Since all of the reverb schemes I've seen have the signal pass thru a large resistor (3M3 in Blackface Fenders, for example) in order to divide off some signal for the reverb circuit, I was thinking of an alternative approach.
Instead of the classic Blackface gain stage > V/T/B > gain stage > divider resistor, I was thinking that the signal could go into both sides of a 12AX7: one for the second gain stage-dry, and one to feed the reverb driver tube. That way, the dry signal never has to pass thru that huge resistor, and the level of that pre-reverb gain can be tailored to taste.
Any thoughts?
Reverb idea--opinions?
Moderators: pompeiisneaks, Colossal
Reverb idea--opinions?
Rich Gordon
www.myspace.com/bigboyamplifiers
"The takers get the honey, the givers get the blues." --Robin Trower
www.myspace.com/bigboyamplifiers
"The takers get the honey, the givers get the blues." --Robin Trower
Re: Reverb idea--opinions?
Aharon
Re: Reverb idea--opinions?
Interesting schematic from Aharon!
Despite what I will say here, it would be really interesting to get more examples of reverb circuits in the sprit of this thread's original question.
Speaking of blackface or silverface Fenders, you can also do the reverb circuit like the schematic I’ve posted here.
Many people don't like the sound of the 3M3/10p solution in blackface and silverface Fenders and think that the Normal channel sounds better, especially with distortion. I think that modifying the reverb circuit according to the schematic here makes a good improvement (if you think that the Normal channel gives a good sound).
The main goal with this reverb circuit is to make the amp sound a little fuller and not so trebly and scooped in the middle that it is with Fender's original 3M3/10p solution.
I also think that big series resistances makes the sound less clear, less dynamic and that the guitar tone is “camouflaged” more. With no or small series resistances I think that the true tone of the guitar and what you make with the strings when you play is more audible and comes thru more.
This mod uses a ECC832 or 12DW7 tube. That’s a double triode, just like the 12AX7, but one of the triodes is like in a 12AX7 and the other is like in a 12AU7. The 12AU7 part is used for the mixing of the dry and wet signals and as it has much less gain than a 12AX7 triode the signal doesn’t have to be lowered with a big series resistor. As you can see in the schematic I’ve used 180k with very good results.
JJ Electronics makes ECC832 today. Remark that if you mod a black-/silverface Fender like this you have to move most of the wires from one of the triode parts to the other.
By choosing tubes from different makers and/or different tubes of the same type I’ve managed to get the dry sound of the Vibrato channel to sound very much like the Normal channel. The wet reverb sound is not changed that much compared to the original circuit.
The dry sound gets a little louder with this mod so you have to set the Reverb control a little higher than before.
Take care,
Sven-Johan
PS: I have tweaked the curcuit a bit but I like to give credit to the person who originally gave me the idea and a circuit solution with the ECC832 (12DW7) in the reverb circuit. It's an amp tech friend of mine called Lars Söderqvist of reVintage Amps here in Sweden. Here’s his web site: http://www.revintage.se/
Despite what I will say here, it would be really interesting to get more examples of reverb circuits in the sprit of this thread's original question.
Speaking of blackface or silverface Fenders, you can also do the reverb circuit like the schematic I’ve posted here.
Many people don't like the sound of the 3M3/10p solution in blackface and silverface Fenders and think that the Normal channel sounds better, especially with distortion. I think that modifying the reverb circuit according to the schematic here makes a good improvement (if you think that the Normal channel gives a good sound).
The main goal with this reverb circuit is to make the amp sound a little fuller and not so trebly and scooped in the middle that it is with Fender's original 3M3/10p solution.
I also think that big series resistances makes the sound less clear, less dynamic and that the guitar tone is “camouflaged” more. With no or small series resistances I think that the true tone of the guitar and what you make with the strings when you play is more audible and comes thru more.
This mod uses a ECC832 or 12DW7 tube. That’s a double triode, just like the 12AX7, but one of the triodes is like in a 12AX7 and the other is like in a 12AU7. The 12AU7 part is used for the mixing of the dry and wet signals and as it has much less gain than a 12AX7 triode the signal doesn’t have to be lowered with a big series resistor. As you can see in the schematic I’ve used 180k with very good results.
