"Like" AX84 on facebook

'A transformerless reverb driver'
Author:John Polstra (registered user: 654 posts )
Date: Fri, Jul 20th, 2012 @ 15:07 ( . )

Recently I've been experimenting with a tube-based reverb driver circuit that doesn't use a reverb transformer. It's designed to deliver a constant signal current (not voltage) to the reverb tank regardless of the frequency. That's what Accutronics recommends in their technical documentation.

The circuit is working out very well, so I wrote a short article about it. I thought some of you might be interested. Here's the link:

[link]

There's a sound clip at the end of the article.

John

-- REPLY: [With No Quote] --- [With Quoted Text]

'A transformerless reverb driver'
Author:Stephen Keller (registered user: 5737 posts )
Date: Fri, Jul 20th, 2012 @ 16:14 ( . )

On 07/20/2012 @ 15:07, John Polstra wrote :
Recently I've been experimenting with a tube-based reverb driver circuit that doesn't use a reverb transformer...

Thanks for the thorough and informative analysis John. That sound sample is golden. I'm thinking that some of its sparkle is due to the improved high-frequency response of the constant-current driver compared to the traditional transformer coupled designs.

Can we use this in our builds?

Stph

-- REPLY: [With No Quote] --- [With Quoted Text]

'A transformerless reverb driver'
Author:John Polstra (registered user: 654 posts )
Date: Fri, Jul 20th, 2012 @ 16:47 ( . )

On 07/20/2012 @ 16:14, Stephen Keller wrote :
: Thanks for the thorough and informative analysis John. That sound sample is golden. I'm thinking that some of its sparkle is due to the improved high-frequency response of the constant-current driver compared to the traditional transformer coupled designs.


Thanks, Stephen. I agree about the high-frequency response. I've noticed that typical Fender reverb circuits roll off the input to the driver at around 300 Hz, and I think that must be partly to compensate for the loss of highs in the driver itself.


: Can we use this in our builds?


Absolutely!

John

-- REPLY: [With No Quote] --- [With Quoted Text]

'A transformerless reverb driver'
Author:Stephen Keller (registered user: 5737 posts )
Date: Fri, Jul 20th, 2012 @ 18:28 ( . )

Thanks.

I think the configuration of that driver is called an "RC parafeed" stage. Bench has an overview here:

[link]

See the section entitled: "Resistor Load, Capacitively Coupled to the Output Transformer."

I think one could accomplish the same effect with a triode having a constant-current sink (CCS) in the load (either the cathode or anode side), so if a pentode isn't handy, a triode and a little sand should suffice, provided the design adhered to your output-impedance guidelines.

Stph

-- REPLY: [With No Quote] --- [With Quoted Text]

'A transformerless reverb driver'
Author:John Polstra (registered user: 654 posts )
Date: Fri, Jul 20th, 2012 @ 19:08 ( . )

I was thinking about how to do it with a triode, and I came up with a couple ways that should work. The easiest would be to make a normal triode amplification stage with a medium-mu tube like a 12AU7A, and then put a resistor R in series with the reverb tank. Choose the resistor so that the output impedance of the triode stage + R was 15K or so. The only thing is, the stage would need to be able to produce a pretty large voltage swing at its plate. It should be doable, though.

Another way to do it would be to use a cascode arrangement. I would just use a MOSFET for the top half of the cascode. It would make the effective plate resistance of the triode practically infinite, and you could design the circuit almost as if you had a real pentode.

I might beef up the article to talk about those possibilities, once I've tried them out.

John

-- REPLY: [With No Quote] --- [With Quoted Text]

'A transformerless reverb driver'
Author:PhilS (registered user: 1576 posts )
Date: Sat, Jul 21st, 2012 @ 08:24 ( . )

The Gibson GA-15RVT uses a 12AU7 cathode follower to drive the reverb tank. There is a divider on the cathode 3.3K and then 10K to ground, with a cap between the divider point and the tank input. Recovery is done with a 6EU7. Have a look!

John, that's a great article. Thanks for posting!

