One-Tube Reverb instead of Fender's, in an AB763

[guitylerham]

Hey everyone,

Here's my second build of this style of amp. Had an old thread but since this is a new build, figured I'd start a new conversation. Built an AB763 Vibrolux Reverb style amp this time but with Tubenit's "one-tube reverb" and went with a compact 4 preamp tubes total. V1A/B first gain stages, V2A reverb driver, V2B reverb recovery, V3A 3rd gain mixer stage, V3B Sluckey's opto trem, V4A/B LTPI, into 6l6's.

I've been reading and trying to figure this out, even after my first similar DR style build was successful enough, I'm still not certain I had a basically stock AB763 circuit but with just a "one-tube reverb". Wise and experienced people suggest omitting the 3rd gain stage altogether and just inserting the reverb between V1B and the LTPI but I'm hesitant to throw out a whole stage so my idea was to finesse the 3.3M/10pf resistor between reverb send/return and the reverb pot value/voltage divider resistors but I am a bit perplexed.

If my goal is to leave the 3.3M resistor and have it dump the 90+% dry signal to ground before that 3rd mixer stage, and then to insert the new one-tube reverb, can I simply keep the stock 100k reverb pot +470k/220k divider?

Or, like Tubenit has mentioned, perhaps I skip all of that and just insert his circuit just before the LTPI... but then I'd really have a redundant gain stage. My goal was just to keep the stock ab763 circuit gain structure the same and then from there I could up it if I wanted (knowing the difference in gain between stock fender Normal and Vibrato channels and that some prefer the higher gain Normal Ch).

Feels like I'm beating my head against a... dead horse but I feel I need to better understand voltage dividers and how the reverb send/returns work and I'm looking for others who might shed a new perspective. Thank you in advance. As always, this place is a blessing.

Geeez, can't figure out how to post a decent picture. Guess I'm old!

I've attached a schematic of Hoffman's AB763 for reference.


One-tube reverb

[pdf64]

...If my goal is to leave the 3.3M resistor and have it dump the 90+% dry signal to ground before that 3rd mixer stage, and then to insert the new one-tube reverb, can I simply keep the stock 100k reverb pot +470k/220k divider? ...

If you want the classic Fender reverb, and the extra kick of gain available that the reverb channel provides, then that's a good way forward. Especially if you're a 'plug straight in' kinda player.
But if you're going to be using the amp as a clean platform, eg to use with a pedalboard, then getting rid of that 3rd gain stage could work out fine.
Your images were too small to be much use.

[tubenit]

So, is the amp built with the one tube reverb?  IF so, how's it sound and what is it doing that may lead you to change what you've done??

I am NOT sure what you're question is?  With the Fender reverb the insertion points have a 3.3M, 100k pot and a 470k to separate R and R2.     With the "one tube reverb", you have a 1M pot, 330k and 220k between the insertion points.

I modified an ExpressSCH schematic for you, to attempt to post a comparison.  This schematic is editable with the FREE ExpressSCH program.    Please note that I am not knowledgeable enough to tell you where to insert the Vactrol one triode tremolo.  Perhaps Sluckey can advise on that?

Can you give us more information about the amp, reverb, concern and give a defined clear question that you want answered?

With respect, Tubenit

FYI,  it is NOT my one tube reverb!  The Dumble clone guys were using that for quite some time before I discovered it.

[echuta13]

Maybe check out this article?  This one looks similar to what Swart does: https://guitar.com/guides/diy-workshop/build-tube-spring-reverb-unit-amplifier/

[pdf64]

...where to insert the Vactrol one triode tremolo...

I think the idea is to use the LED element of the vactrol as the cathode bias for the trem oscillator, and the optoresistor element just the same as a regular BF trem channel, on a 50K trem intensity control pot at the end of the trem channel. Brilliant! 

[tubenit]

Here is a web page of Sluckey's that has the schematic of how he used the Vactrol tremolo on an AB763 Bandmaster.  There is a PDF file there showing both the layout and the schematic.

http://sluckeyamps.com/trem/trem.htm

[guitylerham]

...If my goal is to leave the 3.3M resistor and have it dump the 90+% dry signal to ground before that 3rd mixer stage, and then to insert the new one-tube reverb, can I simply keep the stock 100k reverb pot +470k/220k divider? ...