JJ Electronics makes ECC832 today. Remark that if you mod a black-/silverface Fender like this you have to move most of the wires from one of the triode parts to the other.
By choosing tubes from different makers and/or different tubes of the same type I’ve managed to get the dry sound of the Vibrato channel to sound very much like the Normal channel. The wet reverb sound is not changed that much compared to the original circuit.
The dry sound gets a little louder with this mod so you have to set the Reverb control a little higher than before.
Take care,
Sven-Johan
PS: I have tweaked the curcuit a bit but I like to give credit to the person who originally gave me the idea and a circuit solution with the ECC832 (12DW7) in the reverb circuit. It's an amp tech friend of mine called Lars Söderqvist of reVintage Amps here in Sweden. Here’s his web site: http://www.revintage.se/
You do not have the required permissions to view the files attached to this post.
Re: Reverb idea--opinions?
Aharon,
Thanks for the post. That's the general idea, but with a gain stage/tonestack ahead of the twin triodes. I drew my idea out last night, and I'll compare it to what you posted.
Thanks again.
Thanks for the post. That's the general idea, but with a gain stage/tonestack ahead of the twin triodes. I drew my idea out last night, and I'll compare it to what you posted.
Thanks again.
Rich Gordon
www.myspace.com/bigboyamplifiers
"The takers get the honey, the givers get the blues." --Robin Trower
www.myspace.com/bigboyamplifiers
"The takers get the honey, the givers get the blues." --Robin Trower
Re: Reverb idea--opinions?
Glad it helped a little.I'm also looking for an "better" "easier" reverb. 
Aharon
Aharon
Aharon
Re: Reverb idea--opinions?
The first reverb circuit is the same one Hoffman Amps has been using for years to add reverb to tweed style Fender amps. It's posted on his website.
TT
TT
Re: Reverb idea--opinions?
Of course it is...it says so on the page....Hoffman Amps......... ???!!
Aharon
Aharon
Aharon
Re: Reverb idea--opinions?
> signal pass thru a large resistor (3M3 in Blackface Fenders, for example) in order to divide off some signal for the reverb circuit
Over-simplified.
To add reverb we need a mixer. Mixers have loss. For a minimum loss 2-in mixer, loss is 2:1. If adapting an existing happy signal chain, we must add another low-gain stage to make-up the mix loss. Gain about 2.
An extra tube is more gain than we strictly need to make-up mix loss.
Reverb recovery needs gain, and lots of it.
So the usual Fender (ie AA1069) plan is-
Straight side of signal fed through 15:1 lossy mix input. Fender's usual mix resistor was 220K-270K, so he upped the resistor to 3M3.
Reverb recovery gets one stage gain, then feeds the other mix input. Because the straight input has been un-optimized to give high loss, the reverb input can have loss near 1.1:1 (but often more in Fenders).
The mixed signal feeds an added stage, loaded with the vibrato pot, with gain about 20. This gives equivalent straight-channel gain of 20/15= ~1, and up to extra-gain of 20 on the reverb side.
The high-Meg resistor "shouldn't" be a problem. Except you need to fix-up the stray capacitances. That's what the 10pFd cap does. Depending on layout, you may need a different value.
I'd ask what effect the extra stage has. _I_ suspect that a 3-stage Champ is kinda "just right", the added PI stage in push-pull amps is a necessary evil, and adding more stages to a guitar amp complicates the small-signal distortion quality. Less is best?
To lose the 3Meg resistor: use a "T" network. 220K input, 22K to ground, 220K mix network. This should give the same gain to the straight channel (fiddle the 22K for exact same gain). This will load the reverb signal a bit, AA1069 has a 470K/220K pad after the Reverb pot which you can tinker (swap the 220K and 470K resistors).
Now you lost the 3Meg but still have the added stage. How does that sound?
