Tube tremolo with 12AX7(high voltage)

[gtudoran]

Hey guys,

Just finished to bread-board the full tube / high-voltage tremolo and it sounds like this (very happy with the result i might say).
The schematic was modified and an updated schematic will be available soon. If you like to jump directly to action you can go to min. 2:13 but if you like to hear the explanations you can listen to the whole clip.

Schematic (the one on the clip was modified):


Video:

Regards,
DeX

[Kipper4]

Really nice sound Dexter.
I might just borrow those lfo values for a future transistor pso based tremolo.
Thanks for sharing
Rich

[gtudoran]

Thank you Rich, the values are quite "random" and they will dictate the max freq. available. Also in this case there is a soft-start of the LFO (the reason for the LFO to oscillate is in fact the inherit noise).
Long story short - please do use them and keep me updated. I'm quite interested to see how it goes with a transistor.

Regards,
DeX

[PRR]

C4 100uFd smells high. It damps the LFO signal above say 2Hz, which includes much trem-speed zone. I'd be wondering about values as small as 22uFd, though this invites other trade-offs.

Lot of gain there also.

[gtudoran]

Thank you PRR for your reply. The schematic is not definitive and suffered some modifications (i didn't have time to put it down but i'm working on it).
The cathode of the LFO right now it's bypassed with a 22uF if i remember well  also there is a modification in the anode of the amplification stage.
I wanted to have a little more gain because i would like to use this as a booster also (indeed there is too much gain for now). Also i think i will change C6 with a 22 or 47n as 100n it seems to me to be a little bit too much.

Regards,
DeX

C4 100uFd smells high. It damps the LFO signal above say 2Hz, which includes much trem-speed zone. I'd be wondering about values as small as 22uFd, though this invites other trade-offs.

Lot of gain there also.

[thermionix]

Sounds good, but fairly slow overall compared to most onboard amp trems.  Nothing wrong with that, if that's what you're going for.  But if you wanted to speed it up, decrease the value of any or all of C1, C2, and C3.  Very common to see .02 .01 .01 (mfd).  In my Fender-type amp I think I'm using .03 .01 .01, just for a bit slower than usual.  Still gets plenty fast.

[gtudoran]

Thermionix - indeed it's slower then the amps tremolo and that is what i aimed for. I found a little bit useless the last 20-25% of the speed knob (too fast for my taste and it sounds quite ... unpleasant to my ears). Of course all in all is a matter of taste but i do like the tremllo to be a little more lush and slow-speed.
I've tried with 22n all the way and it was a little bit too much for my taste - but it's a tube device and modifying it should be quite easy.
I hope that i will be able to stick it in a 1590B type enclosure (along with FS).

Regards,
DeX

Sounds good, but fairly slow overall compared to most onboard amp trems.  Nothing wrong with that, if that's what you're going for.  But if you wanted to speed it up, decrease the value of any or all of C1, C2, and C3.  Very common to see .02 .01 .01 (mfd).  In my Fender-type amp I think I'm using .03 .01 .01, just for a bit slower than usual.  Still gets plenty fast.

[thermionix]

Roger that, dig.  As you know, there are other component values that affect speed, but those 3 series caps are the quick and easy way, at least to me.  And they also make it easy to find the LFO circuit if you're opening up an unfamiliar amp with no schematic handy.  Valco, anyone?  Ugh!

[idiot savant]

Cool trem. Reminds me of Rick Campbell's trem-o-drive a bit. Check it out for a good way to add a depth control.

http://www.ricktone.com/tremodrive.shtml

[gtudoran]

(really don't know how to call you ...) That is what i was trying to avoid Using a potentiometer with DC potential, that is why i've choose to use a small toggle switch with 2 depth presets ... especially that the range of the depth pot is not so... great.

Regards,
DeX

Cool trem. Reminds me of Rick Campbell's trem-o-drive a bit. Check it out for a good way to add a depth control.

http://www.ricktone.com/tremodrive.shtml

[thermionix]

In the schematic, the two cathodes are tied together.  Is this an error?

[PRR]

> cathodes are tied together.  Is this an error?

Makes sense. Some wobble appears at V1 LFO cathode. Inject this to V2 cathode, audio wobbles.

There's signs it has not been optimized for profit production. R1 R7, one is redundant, a single 620 Ohms would do both jobs. However in other cathode-couple trems, we often wind up with some or more parts *between* cathodes, so keeping them split makes sense for development and one-off builds.

Connection of R3 to cathode has no purpose here. It is probably taken from an LFO with an OFF switch. Tying one point in the C-R-C-R-C-R network to a DC voltage can "kick-start" the slow-start LFO. That's a feature which might be added later, so may as well leave it connected as shown.