-- REPLY: [With No Quote] --- [With Quoted Text]

'A transformerless reverb driver'
Author:John Polstra (registered user: 654 posts )
Date: Sat, Jul 21st, 2012 @ 15:29 ( . )

That's interesting, Phil. I wish I knew for sure which reverb tank Gibson used in that amp. Googling gave me mixed answers. Some folks said it was a 9BC2C1B (190 ohm), but one guy posted a big list that said 4FB2A1C (1475 ohms).

I just finished refurbishing an ancient Danelectro DM-25 for a guy. It drives the reverb tank capacitively coupled from a pair of medium-mu triodes (6FQ7 or 6CG7) in a self-split push-pull arrangement. It sounds pretty good, too. The reverb tank is like nothing I've ever seen before. It's skinny and sits on a couple of leaf springs on top of the chassis. I read somewhere that its output transducer is piezoelectric. Luckily, the original tank still worked. I don't think I could have found a replacement anywhere.

John

-- REPLY: [With No Quote] --- [With Quoted Text]

'A transformerless reverb driver'
Author:PhilS (registered user: 1576 posts )
Date: Sat, Jul 21st, 2012 @ 15:53 ( . )

John,
We all wish that we could know for sure with the old Gibsons ;-}. I wrote to Accutronics in October 2011 to ask about the correct tank for the GA20RVT, which some people refer to as the "big brother" of the GA15RVT, though I think this is really a cosmetic comment, not a circuit comment. What prompted me to do this was my intention to clone the GA20RVT with a few mods, as I acquired a set of iron from a GA15RVT, which is supposed to be the same iron for both amps. I bought the lot on eBay from a guy in Whitehorse, YT! I wanted the interstage transformer, which went into my GA20RVT. I figured with the leftovers, I'd build another with a tube PI, but never got around to it, or maybe I should say I didn't get to it yet. Anyway, the curious thing is that the lot came with a reverb transformer (which I'm not willing to part with)!

It is hard to know how good the records are, but they replied quickly that the original tank is 4FB2A1B, making it 1475z. Given what I know about Gibson amps (I own two of them and done a bit of reading), I am suspecting the 4FB2 tank was widely used. I'd be willing to go out on a limb and say that's the tank in the GA15RVT, too.

Best,
Phil

-- REPLY: [With No Quote] --- [With Quoted Text]

'A transformerless reverb driver'
Author:brawler (registered user: 290 posts )
Date: Tue, Aug 07th, 2012 @ 12:02 ( . )

I have a '65 GA-15RVT. It's at home so I can't look now. I will try to remember to look tonight when I get home tonite. I will let ya know what tank is there, it is original to the amp.

Jim

-- REPLY: [With No Quote] --- [With Quoted Text]

'A transformerless reverb driver'
Author:PhilS (registered user: 1576 posts )
Date: Tue, Aug 07th, 2012 @ 13:21 ( . )

Hi brawler. This is something people want to know. I'm thinking there is a Gibbs part number that doesn't follow the current Accutronics numbering system and if Accutronics has a parts substituion list. Given that Gibbs was acquired by Accutronics, there is probably a better than even chance such a list exists. Still, it would be good to find out what we can. Thanks.

-- REPLY: [With No Quote] --- [With Quoted Text]

'A transformerless reverb driver'
Author:stringbend (registered user: 1112 posts )
Date: Mon, Jul 23rd, 2012 @ 10:08 ( . )

On 07/21/2012 @ 15:29, John Polstra wrote :
: I just finished refurbishing an ancient Danelectro DM-25 for a guy. It drives the reverb tank capacitively coupled from a pair of medium-mu triodes (6FQ7 or 6CG7) in a self-split push-pull arrangement. It sounds pretty good, too.


Among all the other interesting things in this thread, this one really caught my eye. I had a DM-25 and I LOVED it. Except for the reverb. Mine was nothing but over-the-top cavern echo. Even a reverb junkie like me couldn't find a use for it.


-- REPLY: [With No Quote] --- [With Quoted Text]

'Danelectro DM-25 pics'
Author:John Polstra (registered user: 654 posts )
Date: Mon, Jul 23rd, 2012 @ 10:36 ( . )

Hijacking my own thread . . . Here's a few pictures of the DM-25 before I worked on it. The first scary thing I saw was that somebody had jumpered around the fuse!