I'm especially fretty about that added stage because in AA1069 and Twin Reverb AB763, it shares a cathode network with the reverb recovery amp. Coupling between two sequential cathodes tends to be positive feedback. It is reduced by the mix-loss, and supposedly suppressed by the cathode cap and by reverb pot coupling cap, but it's a cheap trick and not Good Practice. You might simply check if your plan also has this shared cathode network and break it up.
These amps also have a heavy load on the Vibe channel mix-loss make-up stage. 12AX7 with 100K plate resistor and 50K vibrato intensity pot. The poor AX7 is looking into a load less than half its plate resistance.
> Normal channel sounds better, especially with distortion.
I'd have to wonder about that extra stage. Signal level is not high, but it has everything else going against it. Heavy loading and contaminated cathode. To get the same level at the PI, the third stage of the Vibe channel has to swing 3 times the current as the second stage of the Norm channel.
Sven-Johan's plan is interesting. The 12AU7 has less gain, so needs less loss in the straight side of the mixer. He's modded the 470K:220K loss after the Reverb pot, but it may wind up less reverb effect gain than before. If you never turn REV to 10, cool. If you surf-out with infinite reverb, this may disappoint. The 12AU7 is better able to drive the low impedance of the vibrato scheme, but the output of the 12AU7 is lower impedance than the AX7, so the vibrato has to work harder to get the same depth.
Fender may have done it their way so that, in the store, you always could get plenty, even too much, of every effect. "Wow! Gotta buy this!" You are not likely to notice the subtle difference of flavor from the extra stage and convoluted loss/gain system until you've played it a while. And maybe today's players are more critical than we were in 1970.
Too many gimmicks. If you could lose the Vibrato, and accept a small loss of gain, it would be simple to wire two straight channels to the PI via 220K resistors, use a 2-stage reverb recovery amp and dump that to the PI via a 470K or 680K resistor. Same number of tubes, we just moved one to reverb-only. But now there is no obvious place for the vibrato to suck upon. (Except the power tube grids, which is IMHO the only place trem should be done; but then you have trem on the "Norm" channel.)
Over-simplified.
To add reverb we need a mixer. Mixers have loss. For a minimum loss 2-in mixer, loss is 2:1. If adapting an existing happy signal chain, we must add another low-gain stage to make-up the mix loss. Gain about 2.
An extra tube is more gain than we strictly need to make-up mix loss.
Reverb recovery needs gain, and lots of it.
So the usual Fender (ie AA1069) plan is-
Straight side of signal fed through 15:1 lossy mix input. Fender's usual mix resistor was 220K-270K, so he upped the resistor to 3M3.
Reverb recovery gets one stage gain, then feeds the other mix input. Because the straight input has been un-optimized to give high loss, the reverb input can have loss near 1.1:1 (but often more in Fenders).
The mixed signal feeds an added stage, loaded with the vibrato pot, with gain about 20. This gives equivalent straight-channel gain of 20/15= ~1, and up to extra-gain of 20 on the reverb side.
The high-Meg resistor "shouldn't" be a problem. Except you need to fix-up the stray capacitances. That's what the 10pFd cap does. Depending on layout, you may need a different value.
I'd ask what effect the extra stage has. _I_ suspect that a 3-stage Champ is kinda "just right", the added PI stage in push-pull amps is a necessary evil, and adding more stages to a guitar amp complicates the small-signal distortion quality. Less is best?
To lose the 3Meg resistor: use a "T" network. 220K input, 22K to ground, 220K mix network. This should give the same gain to the straight channel (fiddle the 22K for exact same gain). This will load the reverb signal a bit, AA1069 has a 470K/220K pad after the Reverb pot which you can tinker (swap the 220K and 470K resistors).
Now you lost the 3Meg but still have the added stage. How does that sound?
I'm especially fretty about that added stage because in AA1069 and Twin Reverb AB763, it shares a cathode network with the reverb recovery amp. Coupling between two sequential cathodes tends to be positive feedback. It is reduced by the mix-loss, and supposedly suppressed by the cathode cap and by reverb pot coupling cap, but it's a cheap trick and not Good Practice. You might simply check if your plan also has this shared cathode network and break it up.
