Spring reverb 1, DSP 0

[Steve Conner]

Hi all,

A while back I wrote about my attempts to add a DSP reverb to a tube amp. It's all part of my plan to build a super ninja combo amp that looks like a 5E3 but is packed with cool features.

The DSP reverb sounded OK, but I was struggling to get the dynamic range needed for the amp to be playable at low volumes as well as cranked. (It's a non-MV design, so the reverb unit comes after the volume controls.) The DSP would either clip at high volume, or run out of bits and sound rough and unpleasant at low volume.

So I thought I would try a real tube-driven spring reverb. I got an Accutronics #9AB3C1B reverb tank, drove it with the two paralleled halves of a 12AT7 and the output transformer from a small table radio, and used half of a 12AX7 as the recovery amp. I hooked this lot up to the same points where the DSP had been connected.

I was amazed at the difference. The drive amp only clips a little when cranked, and the recovery amp not at all, but it still gives a lovely, rich sounding reverb when the volume is turned right down low. And the reverb pan costs the same as the DSP board :-)

Two things I'm not sure about: It could maybe use being a little brighter, and the 12AT7 runs heck of a hot. I biased it with a 680 ohm Rk to hopefully get about 5mA per side. I tried a 12AU7, and it gives even more reverb, but it drew so much current, the dropper resistor for that stage began to smoke. Should I be thinking about a stronger tube like an EL84 or 6V6? The table radio I got the OT from used an EL84.

[R.G.]

Hi all,
Two things I'm not sure about: It could maybe use being a little brighter, and the 12AT7 runs heck of a hot. I biased it with a 680 ohm Rk to hopefully get about 5mA per side. I tried a 12AU7, and it gives even more reverb, but it drew so much current, the dropper resistor for that stage began to smoke. Should I be thinking about a stronger tube like an EL84 or 6V6? The table radio I got the OT from used an EL84.

You might try using a MOSFET as a current booster to take some of the load off the 12AT7.

There's a cute current mirror based thing that came up for boosting germanium transistors that ought to work for a triode.

You could also diddle the biasing to get the 12AU7's stage current lower.

[Steve Conner]

Hi R.G.,

If I'm going to let the silicooties in, I guess I could just use a smallish TO-220 MOSFET in place of the triode. Or maybe a cascode of a JFET and a MOSFET, which would let me "cathode" bias it like a tube. That would save even more space since the chassis will be a bit cramped, and it would also help with heater current, which I'm running short of if I want to keep the option of using EL34s.

I was specifically wondering though, whether the small triode gives enough drive for that reverb tank (biasing it cooler would just give even less drive) and whether using a pentode instead would make the reverb brighter. A pentode has higher impedance, and Accutronics recommend high impedance drive. I guess the cascode I mentioned above would function like a pentode.

[Enzo]

Your DSP is limited by the power supplies it runs off of. If it has +/-15 or 12 for the analog parts (and I assume +5 and maybe +3.3 for the logic) then those voltages are the peak it can be fed. You slam tons more signal into a tube before it limits you, so I don't doubt the tube clips less easily.

If a 12AT7 can't drive a reverb pan, there are a whole lot of Fenders out there that don't realize they don't work...

A quick glance at a Fender circuit looks like about 3-4ma through the tube.

[Mark Hammer]

http://img.photobucket.com/albums/v4...er/Reverb1.png
http://img.photobucket.com/albums/v4...er/Reverb2.png

[Steve Conner]

Awesome article Mark, thanks! I feel all motivated now to get a couple more reverb tanks and hook them up to my synths. I'm always looking for more waterfall and less slab of plexiglass :-)

Enzo: Sure, the dynamic range of the DSP is limited at the high end by its supply rail. But that wouldn't matter (the signal could just be attenuated before the DSP, then amplified again afterwards, with a tube of course...) if it weren't for the fact that it's limited at the low end too.

And the low end of a digital reverb's dynamic range is limited in the most unpleasant sounding way you can imagine, by things like limit cycles and arithmetic rounding errors. The tails of reverb decays go missing, and you start to hear strange drones, blips, and hash. I always wonder if Line6's stomp boxes do the same when used in front of a dirty amp with a lot of gain.

On the other hand, a spring reverb's dynamic range is limited at the bottom by hum pickup in the receiver coil. Hum can be annoying, but it doesn't seem to spoil the character of the reverb effect in the same way that a DSP running out of resolution does.

Here's a link to some pics of my progress on this project:
http://scopeboy.com/gallery/index.php?showimage=297

I think it's almost finished now, except I might add one reverb level control per channel, and a 20dB attenuator switch post-PI. Same idea as a post-PI MV, but just a switch with two settings, 0 and -20.

[R.G.]

I'm reminded of a quote in Russell O. Hamm's article on Tubes vs Transistors where he noted that a tube overloading was much like an ideal compressor.

I suspect your results would get better if you companded around the digital reverb, and perhaps soft limited at some signal level.

[Mark Hammer]

One of the tricks that Craig Anderton uses in the PAiA/Anderton "Hot Springs" reverb project is two pans wired out of phase so as to cancel the hum. Essentially a humbucker pickup, but with reverb pans. Throw some creative use of downward expansion on top of that, and you should have quite a clean reverb path.

One of the interesting things about the Hot Springs circuit/configuration is that it tolerates use of divergent pans, as long as the sensing transducers are impedance matched so as to yield hum-rejection. In theory, a person could have two different driving circuits with independent dwell controls, one driving a long 3-spring pan and the other driving a shorter 2-spring pan. Throw in a simple toggle-based capacity to limit low-end to each pan, and there is no end to the reverb textures attainable.

One of the great things about spring reverb, as opposed to digital, is how easy it is to generate a unique and idiosyncratic sound. Of course, you can do that in digital too...the creation part that is. With springs, though, radical change can be as simple as a quarter turn of the "Dwell" pot.

Sometimes, low-tech is more sophisticated than high-tech.