If you want the classic Fender reverb, and the extra kick of gain available that the reverb channel provides, then that's a good way forward. Especially if you're a 'plug straight in' kinda player.
But if you're going to be using the amp as a clean platform, eg to use with a pedalboard, then getting rid of that 3rd gain stage could work out fine.
Your images were too small to be much use.

Yeah, I apologize for those crappy images, had limited time and internet to fix it last night. I am definitely a plug straight in/boost/OD for solos kinda guy. So I’m all for some more character in my amps. Thank you for the direction!

[guitylerham]

So, is the amp built with the one tube reverb?  IF so, how's it sound and what is it doing that may lead you to change what you've done??

I am NOT sure what you're question is?  With the Fender reverb the insertion points have a 3.3M, 100k pot and a 470k to separate R and R2.     With the "one tube reverb", you have a 1M pot, 330k and 220k between the insertion points.

I modified an ExpressSCH schematic for you, to attempt to post a comparison.  This schematic is editable with the FREE ExpressSCH program.    Please note that I am not knowledgeable enough to tell you where to insert the Vactrol one triode tremolo.  Perhaps Sluckey can advise on that?

Can you give us more information about the amp, reverb, concern and give a defined clear question that you want answered?

With respect, Tubenit

FYI,  it is NOT my one tube reverb!  The Dumble clone guys were using that for quite some time before I discovered it.

Fair enough, I tend to ramble/thought dump sometimes!

I think my question premise is: the stock fender Reverb mixing circuit (3.3m, 100k pot, 470k resistor, 220k resistor in the one channel AB763) separates and balances the dry/wet ratio by severely attenuating the majority of dry signal first, then brings in some amount of wet signal (determined by the 100k Reverb pot and 470k/220k to ground), the small dry signal is now balanced—as we’ve come to recognize as Fender—with the wet signal and fed to the 3rd mixer gain stage to recoup the loss from a 3.3m/220k and Reverb circuit.

Now the Dumble style Reverb (my apologies for misappropriating it to to you, that happens a lot on the interwebs!) to me, looks like it’s more “standalone” and modularly insertable  i.e. maybe doesn’t severely dump dry signal to ground, can be implemented in many amp circuits without causing much of a headache, unlike my assumption of the Fender style Reverb, which requires an additional makeup gain stage after the recovery stage—I’m narrowing in on my question, I swear! How does one alter the one-tube Reverb mixing resistors to be able to achieve a similar result if implemented into a fender circuit while keeping the same gain behavior with that now slightly unnecessary 3rd gain stage in the AB763?

Currently, I’ve yet to fire the amp up but have left several components empty in order do some more research. I’ve built in the “one-tube” verb per your posted scheme and have left out the 3.3m and 220k to ground, planning to install corrected values...

Hope that made a bit more sense. I’m certainly missing some fundamental knowledge but I truly appreciate you all offering your help!

[sluckey]

How does one alter the one-tube Reverb mixing resistors to be able to achieve a similar result if implemented into a fender circuit while keeping the same gain behavior with that now slightly unnecessary 3rd gain stage in the AB763?

Just copy the Fender circuit IAW the original schematic... except, instead of using a parallel 12AT7 for the driver you'll just use 1/2 of the 12AX7. The recovery stage remains the same as the Fender. And the 3.3M with wet/dry mixer stage remains the same.

I would use a 12DW7 instead of a 12AX7.

[guitylerham]

How does one alter the one-tube Reverb mixing resistors to be able to achieve a similar result if implemented into a fender circuit while keeping the same gain behavior with that now slightly unnecessary 3rd gain stage in the AB763?

Just copy the Fender circuit IAW the original schematic... except, instead of using a parallel 12AT7 for the driver you'll just use 1/2 of the 12AX7. The recovery stage remains the same as the Fender. And the 3.3M with wet/dry mixer stage remains the same.

I would use a 12DW7 instead of a 12AX7.

Jeez, that simple, huh? I figured had I tried that, I wouldn’t know what I was comparing the results to, no benchmark to go off of. I’ll try this tonight. I might even have a 12DW7 someplace. Thank you, sir. I’ll report back with some results once I get the time.