These amps also have a heavy load on the Vibe channel mix-loss make-up stage. 12AX7 with 100K plate resistor and 50K vibrato intensity pot. The poor AX7 is looking into a load less than half its plate resistance.
> Normal channel sounds better, especially with distortion.
I'd have to wonder about that extra stage. Signal level is not high, but it has everything else going against it. Heavy loading and contaminated cathode. To get the same level at the PI, the third stage of the Vibe channel has to swing 3 times the current as the second stage of the Norm channel.
Sven-Johan's plan is interesting. The 12AU7 has less gain, so needs less loss in the straight side of the mixer. He's modded the 470K:220K loss after the Reverb pot, but it may wind up less reverb effect gain than before. If you never turn REV to 10, cool. If you surf-out with infinite reverb, this may disappoint. The 12AU7 is better able to drive the low impedance of the vibrato scheme, but the output of the 12AU7 is lower impedance than the AX7, so the vibrato has to work harder to get the same depth.
Fender may have done it their way so that, in the store, you always could get plenty, even too much, of every effect. "Wow! Gotta buy this!" You are not likely to notice the subtle difference of flavor from the extra stage and convoluted loss/gain system until you've played it a while. And maybe today's players are more critical than we were in 1970.
Too many gimmicks. If you could lose the Vibrato, and accept a small loss of gain, it would be simple to wire two straight channels to the PI via 220K resistors, use a 2-stage reverb recovery amp and dump that to the PI via a 470K or 680K resistor. Same number of tubes, we just moved one to reverb-only. But now there is no obvious place for the vibrato to suck upon. (Except the power tube grids, which is IMHO the only place trem should be done; but then you have trem on the "Norm" channel.)
Re: Reverb idea--opinions?
Maybe this thread isn't about the type of reverb in Fender black- and silverface amps, as I thought...
Despite that, I have some questions about the 18 Watt Stout Reverb:
How does it affect the dry sound? Isn't that the important question? As I understand it this thread’s initial question was how to make a reverb circuit that affects the dry sound as little as possible.
How well does the Mix control work? I may be wrong, but I have easy to think that the overall sound volume depends a little on how it’s set. If you turn up the Mix control for more reverb, is the perceived overall volume then the same or do you want to turn the Volume control to compensate?
If anyone have tested the 18 Watt Stout Reverb it would be nice to get some information about its sound. Both the dry and the wet sound.
Take care,
Sven-Johan
Despite that, I have some questions about the 18 Watt Stout Reverb:
How does it affect the dry sound? Isn't that the important question? As I understand it this thread’s initial question was how to make a reverb circuit that affects the dry sound as little as possible.
How well does the Mix control work? I may be wrong, but I have easy to think that the overall sound volume depends a little on how it’s set. If you turn up the Mix control for more reverb, is the perceived overall volume then the same or do you want to turn the Volume control to compensate?
If anyone have tested the 18 Watt Stout Reverb it would be nice to get some information about its sound. Both the dry and the wet sound.
Take care,
Sven-Johan
Re: Reverb idea--opinions?
Isn't that a kind of version of the, on this forum so called "one tube reberb"?PRR wrote:...Too many gimmicks. If you could lose the Vibrato, and accept a small loss of gain, it would be simple to wire two straight channels to the PI via 220K resistors, use a 2-stage reverb recovery amp and dump that to the PI via a 470K or 680K resistor. Same number of tubes, we just moved one to reverb-only...
Do you mean almost like in the schematic I've posted down here? That's my own thoughts of that kind of reverb, but I haven't tested or tweaked my own circuit yet...
I don't like how I have drawn the tremolo, but I've tried to be honest to Fender's solutions and have done it like it is on the blackface amps without reverb.
Take care,
Sven-Johan
You do not have the required permissions to view the files attached to this post.
Re: Reverb idea--opinions?
> like in the schematic I've posted down here?
Interesting.
