn how circuits work and save some tedious work with algebra
I believe the 10meg pot can be adjusted by a 3megRA one, but multiple values have to be adjusted. I'd like to retain the original sweep. Maybe gain a bit of extra output from the oscillator circuit, but that's not the most important thing.
I adjusted some values in the crossover circuit, to make it mostly frequency flat.
The only way to get some more headroom out of the amp I found

Now all these changes simulate ok. But the big question is: did I change the right things, or are there (much) better ways to achieve my goals?

[Guy77]

Harmonic trem is my favorite trem.
I was at a local family owned restaurant and a band was playing and the guitar player was plugged into this custom amp that had no name on it and the cab was poorly made. It had the same controls as a Princton Reverb and about the same size. He told me it had harmonic trem and that was about all he knew.
The sound of the harmonic trem was amazing. It vibrated through the room endlessly.

Cant wait to see your finished project!
There is also a nice PCB here with the Princeton circuit but different trem.

https://ampgarage.com/forum/viewtopic.php?f=17&t=33470



Cheers
Guy

[sluckey]

Replacing the LFO oscillator cathode RC with a red LED will give considerably more oscillator signal amplitude. This should allow the use of a 3M INT pot. You would also want to reduce that 4.7M resistor to about 1M to 1.5M. I haven't tried this idea mainly because I had a 10M pot.

My Revibe has the same vibrato circuit as your circuit. I made several changes for different reasons than your list. May be useful as a reference...

https://sluckeyamps.com/revibe/revibe.pdf

Page 15 of my scrapbook has a function description of the harmonic vibrator circuit. Sounds like you may already know this stuff.

https://sluckeyamps.com/misc/Amp_Scrapbook.pdf

[rootz]

That's a great story Guy! Yes, it's my favourite trem too, especially the harmonic trem of Fender. Thank you for the link to the pcb. I do have one with the harmonic trem already. It's pretty basic. Preamp --> harmonic trem --> reverb with purely passive mixer (like the Princeton pretty much) --> Princeton Reverb power amp.

Thanks for the links sluckey. Lots of useful information. I had no idea it is the idea of the mixer stage to filter out the LFO signal. If so, it's doing a terrible job. The phase inverter looks somewhat like a floating paraphase to me, though really unbalanced. The inverter has less gain than the oscillator side and signals on both sides of the mixer grids differ by a big margin. Not that it really matters though, I care how it sounds.

Still hope there's at least someone who can share experience on the depth pot, or better getting rid of that 10 meg pot. It's a rather big hurdle. I'd hate the shipping, tax, duties for just one simple pot...

[Reeltarded]

Someone may have a torn down circuit and a used pot?

I worked my 63 Bandmaster over with the red led and doubled the caps for slow speed. Freaking sounds like a tube univibe.

https://www.dropbox.com/s/e2rwjh2tvrppp ... C.MOV?dl=0

https://www.dropbox.com/s/tua5397xb3xi6 ... 6.MOV?dl=0


I think that should work. The tweaks make it mental! It is like the water swimming IN YOU!

Yeah, so the mid scoop is part of the action.. if you flatten it too far there won't be the wobble transition between phases.

[martin manning]

...or better getting rid of that 10 meg pot.

This configuration does away with the 10 Meg pot. It's the 6G4-A circuit, though. Maybe you can do something similar.

[Reeltarded]

Martin is a tricky man.

[martin manning]

Is there any consensus on which circuit is "better," 6G4 vs. 6G4-A? The -A needs another triode, which is inconvenient.

[Reeltarded]

Unfortunately, the -A is better, but you would never know.. honestly.

Both versions are great.

[rootz]

Thanks all for chiming in here!

The A is better if you want a more deep trem than the other circuit can provide. The A is also more tilted towards the high end; it has some significant imbalance between the low and high end amplification in the crossover circuit before the mixer triodes.

That is a nice trick there Martin! I have absolutely no idea how to implement it in Fender's two tube design though.

Did some more tweaking with resistor and cap values and I think I'ver got some working solution, essentially doubling or tripling a couple cap values around the pot. Those set the sweep of the amp.

Another observation though, when lowering the 4.7 meg resistor coming of the oscillators anode. That is a great place to tweak for a bigger oscillator signal going into the mixer. The LED in place of the 4k7/4u7 combo is another.

Thanks for all the great ideas!

[martin manning]

That is a nice trick there Martin! I have absolutely no idea how to implement it in Fender's two tube design though.

Here is another way to skin it: Start with the -A, and eliminate the cathode follower. As above, the 10M RA is eliminated by controlling the intensity at the input to the cathodyne splitter. The max LFO signal at the mixer grids is easily adjusted with the large resistor feeding the intensity pot.

[rootz]

Hi Martin, also a good way to skin the cat! Another would be to use a dual ganged 1meg pot instead of the 1meg grid leaks of the mixer triodes. Just wire it with the wipers to the PI's 1 meg resistors. Might need to replace the coupling caps coming from the PI with 100n to maintain -3dB point.

I skinned the cat with some adjusted cap and resistor values. Might look into the different ways to do that when this photo I'm building fails. Attached a pic of the almost finished board and chassis it is going in.

Speaking of which: how would you ground the harmonic trem Martin? Would you keep the ground separate from the preamp and reverb grounds and join them at a star ground? I'm planning to use a 40/20/20/20 JJ cap for all nodes but the trem. So I have the possibility of separating them.

[martin manning]

Looks like your project is coming along nicely!

A dual pot might very well work. I considered that, but I like the cathodyne phase splitter with a single gang pot at the input. That way the crossover filter is left exactly as-is.

I'd take the LFO and phase splitter grounds back to the negative terminal of the filter that's feeding them, and then run that to the main reservoir ground. I would't worry too much about it though, in this case the LFO isn't firing a neon lamp.

Re DR heater current, it has six 12A_7 @ 0.3A and two 6V6 @ 0.45A... 2.7 total.

[JD0x0]

That is a nice trick there Martin! I have absolutely no idea how to implement it in Fender's two tube design though.

Here is another way to skin it: Start with the -A, and eliminate the cathode follower. As above, the 10M RA is eliminated by controlling the intensity at the input to the cathodyne splitter. The max LFO signal at the mixer grids is easily adjusted with the large resistor feeding the intensity pot.

What about a source follower in place of the cathode follower? It would allow you to do the full circuit with just two tubes.

[martin manning]

That is a nice trick there Martin! I have absolutely no idea how to implement it in Fender's two tube design though.

Here is another way to skin it: Start with the -A, and eliminate the cathode follower. As above, the 10M RA is eliminated by controlling the intensity at the input to the cathodyne splitter. The max LFO signal at the mixer grids is easily adjusted with the large resistor feeding the intensity pot.

What about a source follower in place of the cathode follower? It would allow you to do the full circuit with just two tubes.

That would work, but it isn't needed. In the "Another way to skin it" version just above, the waveforms going into the phase splitter will be virtually identical to the original.