[link]

Does anybody know what those metal sleeves are around two of the tubes? I don't even know if they are original or not. They'd be worthless as shields, since they're not grounded. Could they be heat radiators intended to be put on the power tubes? The amp runs the power tubes very hot.

John

-- REPLY: [With No Quote] --- [With Quoted Text]

'Danelectro DM-25 pics'
Author:Stephen Keller (registered user: 5737 posts )
Date: Mon, Jul 23rd, 2012 @ 11:43 ( . )

On 07/23/2012 @ 10:36, John Polstra wrote : Does anybody know what those metal sleeves are around two of the tubes? I don't even know if they are original or not. They'd be worthless as shields, since they're not grounded. Could they be heat radiators intended to be put on the power tubes? The amp runs the power tubes very hot.

Noting the socket on the second tube from the left, I expect the sleeve is supposed to be an electrostatic shield of some sort and that it should make some sort of physical connection with the socket. As you point out it obviously isn't serving in that function now. Can't say for the other, but it is possible that one or more sockets were replaced since this was new. If so, it would have to have been done a while back as the socket rivets have a distinctly "factory" look about them. It's also possible the sleeves aren't original and that there was only one shield to begin with.

Given their shape, I doubt those sleeves are intended to radiate heat from the power tubes.

Stph

-- REPLY: [With No Quote] --- [With Quoted Text]

'Danelectro DM-25 pics'
Author:Celeste hall (registered user: 2252 posts )
Date: Mon, Jul 23rd, 2012 @ 17:09 ( . )

I would say someone put taller tubes in then the tube shields were designed for. As shown they are doing nothing useful

-- REPLY: [With No Quote] --- [With Quoted Text]

'A transformerless reverb driver'
Author:Paul Fawcett (registered user: 3028 posts )
Date: Sat, Jul 21st, 2012 @ 08:29 ( . )

On 07/20/2012 @ 19:08, John Polstra wrote :
I was thinking about how to do it with a triode, and I came up with a couple ways that should work. The easiest would be to make a normal triode amplification stage with a medium-mu tube like a 12AU7A, and then put a resistor R in series with the reverb tank. Choose the resistor so that the output impedance of the triode stage + R was 15K or so. The only thing is, the stage would need to be able to produce a pretty large voltage swing at its plate. It should be doable, though.
:
: Another way to do it would be to use a cascode arrangement. I would just use a MOSFET for the top half of the cascode. It would make the effective plate resistance of the triode practically infinite, and you could design the circuit almost as if you had a real pentode.
:
: I might beef up the article to talk about those possibilities, once I've tried them out.
:
: John
:
--



The sound clip sounds really nice.

As for using a triode with a series resistance, Merlin's been there and done that: [link] In fact, I used a circuit very much like his parallel 12AU7 version (in conjunction with a 1.5K Zin tank) in my Grail amp. It works well -- the quality of the reverb is fairly similar to the one in your clip.




-- REPLY: [With No Quote] --- [With Quoted Text]

'A transformerless reverb driver'
Author:John Polstra (registered user: 654 posts )
Date: Sat, Jul 21st, 2012 @ 15:35 ( . )

For some reason I can't connect to Merlin's site at the moment, but I'll check it out later. It's good to know that the triode + resistor approach works. I may try changing the amp I recorded to use the triode from the 6GH8A to drive the reverb. Then I could use the pentode for the recovery stage. I'd like to be able to get just a little bit more gain out of the recovery side.

John

-- REPLY: [With No Quote] --- [With Quoted Text]

'A transformerless reverb driver'
Author:printer2 (registered user: 115 posts )
Date: Sat, Jul 21st, 2012 @ 22:06 ( . )

Found this schematic with a 9 pin 6U8A pentode-triode driving a capacitor coupled reverb. Might be of interest.

[link]

-- REPLY: [With No Quote] --- [With Quoted Text]

'A transformerless reverb driver'
Author:John Polstra (registered user: 654 posts )
Date: Sat, Jul 21st, 2012 @ 22:33 ( . )

Thanks for that link. I actually started with a 6U8A, but I couldn't get quite 4.5 mA of drive current out of it for my 600 ohm tank. I bet it works just fine with that higher impedance tank, though.

I really like those triode-pentode tubes. They give you so many options in an amp.