[shooter]

Jeez, that simple, huh?

most complex things turn out just so, humans tend not to think simple though 

[d95err]

You could use the regular AB763 reverb driver stage and then use an LND150 mosfet instead of a 12AX7 for the reverb return stage.

Reverb return is just amplifying a very low amplitude signal, so using a solid state stage should have no impact on tone (except maybe be less hum and hiss).

[guitylerham]

You could use the regular AB763 reverb driver stage and then use an LND150 mosfet instead of a 12AX7 for the reverb return stage.

Reverb return is just amplifying a very low amplitude signal, so using a solid state stage should have no impact on tone (except maybe be less hum and hiss).

This is interesting. I’m not against utilizing various technology for a good cause in tube amps but have generally stayed away from solid state in the signal path. It’s not like we get any tube saturation from that Reverb recovery stage and there shouldn’t be noticeable tone imparted by a transistor, it’s not like we don’t put our guitar signal through several of them before even getting to the amp...

I’m going to build the fender circuit and see what that leaves me with. From there I’ll look into implementing your idea. Thank you!

[clbraddock]

I'm in a similar boat as the original poster.

I have an AB763 based amp (Rob Robinette Blackvibe Micro EF80) that I want to add reverb too. My plan was to try the one tube 12DW7 reverb in lieu of the full blown Fender reverb circuit. I'm trying to decide the best way to insert it.

See attached schematic. The options seem to be:

A) Leave everything as is and have the reverb send before the 220K mixing resistor and the reverb return immediately after the 220K mixing resistor prior to the 270K (vib ch load) to ground
B) Leave everything as is and have the reverb send before the 220K mixing resistor and the reverb return immediately after the 270K (vib ch load) to ground (i.e. right before .001 coupling cap going to the phase invertor.
C) Reverb send before 220K mixing resistor. Remove 270K (vib ch load) to ground and Reverb return immediately after 220K mixing resistor - thought process being that the reverb level pot provides a load an path to ground, so the 270K resistor to ground may not be needed.
D) Replace 220K mixing resistor with 3.3M resistor. Reverb send immediately before 3.3M resistor. Return immediately after 3.3M resistor (ala classic AB763) and add additional 12AX7 triode after reverb return - this would basically be copying the original AB763 circuit just using the one tube reverb in place of the AT7 type fender reverb.

I have a spare triode from the tremolo I added. Originally, my thought was to add it as a cold clipper that could be switched in or out early in the circuit (to get a JCM800 effect), but if option D above will have the best clean sound then I can do that. Just not sure if it is necessary with the one tube reverb. Most versions Ive seen posted do not have a large mixing resistor and additional gain stage.

[PRR]

> A) Leave everything as is and have the reverb send before the 220K mixing resistor and the reverb return immediately after the 220K mixing resistor prior to the 270K (vib ch load) to ground
B) Leave everything as is and have the reverb send before the 220K mixing resistor and the reverb return immediately after the 270K (vib ch load) to ground (i.e. right before .001 coupling cap going to the phase invertor.

What is the difference? It is a wire. There is no "before and after".

If I am mis-understanding, please understand that I can't count to three without fingers, so four text descriptions of a map-object sprains my brain. You might go ahead and draw what you are thinking.

[clbraddock]

I was thinking of the instrument input being the beginning and the speaker out being the end. So In that sense, with a beginning and end, there could be a before and after. I can see how that could come across confusing though.

Maybe a better explanation would be my use of “before” means “to the left of” on the schematic and “after” means “to the right of”

I will try to print them out and draw on them though to make it more clear

[sluckey]

What's your point? You built an AB763. Just build the stock reverb circuit and quit beating your head on the floor.

[PRR]

Beating your head about a distinction without a difference. A and B are the same point, electrically.

[clbraddock]

Beating your head about a distinction without a difference. A and B are the same point, electrically.

Ah, I see now. Yeah that was a dumb distinction on my part. Thanks for clarifying for me.

That simplifies it then. I guess question then is just whether the additional recovery tube is needed/beneficial with the one tube circuit

[clbraddock]

What's your point? You built an AB763. Just build the stock reverb circuit and quit beating your head on the floor.