You do get Trem on the reverb, but on the Send. The reverb return trem will be out of sync with the trem on the straight path, delayed by reverb time. The trem-dips in the straight path will have reverb in them. Might be fun.
In short: you can have a simple "pure" amp, or a lot of frills. Adding either reverb or trem is a minor compromise with purity; adding both seems to lead to "Many people ...think that the Normal channel sounds better".
So while the "3Meg problem" is in the reverb, on those amps with photocell tremolo the fix has to include the trem too.
But to return to your plan:
The shared cathode R-C network should, IMHO, be broken-up.... we are not out to shave every penny like a manufacturer has to. At least leave room in the layout for two cathode networks. There may be "a sound" in the shared cathode; but I think separate networks are more stable and more versatle.
Reverb hiss is a perennial problem; I'd be inclined to lose the 330K+180K pad between reverb recovery stages. That will leave too much gain; let the second stage cathode resistor be un-bypassed, then increase the size of the reverb mix resistor from 220K to 680K or more. The larger mix resistor reduces the added loss on the straight channels. The two casecaded stages won't gross-overload because you have the reverb pot between them. If you want reverb to be as large as (or larger than) the straight signal, at high power levels, we need to ask if the reverb amp can saturate the output stage through this large loss. With 350V supply and 220K+220K+680K mixer, it surely can put 10V into the PI which only needs like 2V to saturate. But I'm sure the values need tweaking and it could wind up marginal.
Another, blasphemous, approach is to isolate the reverb and trem functions between some high-purity (i.e. "chip") buffers. Possibly with hard bypass (relays) so that when each FX function is disengaged, the sound runs straight from tube to tube... the SS pollution only happening when effects are in use.
> interesting to get more examples of reverb circuits in the sprit of this thread's original question.
Fender invented (a dozen kinds of) Tremolo, and also pioneered Reverb. And many other makers "studied" (or just stole) Fender's ideas.
Ampeg has some variations. Try the 1971 VT40. This IS different. Channel mixing is done on a shared 68K plate resistor, both channels run thru a James tone control. Straight path runs through V201 which is an over-elaborate gain-of-10 amp with a Mid control. Reverb bypasses this and mixes back with the 180K+270K network at V3 pin 2. There's no Tremolo.
ReverbRocket is a one channel rev+trem amp. Trem is applied early (and opposite to Fender: lamp bright makes sound louder). Straight and reverb are mixed with 2.2Meg and 150K, not very different from AB763. Huh.... I never noticed that: this amp has two trem oscillators. One just for reverb.
Interesting.
You do get Trem on the reverb, but on the Send. The reverb return trem will be out of sync with the trem on the straight path, delayed by reverb time. The trem-dips in the straight path will have reverb in them. Might be fun.
In short: you can have a simple "pure" amp, or a lot of frills. Adding either reverb or trem is a minor compromise with purity; adding both seems to lead to "Many people ...think that the Normal channel sounds better".
So while the "3Meg problem" is in the reverb, on those amps with photocell tremolo the fix has to include the trem too.
But to return to your plan:
The shared cathode R-C network should, IMHO, be broken-up.... we are not out to shave every penny like a manufacturer has to. At least leave room in the layout for two cathode networks. There may be "a sound" in the shared cathode; but I think separate networks are more stable and more versatle.
Reverb hiss is a perennial problem; I'd be inclined to lose the 330K+180K pad between reverb recovery stages. That will leave too much gain; let the second stage cathode resistor be un-bypassed, then increase the size of the reverb mix resistor from 220K to 680K or more. The larger mix resistor reduces the added loss on the straight channels. The two casecaded stages won't gross-overload because you have the reverb pot between them. If you want reverb to be as large as (or larger than) the straight signal, at high power levels, we need to ask if the reverb amp can saturate the output stage through this large loss. With 350V supply and 220K+220K+680K mixer, it surely can put 10V into the PI which only needs like 2V to saturate. But I'm sure the values need tweaking and it could wind up marginal.
