12DW7 Reverb

[Natman]

Yeah that definitely crossed my mind but the tank is an oddball and they're a little hard to find. The jacks measure DC resistance in the right range, but that doesn't mean it's not damaged. The tank's high impedance and parallel circuit could mean that a simple component mismatch would send the current away from the input. I haven't had time to get into more troubleshooting this week even though I was home 3 days with the flu. I do intend to order another tank eventually, but I had it working a little some time ago. The problem then was a gated effect but lately I get nothing at all. Did I fry something?

Octal: R8 and R9 are definitely wired this way in the original circuit! What I haven't mentioned is that this design is based on a boutique amp which in turn was almost entirely based on the Gibson Scout. Some of what I came up with is from gut shots (pics) of one of those.

[dchang0]

Well, here's where the cheap pair of computer speakers will come in handy. You can test the reverb driver side without the tank and prove that it is working. If you do, then test the recovery circuit with a guitar, and then you have proven that the tank is at fault--either because it is broken or because it is impedance mismatched.

Get better soon, man. Of course, staying at home with the flu would be an opportunity for some to spend more time debugging their amps, LOL.

[octal]

>
Octal: R8 and R9 are definitely wired this way in the original circuit! What I haven't mentioned is that this design is based on a boutique amp which in turn was almost entirely based on the Gibson Scout. Some of what I came up with is from gut shots (pics) of one of those.
>
Do you happen to know which model Gibson Scout? The only Scout I could find a schematic for used a transformer driven reverb and was nothing alike. Gibson schematics are notoriously error ridden too. I guess what I'm saying is that the boutique amp builder may have copied a mistake... presumably it would at least work with the error but it may not be optimized. For signal to noise reasons, it doesn't make sense to attenuate the output of the tank before amplifying it, and I very much doubt the reverb tank can generate enough output to overload the following gain stage.

Checking the reverb recovery circuit with a guitar as a signal generator is a good idea. I'm not so sure about checking the drive side with a computer speaker... there should be a pretty large voltage output there- too much for an active computer speaker.

[dchang0]

Oh, yeah. Use a passive computer speaker. You know, those cheap ol' $5 ones that most everyone tosses out. Something throwaway, that you aren't afraid to fry if you stuff too much voltage/current through it.

Or possibly an old abandoned guitar amp speaker that you don't care if you blow up, even though its impedance is way lower than that tank.

As for R8 and R9, I've seen something similar in the FBT 500 R2 schematic. They used a 27K resistor to dump signal to ground, then a 2.2kpF cap to couple to a 10M(!) grid leak resistor for a 1/2 ECC83 recovery stage. Strange that they did it that way--why not just use a lower grid leak? I figure they must've been trying to use the impedance of the tank itself with the 27K resistor as a voltage divider of sorts.

But I borrowed and then frankensteined the idea anyway (by using only a 22K grid leak) and it seems to work fine.

500-R2.pdf

[octal]

Oh, yeah. Use a passive computer speaker. You know, those cheap ol' $5 ones that most everyone tosses out. Something throwaway, that you aren't afraid to fry if you stuff too much voltage/current through it.

Or possibly an old abandoned guitar amp speaker that you don't care if you blow up, even though its impedance is way lower than that tank.

As for R8 and R9, I've seen something similar in the FBT 500 R2 schematic. They used a 27K resistor to dump signal to ground, then a 2.2kpF cap to couple to a 10M(!) grid leak resistor for a 1/2 ECC83 recovery stage. Strange that they did it that way--why not just use a lower grid leak? I figure they must've been trying to use the impedance of the tank itself with the 27K resistor as a voltage divider of sorts.

But I borrowed and then frankensteined the idea anyway (by using only a 22K grid leak) and it seems to work fine.

[ATTACH]11960[/ATTACH]

That FBT 500 schematic is a totally different scenario than the Scout one though- the 27K is to load the tank (probably so the coil doesn't ring too much) and then the 10M is a grid leak bias resistor. It's not a voltage divider- the 27K and the 10M are are parallel loads that are AC coupled together. You can essentially ignore the 10M... the tank sees a 27K load.

In the OP's schematic, the 10K in series with the tank feeds 22K... so it's throwing away 1/3 MINIMUM signal from the tank- actually more due to the source impedance of the tank.

Nathan

[dchang0]

Ah, that makes sense. Thanks for the explanation.

Yeah, the 10K seems like a mistake.

[Natman]

Ok quick update: I dug up an old Blues Jr. tank (8DB2...) and everything worked FINE! I was quite surprised. That means that either the 8FB tank was damaged in shipping or its input impedance is too high and blocks the signal. For now, I'm prepared to just use the amp as is. I would have preferred long decay but it's better to have something than nothing at all! I have to say that I'm quite relieved. There are surely issues which should be addressed but right now I have a sweet-sounding functional amp and I'm quite glad. I want to thank you guys for the help! I'll keep trying to optimize each stage. I'll take new readings all around too.

I'm not sure why the original amp used this particular model of tank, maybe because the old Ampegs & Traynors did?
FWIW I have not found ANY amps that inporporate the reverb quite like this one does.

So if I want more reverb signal I gotta drop R9?

[dchang0]

Awesome--glad you fixed it!

The 8F tank should still have worked. It's probably broken.

Yes, drop R9 to improve your reverb signal. You might have too much signal, but you can probably dial that down other, better ways.

[Natman]

...And yet another update: I bought a 9FB2*** MOD tank off ebay because it was the cheapest I could find and boy the reverb just came alive! It's now as full & lush as any BF Fender basically! That input impedance makes all the difference. I don't know if a real USA Accutronics would be any better, but they're selling their old stock for over $50 +shipped. Pretty steep!

[tubeswell]

Hi Natman.

Out of curiosity, did that first tank's input transducer test okay for continuity with your R-meter?

[allyshake]

The circuit is Gibson Scout all the way except with modern tubes essentially. The tank is high impedance 8FB3C1C and it is cap-coupled.

Doesn't the Scout use an OT to drive the reverb? Did you ever get this working well with cap coupled single triode stage driving the tank? How did it sound. I am working on a similar problem here: https://music-electronics-forum.com/...d=1#post529115

Is there a link to more discussion by Swart or the Atonic project? Cheers