John

-- REPLY: [With No Quote] --- [With Quoted Text]

'A transformerless reverb driver'
Author:RickTown (registered user: 792 posts )
Date: Mon, Aug 06th, 2012 @ 13:01 ( . )

Looking at the schematic that printer2 posted, [link]
I'd like to add a version of this to my stock HO. At this point I would think to add it to my first stage and reinsert it after the third stage. I may have to tweak a few values on couplers and grid resistance. The tank I want to use is 800R input so I was going to use a 6GH8 like John did to make sure I have enough current.

What modifications would any of you recommend?

-- REPLY: [With No Quote] --- [With Quoted Text]

'A transformerless reverb driver'
Author:Celeste hall (registered user: 2252 posts )
Date: Mon, Aug 06th, 2012 @ 17:28 ( . )

You may have some level matching troubles mixing the wet signal back in that late, unless you some serious recovery. Were it me, I would try it, but would be prepared to do my initial recover with a LND150 cascaded into the existing tube recovery

-- REPLY: [With No Quote] --- [With Quoted Text]

'A transformerless reverb driver'
Author:RickTown (registered user: 792 posts )
Date: Mon, Aug 06th, 2012 @ 18:34 ( . )

Should I be concerned about dirtying up the recovered reverb by inserting it sooner. I plan to use it mainly in lower gain tones but you never know what I might like.

I've never experimented with adding a reverb before so I don't know how much signal I'm going to have.

Also I'm curious about the size of the cathode bypass caps on both ends of the tank. I'm not used to seeing caps that big for bypass. What's the reason for the size?

-- REPLY: [With No Quote] --- [With Quoted Text]

'A transformerless reverb driver'
Author:Celeste hall (registered user: 2252 posts )
Date: Mon, Aug 06th, 2012 @ 19:13 ( . )

The bypass on the pentode driver is much bigger then it needs to be, likely corner frequency below 1ha (just a swag there) that is a lot lower then the coupling cap is going to pass into a 800 ohm tank. The 22uf on the recover stage is a bit larger then it needs to be, but is likely designed by copying an existing gain stage, without thinking what this stage really need,

-- REPLY: [With No Quote] --- [With Quoted Text]

'A transformerless reverb driver'
Author:RickTown (registered user: 792 posts )
Date: Mon, Aug 06th, 2012 @ 19:42 ( . )

Ok so I'm so far off when assuming that I can still adjust those to more common sizes.

Thanks. Any other advice on making the most of this?

-- REPLY: [With No Quote] --- [With Quoted Text]

'A transformerless reverb driver'
Author:John Polstra (registered user: 654 posts )
Date: Mon, Aug 13th, 2012 @ 20:09 ( . )

On 08/06/2012 @ 19:13, Celeste hall wrote :
The bypass on the pentode driver is much bigger then it needs to be, likely corner frequency below 1ha (just a swag there) that is a lot lower then the coupling cap is going to pass into a 800 ohm tank. The 22uf on the recover stage is a bit larger then it needs to be, but is likely designed by copying an existing gain stage, without thinking what this stage really need,
--



Yes, I just chose bypass capacitors that I knew would be "plenty big enough".

John

-- REPLY: [With No Quote] --- [With Quoted Text]

'A transformerless reverb driver'
Author:tubeswell (registered user: 107 posts )
Date: Sun, Aug 19th, 2012 @ 14:02 ( . )

On 07/20/2012 @ 19:08, John Polstra wrote :
I was thinking about how to do it with a triode, and I came up with a couple ways that should work. The easiest would be to make a normal triode amplification stage with a medium-mu tube like a 12AU7A, and then put a resistor R in series with the reverb tank. Choose the resistor so that the output impedance of the triode stage + R was 15K or so. The only thing is, the stage would need to be able to produce a pretty large voltage swing at its plate. It should be doable, though.
:
: Another way to do it would be to use a cascode arrangement. I would just use a MOSFET for the top half of the cascode. It would make the effective plate resistance of the triode practically infinite, and you could design the circuit almost as if you had a real pentode.
:
: I might beef up the article to talk about those possibilities, once I've tried them out.
:
: John
:
--



A 12AU7 cascode might work very well

-- REPLY: [With No Quote] --- [With Quoted Text]

'A transformerless reverb driver'
Author:tubeswell (registered user: 107 posts )
Date: Sun, Aug 19th, 2012 @ 14:07 ( . )

Actually strike that - you want high output impedance.