Fully stock circuit would require 2 more tubes. I.e a full 12at7 and a full 12ax7. It would be very tight trying to fit two more tubes in. That’s why I’m wanting to try a one tube 12DW7. I do have 1 triode of a 12ax7 free though, so I could use it as an additional gain stage similar to a ab763.

The question then (which I made overly complicated) is do I keep the smaller 220k mixing resistor and leave off the extra gain stage or do I use a 3.3M mixing resistor and add the additional gain stage. I know the 3.3M with additional gain stage is what a stock AB763 does, but it has a different reverb circuit. None of the successful tubnits circuits use a such a large mixing resistor and/or additional gain stage.

[clbraddock]

I guess I can just build it with the existing 220k mixing resistor and no additional gain stage and see how it sounds. If it’s sucks, it wouldn’t be that hard to swap the resistor and add in the gain stage.

[tubenit]

One option to consider


With respect, Tubenit

[clbraddock]

Thanks Tubenit!

I will connect it like that.

As far as the one tube reverb circuit that is a little different than the schematic I was working off. I was working off the attached. Main difference would be the cathode bias resistors. What you posted has a 2.2K bias resistor on the reverb driver. The one I was working off is 1.5k (and suggests trying 1k).

This is a micro amp, so my max B+ is only 295 VDC (which would drop further going through the reverb transformer). I'd be using a 12DW7, so the driver would be the equivalent of a 12AU7. Based on that do you have any advice concerning the cathode bias resistor value and bypass cap on the reverb driver.

Thanks!

[sluckey]

The driver cathode resistor value is not critical (within reason). I'd probably use a 820Ω to 1.5K. Lower value provides more current to drive the tank. Using a 12DW7, be sure to use pins 1,2,3 for the driver.

[clbraddock]

Finished adding the one tube reverb. My thoughts:

The good: It works.

The bad: Dwell pot doesn’t seem to do anything. Turning intensity to 0 causes a ground loop and sucks out all the highs. Turning it up just a tiny bit and things go back to normal. I need to investigate these issues further.

The subjective: Not sure that I really like it. There is no “sproing” and it almost sounds more like a plate reverb, I.e. dense without many distinct reflections

It’s possible I have something connected wrong, but maybe I would be better off just figuring out a way to cram an AB763 reverb circuit into the space.

Edit: Just realized heater wiring is wrong. Like an idiot forgot to jump pin 4 to pin 5. I will fix that and see if that fixes the problems. The driver tube not functioning would explain the reverb sounding pretty shite. Surprised it was generating reverb at all based on that.

[Williamblake]

One option to consider


With respect, Tubenit

Thanks for sharing. Small glitch at the oscillator i would guess.

[sluckey]

One option to consider


With respect, Tubenit

Thanks for sharing. Small glitch at the oscillator i would guess.

What small glitch?

[pdf64]

A cap is in series with the trem oscillator anode.

[sluckey]

Ahh, yes.

[clbraddock]

Ok, fixed the 12DW7 heater wiring and there’s (obviously) a major improvement. Reverb sounds not to bad now. Maybe not quite as nice as a true AB763 but close enough.

Still having a problem where turning the reverb intensity pot down inversely increases ground noise though. Once reverb pot is at about 4 the noise goes away, but that’s more reverb than I really want.

[tubeswell]

... turning the reverb intensity pot down inversely increases ground noise... Once reverb pot is at about 4 the noise goes away, but that’s more reverb than I really want.

The reverb recovery stage has a high S:N cos it's amplifying a really weak input signal from the pan's output transducer. Some suggestions:

1) Orient the pan so that the output transducer (the one connected to the reverb recovery stage input) is as far away as possible from the EM flux density field produced by the PT
2) Try 'heavy duty' (high quality) pan send/return cables (that have better shielding)
3) try ferrite beads on the pan return cable
4) try a highish (34k - 68k) grid stopper, or a small HF ferrite inductor, or small Cg-k shunt, at the grid pin for the reverb recovery stage
5) Use a high capacitance cap to bypass the recovery stage's cathode resistor (e.g. 220uF), this may help with the S:N ratio
6) improve your lead dress by shortening long run signal wires, keeping wires close to the chassis (to eat up stray EMF), crossing other wires at right angles or keeping 1" away from other wires, and shielding the signal wires to and from the reverb level pot

The amount of wet/dry in the mix can be adjusted by changing the voltage divider that is attenuating the dry signal.