I still think that is a lot of gain. If B+ is over 100V and signal levels are guitar-size, R6 could perhaps be reduced to 22K even 10K for gain under 10 instead of near 50. THD will rise a bit but nothing to offend an e-guitarist.

There is a lot of sub-sonic wobble in the audio output. Some next-devices will be OK, others will be slammed into rude sounds. It is generally "polite" to not have large sub-sonics on an audio output. The 100nFd+500K load is flat to 3Hz, so doesn't cut much. Also 500K pot can be a fairly large output impedance. I'd look to a several-stage C-R-C-R low-cut filter with R near 100K and C say 0.02uFd (22nFd). The 100K||100K loading also drops-down the tube's high gain.

But low-level trem injection seems generally problematic. You have to really beat the audio tube to get significant trem, which is not so easy; and that heavy LFO signal causes troubles downstream. Push-pull injection at high level is generally smoother, but FAR more expensive (and no 1590B).

[thermionix]

^Yeah what he said.   

I'm used to seeing one triode anode follower-connected to modulation point, or two triodes in LFO/CF configuration.

But I don't consider myself an expert by any means.  When I build trem into an amp, I just copy circuits that I already know I like, and will function.  Adapt as necessary.

[gtudoran]

Well... again PRR offered the most complete explanation possible.

- indeed the cathode resistors could be turned from 2 to 1 ... but i like to keep them separate (not for a higher purpose reason but for the "visual" flow and adding another 2c would not be a problem even in a run of 100 units and i doubt that i will sell so much)
- i will try to add some DC potential to kick-start the LFO, right now it takes about 2 seconds
- and yes indeed the signal level is quite high, that's why i've said that the schematic is not quite updated with the last modifications, right now i'm using 50K for the anode resistor but yes, i could go way way lower then 50.
- the output stage at the moment is buffered using a high-voltage fet and indeed 100nF seemed a little too boomy. The output filter seems like an excellent idea - i will try it ASAP.

And yes there are other methods that can be more effective, but not inside of a pedal. I think the final result turned out quite ok and also there is the advantage of having a true tube tremolo inside a pedal (not to mention the booster feature).

Thank you again PRR for taking the time to write such a complete explanation.

Best regards,
DeX

[merlinb]

This sort of tremolo works best with a ECC81 / 12AT7. It is more non-linear so you get a deeper trem, and it's voltage gain is a bit lower which is also the direction you're going. Yes, this tube works for the oscillator too.

[duck_arse]

is a 6j6 a candidate for this type circuit?

[gtudoran]

The common cathode could be a drawback and as far as i can see it's a VHF tube and i don't think it would be ok in audio domain ... at least not for classical applications.

Regards,
Dex

is a 6j6 a candidate for this type circuit?

[PRR]

6J6 is a fine audio tube (WHEN you can use the common cathode).

It is in fact a derivative of an old octal twin-triode which was used for large class-B power and also as a driver and general amplifier.

However the Mu may not be enough for the phase-shift oscillator. We need final gain of >27, 6J6 is IIRC 36. Even the 12AT7 Mu=60 can be tricky.

12AU7 is sometimes a better gain-controlled stage. Won't self-wobble with just a CRCRCR network. There is a tube which is half a 12AX7 and half a 12AU7. Has been scarce but I think someone is new-making these. Dang, I even remember the name: 7247. Not dirt cheap.

[vigilante397]

This thread has several of my favorite people discussing things I find very interesting, so I am commenting in order for it to show up in my "replies to your posts." Also, Dex, big fan of everything you've ever done (been stalking you since GTFO first popped up), including this build. Really great sound.

Carry on.

Real quick off-topic: Merlin, Designing Preamps was an amazing resource for a research paper I did on tubes for an engineering class. Really excellent book, thanks for all your work. Cheers.

[gtudoran]

@PRR
HA! Didn't know about it ... i will look into it. A little bit tricky with the common cathode but i do think that can be a very nice parallel input stage for a tube amp.
The heptal base ... it's not standard but i do think that it's something you can accept.

@Nathan - you are way too kind brother ... way too kind. Thank you a lot.

Regards,
DeX

6J6 is a fine audio tube (WHEN you can use the common cathode).

It is in fact a derivative of an old octal twin-triode which was used for large class-B power and also as a driver and general amplifier.

However the Mu may not be enough for the phase-shift oscillator. We need final gain of >27, 6J6 is IIRC 36. Even the 12AT7 Mu=60 can be tricky.

12AU7 is sometimes a better gain-controlled stage. Won't self-wobble with just a CRCRCR network. There is a tube which is half a 12AX7 and half a 12AU7. Has been scarce but I think someone is new-making these. Dang, I even remember the name: 7247. Not dirt cheap.