-- REPLY: [With No Quote] --- [With Quoted Text]

'A transformerless reverb driver'
Author:John Polstra (registered user: 654 posts )
Date: Sun, Aug 19th, 2012 @ 16:53 ( . )

I haven't modeled or tried a 12AU7 cascode, but I think it might work just fine. The plate resistance of the upper tube effectively gets multiplied by something like (mu + 2), so I think the output impedance would be primarily determined by the plate resistor of the top tube. I guess you'd need a pretty high supply voltage so the plate resistor could be reasonably large. Probably not worthwhile, when you could simply use a pentode instead.

John

-- REPLY: [With No Quote] --- [With Quoted Text]

'A transformerless reverb driver'
Author:tubeswell (registered user: 107 posts )
Date: Sun, Aug 19th, 2012 @ 18:24 ( . )

From Merlin's 1st book, the output impedance of a cascode is Ra||ra2+ra1(u2+1)

where:

ra2 = upper triode's plate resistance
ra1 = lower triode's plate resistance
u2 = the amplification factor of the upper triode

For most triodes ra+ra(u+1)>Ra, therefore Zout~=Ra

which is why I was subsequently thinking that the output impedance wouldn't be high enough

-- REPLY: [With No Quote] --- [With Quoted Text]

'A transformerless reverb driver'
Author:Chris A (registered user: 751 posts )
Date: Sat, Jul 21st, 2012 @ 23:16 ( . )

I like the sound. It sounds more like a real "space" with lots of room between hard walls. I think you've got a great design. The Fender reverb sound is very different from yours. The key to the Fender sound is the tiny transformer that acts like a filter limiting the lows and that Fender used such a powerful driver tube as the 6K6 (now 6V6 in the current production units) Yours is a more "HiFi" reverb sound.

It's good to see people building new ideas. Thanks.

-- REPLY: [With No Quote] --- [With Quoted Text]

'A transformerless reverb driver'
Author:RickTown (registered user: 792 posts )
Date: Mon, Aug 13th, 2012 @ 19:54 ( . )

John, I am trying to spice this reverb for my HO but would be a lot further along if you might lend me your 6GH8 and reverb tank models. I hope I'm not asking too much.
If it's not a problem, my email is rtownsendATcuttingedgediamondDOTcom
Thanks in advance

-- REPLY: [With No Quote] --- [With Quoted Text]

'A transformerless reverb driver'
Author:John Polstra (registered user: 654 posts )
Date: Mon, Aug 13th, 2012 @ 20:07 ( . )

Rick, I actually modeled the tube as a 6U8 rather than 6GH8. They are pretty similar, although in practice I found that the 6GH8 drives the tank a little bit better. You can find the 6U8 model in the file "Koren_Tubes.inc", which I think is from the LTSpice users group. I found it by googling the filename. Let me know if you can't find it, and I'll send it to you.

The 600 ohm tank was modeled by a 58 ohm resistor in series with a 95.046 mH inductor. That gives the correct DC resistance and also the correct impedance (600 ohms) at 1KHz.

John

-- REPLY: [With No Quote] --- [With Quoted Text]

'A transformerless reverb driver'
Author:John Polstra (registered user: 654 posts )
Date: Mon, Aug 13th, 2012 @ 20:18 ( . )

About modeling the reverb tank: If you have a tank with a DC resistance of R and a 1KHz impedance Z (with values taken directly from the Accutronics website), you can model it as a resistor R in series with an inductor L. The value of the inductor is:

L = SQRT(Z^2 - R^2) / (2*pi*1000)

in Henrys.

For the 800 ohm tank you mentioned, R is 58 ohms and L is 127 mH.

