Adding Two Tube Reverb to the 5F6A/Marshall Topology (Western Electric Circuit)

Phrygian77

Poster Extraordinaire
Gold Supporter
Joined
Apr 30, 2016
Posts
5,748
Location
Crawfordville, FL
I'm looking at doing something like this on an upcoming 50 watt Marshall-esque 2x12 combo amp build.

t-add-reverb.jpg



This has been floating around for a while, and I'm sure some of you may have seen it posted in discussions about adding reverb over at the el34world.com forum.

I like the approach of using an extra gain stage just for the reverb here, and then mixing with the dry going into the PI. The separate power node also seems to be a good idea. The one thing that really doesn't make sense to me though is the 500pF coupling cap on the final stage. That's going to kill some treble (seems to be about -3dB around 2.3-2.4k) on the dry signal through the mixing resistors and the output impedance of that stage. I don't know why you wouldn't use a larger cap here to keep it from affecting the tone. If the person who drew this up did it intentionally, I don't understand why. You could high pass the preceding stage more if needed. Probably should add some low passing too to reduce hiss.

I don't want to get into a discussion here about simpler circuits with less tubes/stages. I want big full reverb, and I'm not trying to cram this into an existing layout. I'll have plenty of room and the heater current to support it.
 

andrewRneumann

Friend of Leo's
Joined
Mar 22, 2020
Posts
2,027
Location
Cincinnati, OH, USA
I'm looking at doing something like this on an upcoming 50 watt Marshall-esque 2x12 combo amp build.

View attachment 984820


This has been floating around for a while, and I'm sure some of you may have seen it posted in discussions about adding reverb over at the el34world.com forum.

I like the approach of using an extra gain stage just for the reverb here, and then mixing with the dry going into the PI. The separate power node also seems to be a good idea. The one thing that really doesn't make sense to me though is the 500pF coupling cap on the final stage. That's going to kill some treble (seems to be about -3dB around 2.3-2.4k) on the dry signal through the mixing resistors and the output impedance of that stage. I don't know why you wouldn't use a larger cap here to keep it from affecting the tone. If the person who drew this up did it intentionally, I don't understand why. You could high pass the preceding stage more if needed. Probably should add some low passing too to reduce hiss.

I don't want to get into a discussion here about simpler circuits with less tubes/stages. I want big full reverb, and I'm not trying to cram this into an existing layout. I'll have plenty of room and the heater current to support it.

So you would rather have even attenuation across the board, even if it meant more dry signal was dumped in the process? Ie -5db across the board instead of a -2.5db shelf dropping to -5db shelf at higher freqs?
 

Phrygian77

Poster Extraordinaire
Gold Supporter
Joined
Apr 30, 2016
Posts
5,748
Location
Crawfordville, FL
So you would rather have even attenuation across the board, even if it meant more dry signal was dumped in the process? Ie -5db across the board instead of a -2.5db shelf dropping to -5db shelf at higher freqs?

I don't see that as a problem. There has to be some trade off involved in mixing two stages together. In this amp I'm going to an have extra (cascade) preamp gain stage. Think Blenclowe's medium gain preamp ... bypassed input stage, a slightly cold biased stage, a warm biased stage, and a cathode follower. So, I'm going to have a pre PI volume, and possibly a post PI master.
 

Phrygian77

Poster Extraordinaire
Gold Supporter
Joined
Apr 30, 2016
Posts
5,748
Location
Crawfordville, FL
I'm starting to think you guys think I'm crazy.

Hell yeah, buddy. I think all amps should have reverb.

Well, I'm not going to go as far as saying that. My amp that I gig with doesn't have reverb, but I do miss having real spring reverb when it's called for. If you're going to add reverb to an amp, there's no reason to do it for anything other than that spring reverb sound. Pedals just don't cut it. Sure, they can sound good, but no one has made a digital pedal that truly sounds like a spring, and responds like a spring. The tank/springs are affected the speakers and vibrations in the amp cabinet for one. I imagine it's been left out of gainier amps partly that reason, not to mention the added complexity, extra tubes, and added noise.
 

andrewRneumann

Friend of Leo's
Joined
Mar 22, 2020
Posts
2,027
Location
Cincinnati, OH, USA
I don't see that as a problem. There has to be some trade off involved in mixing two stages together. In this amp I'm going to an have extra (cascade) preamp gain stage. Think Blenclowe's medium gain preamp ... bypassed input stage, a slightly cold biased stage, a warm biased stage, and a cathode follower. So, I'm going to have a pre PI volume, and possibly a post PI master.

I think your analysis is correct, but I’m not sure what the right answer is. Turn down the mid and bass in the tone stack is the easy way without changing anything. Raising the cap value and then decreasing the value on the previous stage, as you say, could maintain a similar response from the reverb circuit. If you do raise the value, does R8 become redundant? Looks like you’d have yet another coupling cap and grid leak resistor for the PI according to the schematic.

My experience is that every coupling cap is put on the substitution box and set by ear as the final. That way, even if everything else in the circuit is a copy, I can tweak one cap and say I “modded” the circuit. ;) Go for it.
 

Phrygian77

Poster Extraordinaire
Gold Supporter
Joined
Apr 30, 2016
Posts
5,748
Location
Crawfordville, FL
I think your analysis is correct, but I’m not sure what the right answer is. Turn down the mid and bass in the tone stack is the easy way without changing anything. Raising the cap value and then decreasing the value on the previous stage, as you say, could maintain a similar response from the reverb circuit. If you do raise the value, does R8 become redundant? Looks like you’d have yet another coupling cap and grid leak resistor for the PI according to the schematic.

My experience is that every coupling cap is put on the substitution box and set by ear as the final. That way, even if everything else in the circuit is a copy, I can tweak one cap and say I “modded” the circuit. ;) Go for it.

Yeah, you're absolutely correct about R8, it's just bringing down the impedance of everything. For my initial analysis, I left it as is because I wanted to see what may have been intended by whoever designed it, since removing it will obviously alter the frequency response.
 

andrewRneumann

Friend of Leo's
Joined
Mar 22, 2020
Posts
2,027
Location
Cincinnati, OH, USA
Yeah, you're absolutely correct about R8, it's just bringing down the impedance of everything. For my initial analysis, I left it as is because I wanted to see what may have been intended by whoever designed it, since removing it will obviously alter the frequency response.

Yeah, I can't really say what the total effect would be without modelling it. Too much interaction with the 2 channels right there. Seems like the design is high-passing the reverb signal on purpose. When you get the schematic squared away with both wet and dry channels and PI, let's have a look at it again!
 

Phrygian77

Poster Extraordinaire
Gold Supporter
Joined
Apr 30, 2016
Posts
5,748
Location
Crawfordville, FL
@andrewRneumann remove the 470k and you get a 3dB increase in the dry 'channel' bass frequencies. I didn't expect that, but it should have been obvious from the start.

I'm imagining a 5E3 with reverb, and the reverb is mixed in the same interactive way as the two channels. I'll have nightmares about that now.
 
Last edited:

andrewRneumann

Friend of Leo's
Joined
Mar 22, 2020
Posts
2,027
Location
Cincinnati, OH, USA
@andrewRneumann remove the 470k and you get a 3dB increase in the dry 'channel' bass frequencies. I didn't expect that, but it should have been obvious from the start.

It would also depend on what the PI input looks like. Though if it's the standard LTP with 1M grid leak with bootstrapped input impedance, maybe not that much difference.
 




New Posts

Top