[clbraddock]

Thanks tubeswell, appreciate all of the info. I will try some of what you suggest.

I'm not so sure that it is a S/N or RFI/EMI issue though. With reverb intensity on 0 it very much sounds like what a loose ground connection sounds like. Its weird though since it goes away as the reverb intensity is increased.

On a separate note, I'd still like a bit more drip/sproing? I only have about 290VAC to work with. Is there a tube that I could swap in for the 12DW7 to increase the current to the spring? I know current is what the reverb spring needs, but would using a 12AX7 (more voltage gain) through the reverb transformer (converting the voltage to more current) give me more current to the spring and therefore more drip? Or, perhaps a 6BM6 like some of the other tubnits designs.

Another option is that I have an unused 12ax7 triode already. So, I have 3 triodes to work with if I take out the 12DW7. Would using a parallel 12AT7 or 12AU7 as the driver and then the singe 12AX7 triode as the recovery give me more drip/sproing? This would be sort of like a princeton reverb, except instead of a 3rd preamp triode into a cathodyne phase invertor after the reverb recovery, I would instead have the reverb recovery directly into a long tail pair. My thinking is that the long tail pair could provide the make up gain of the 3rd preamp triode. However, I still have the issue that I only have 290VAC to work with and I believe a princeton is more in the 400VAC range.

[tubeswell]

I'm not so sure that it is a S/N or RFI/EMI issue... With reverb intensity on 0 it very much sounds like what a loose ground connection sounds like. Its weird though since it goes away as the reverb intensity is increased

Well then, check your grounds. How have you grouped the ground returns for the reverb recovery stage? (You should avoid mixing up high current and low current returns).


The pentode section of an ECL82 or ECL86 will give more driving current. The latter type has a higher gain triode stage.

[dblgun]

How does one alter the one-tube Reverb mixing resistors to be able to achieve a similar result if implemented into a fender circuit while keeping the same gain behavior with that now slightly unnecessary 3rd gain stage in the AB763?

Just copy the Fender circuit IAW the original schematic... except, instead of using a parallel 12AT7 for the driver you'll just use 1/2 of the 12AX7. The recovery stage remains the same as the Fender. And the 3.3M with wet/dry mixer stage remains the same.

I would use a 12DW7 instead of a 12AX7.

One question sir. The circuit has about 440vdc on the parallel plates of the 12AT7. If one was to replace this with one of the triodes in a 12AX7 wouldn't the voltage need to be reduced? My tube manual shows a maximum of about 330vdc for the 12AX7.

[sluckey]

I wonder what your tube manual says about the 12AT7?

[thetragichero]

I wonder what your tube manual says about the 12AT7?

oh don't even look at some of those 6V6 while you're at it

[dblgun]

I'm pretty sure you know the answer. The 12AT7 shows a max of 300vdc. I've never had cause to put a consistent 300v to the plates of either, not to mention 400+. I have also never paralleled the plates so was wondering if there was something I missed when it is done.
 

[tubenit]

Ok, fixed the 12DW7 heater wiring and there’s (obviously) a major improvement. Reverb sounds not to bad now. Maybe not quite as nice as a true AB763 but close enough.Still having a problem where turning the reverb intensity pot down inversely increases ground noise though. Once reverb pot is at about 4 the noise goes away, but that’s more reverb than I really want.

It would be helpful to see an exact (not just sort of exact) schematic for your amp. And then a high-resolution photo of the chassis interior.

I don't remember any of the one tube reverbs I've done increased noise as the potentiometer was dialed down. I suspect a soldering or wiring mistake.

with respect, Tubenit

[AmberB]

(The pentode section of an ECL82 or ECL86 will give more driving current. The latter type has a higher gain triode stage.)

It never occurred to me to use an ECL-82/6BM8 or an ECL-86/6GW8 as a reverb driver and recovery tube.  Definitely something to think about for a 1 tube reverb...