John

-- REPLY: [With No Quote] --- [With Quoted Text]

'A transformerless reverb driver'
Author:Paul Fawcett (registered user: 3028 posts )
Date: Mon, Aug 13th, 2012 @ 21:35 ( . )

Cool, thanks John. Also good measured data on several types of tank here: [link]

-- REPLY: [With No Quote] --- [With Quoted Text]

'A transformerless reverb driver'
Author:RickTown (registered user: 792 posts )
Date: Mon, Aug 13th, 2012 @ 21:37 ( . )

Thanks John.
Just making sure I can get the recovery i want before I start rippin-n-tearin into my amp. I put the series resistance setting in the inductor properties. I'm assuming this is the same. I guess I'll try it both ways and see if the output changes. Thanks again.
Rick

-- REPLY: [With No Quote] --- [With Quoted Text]

'A transformerless reverb driver'
Author:RickTown (registered user: 792 posts )
Date: Tue, Aug 14th, 2012 @ 13:09 ( . )

Well I have ltspice working on this and I see something in spice that I don't know is accurate. As I draw the reverb circuit and sim it, all seems well. When I attach the return signal to the dry path, it seems to clip in a strange way in the reverb output. I'm skipping 1 preamp stage with 2 reverb stages so the wet signal is out of phase with the dry signal. Is this the cause?

-- REPLY: [With No Quote] --- [With Quoted Text]

'A transformerless reverb driver'
Author:foulowl (registered user: 341 posts )
Date: Fri, Aug 17th, 2012 @ 01:09 ( . )

What kind of cap are you using for that 1uF coupling cap? Wouldn't it be a bad idea to use a polarized cap there?

Sounds fantastic though!

-- REPLY: [With No Quote] --- [With Quoted Text]

'A transformerless reverb driver'
Author:John Polstra (registered user: 654 posts )
Date: Fri, Aug 17th, 2012 @ 16:20 ( . )

I used a polarized electrolytic, although you could use a film capacitor if you wanted to. The plus side is at plate potential (around 186 volts) and the minus side is at ground potential, so there's no reason to use a non-polarized capacitor.

I also doubt there'd be any advantage to using a film capacitor instead of a small, cheap electrolytic. Remember, reverb tanks are kind of the opposite of "hi-fi".

John

-- REPLY: [With No Quote] --- [With Quoted Text]

'A transformerless reverb driver'
Author:foulowl (registered user: 341 posts )
Date: Fri, Aug 17th, 2012 @ 17:13 ( . )

I don't have too much experience with reverb tanks, I have just read that electrolytic caps can produce ugly tone artifacts. But your clip sounds fantastic currently. I just haven't ever seen an electrolytic as a coupling cap before. :) Cheers!

-- REPLY: [With No Quote] --- [With Quoted Text]

'A transformerless reverb driver'
Author:printer2 (registered user: 115 posts )
Date: Fri, Aug 17th, 2012 @ 19:07 ( . )

On 08/17/2012 @ 17:13, foulowl wrote :
I don't have too much experience with reverb tanks, I have just read that electrolytic caps can produce ugly tone artifacts. But your clip sounds fantastic currently. I just haven't ever seen an electrolytic as a coupling cap before. :) Cheers!
--



Look at some single supply SS amps.

-- REPLY: [With No Quote] --- [With Quoted Text]

'A transformerless reverb driver'
Author:John Polstra (registered user: 654 posts )
Date: Fri, Aug 17th, 2012 @ 19:10 ( . )

On 08/17/2012 @ 17:13, foulowl wrote :
I don't have too much experience with reverb tanks, I have just read that electrolytic caps can produce ugly tone artifacts. But your clip sounds fantastic currently. I just haven't ever seen an electrolytic as a coupling cap before. :) Cheers!
--



Electrolytic caps are definitely not the best tone-wise, but on the other hand, they're not all that bad. If you ever look at any solid state schematics, you'll find that electrolytics are used all over the place in the signal path. They pretty much have to be used, because the impedances are so low in solid state gear. (I'm talking about solid state gear using discrete transistors. Op amp circuits use fewer coupling caps of any kind, because they're mostly direct-coupled.)

Also, by the time the sound has bounced around for awhile in the springs of a reverb unit, any tone artifacts from an electrolytic coupling cap will be undetectable.

Anyway, I'm glad you liked the sound of the thing!

John

-- REPLY: [With No Quote] --- [With Quoted Text]


--- 26 Users Online --- 24 Recent Unique Posters

Q497=1406917109 - Threads: / 1406917109