Another, blasphemous, approach is to isolate the reverb and trem functions between some high-purity (i.e. "chip") buffers. Possibly with hard bypass (relays) so that when each FX function is disengaged, the sound runs straight from tube to tube... the SS pollution only happening when effects are in use.
> interesting to get more examples of reverb circuits in the sprit of this thread's original question.
Fender invented (a dozen kinds of) Tremolo, and also pioneered Reverb. And many other makers "studied" (or just stole) Fender's ideas.
Ampeg has some variations. Try the 1971 VT40. This IS different. Channel mixing is done on a shared 68K plate resistor, both channels run thru a James tone control. Straight path runs through V201 which is an over-elaborate gain-of-10 amp with a Mid control. Reverb bypasses this and mixes back with the 180K+270K network at V3 pin 2. There's no Tremolo.
ReverbRocket is a one channel rev+trem amp. Trem is applied early (and opposite to Fender: lamp bright makes sound louder). Straight and reverb are mixed with 2.2Meg and 150K, not very different from AB763. Huh.... I never noticed that: this amp has two trem oscillators. One just for reverb.
Re: Reverb idea--opinions?
As I said I haven’t tested or tweaked the SJB-Mod myself (yet). I drew this schematic just as a basis for discussion on another forum than this. One of its purposes (in the other discussion) was to make the mod as simple as possible in a black- or silverface Fender amp.
The circuit isn’t too different from the so called “one tube reverb” that has been discussed on the Dumble forums here on Amp Garage. (I may be wrong, but I have easy to think that those who have tried versions of the “one tube reverb” are not totally satisfied).
For myself I am very satisfied with the sound from Fender blackface amps with the LS-Mod that I posted the schematic for before.
As I am a tube amp tech I have tried many amps with spring reverb, both vintage and new, but still haven’t found one where the reverb sounds better than the Fender blackface circuit. That’s, of course, my own, personal opinion. The LS-Mod doesn’t affect the wet sound much but makes the dry sound even better (still my own, personal opinion).
I agree that the sound becomes better with separate resistors and (where appropriate) decoupling capacitors on each cathode. I have done modification for that in many, many amps. (That is when each triode is for different amp stages. With paralleled triodes it’s another thing, as I see it).
The 330K+180K pad between reverb recovery stages is just to balance the wet with the dry sound, and to make the reverb control to work in a good way. But, as I said, the circuit isn’t tweaked or calculated, so there are certainly things to do here.
Take care,
Sven-Johan
PS: Probably the best spring reverb sound I've ever heard is in one of my own amps. It’s based on a modified LS-Mod and has a pentode as the driver, bigger transformer that makes a good impedance match between the pentode and the tank (which is a six spring Accutronics tank)...
The circuit isn’t too different from the so called “one tube reverb” that has been discussed on the Dumble forums here on Amp Garage. (I may be wrong, but I have easy to think that those who have tried versions of the “one tube reverb” are not totally satisfied).
For myself I am very satisfied with the sound from Fender blackface amps with the LS-Mod that I posted the schematic for before.
As I am a tube amp tech I have tried many amps with spring reverb, both vintage and new, but still haven’t found one where the reverb sounds better than the Fender blackface circuit. That’s, of course, my own, personal opinion. The LS-Mod doesn’t affect the wet sound much but makes the dry sound even better (still my own, personal opinion).
I agree that the sound becomes better with separate resistors and (where appropriate) decoupling capacitors on each cathode. I have done modification for that in many, many amps. (That is when each triode is for different amp stages. With paralleled triodes it’s another thing, as I see it).
The 330K+180K pad between reverb recovery stages is just to balance the wet with the dry sound, and to make the reverb control to work in a good way. But, as I said, the circuit isn’t tweaked or calculated, so there are certainly things to do here.
Take care,
Sven-Johan
PS: Probably the best spring reverb sound I've ever heard is in one of my own amps. It’s based on a modified LS-Mod and has a pentode as the driver, bigger transformer that makes a good impedance match between the pentode and the tank (which is a six spring Accutronics tank)...