News:

SMF for DIYStompboxes.com!

Main Menu

Analogue Revberrb

Started by lightningfingers, April 01, 2004, 12:14:26 PM

Previous topic - Next topic

lightningfingers

how does one buy/make a reverb tank?

i only ave a time based reverb unit and it SO totally sucks
U N D E F I N E D

Mark Hammer

"Reverb" needs to mimic a few things about natural spaces:

1) Unless you are in the middle of the desert in front of one very reflective wall, there will be multiple reflection times, so iterations of the original input signal can not be regularly spaced.  This is why analog delay lines, or any delay with a single fixed echo rate, has a hard time simulating reverberant space.

2) The loss of acoustic energy in reverberant spaces is not constant across the spectrum - one loses highs faster than one loses lows because reflecting surfaces tend to eat up high frequency content and it wasn't that high amplitude to begin with.

One of these properties, you can sort of fix in analog delay lines by inserting some simple lowpass filtering into the regeneration loop so that each iteration becomes progressively muter.  The other property of multiple unrelated reflection times is much more difficult to mimic using a single BBD/delay-time.  The old (now discontinued) MN3011 and MN3214 chips from Matsushita tried to fake this by having multiple unrelated taps along the delay, whose outputs you could mix together.  I've never knowingly heard one of these so I couldn't comment on how good a job they might be able to do.  There are posted projects with the MN3011, and there may be the odd one floating around out there, but you won't easily obtain one (keep your eyes peeled for busted karaoke machines, though, since I figure they were developed for the echo add-in market).

The simplest non-digital way to produce an emulation of reverberant spaces is by means of springs.  The amp treats the spring input as if it was a speaker, causing the springs to shake.  Much like splashing your hand in the bathtub, though, the pressure wave initiated at the start of the spring does not simply proceed from one end to the other, but bounces around the spring until it loses energy.  This produces multiple unrelated iterations of the input signal, and also loses high-end energy PDQ.  The "wiggles" at the end of the spring are detected by a sensor much the way a contact mic would.

Reverb pans provide the springs, plus send and receive transducers.  Generally speaking, the longer the springs, and the more of them, the more pleasing a reverb sound you get since more springs implies greater diversity in reflections and delay time.  Some old car radios used to come with single short-spring reverb units that almost looked like heater coils.   These would "work" but the sound was too "boingy" and not diffuse enough to mimic a large space.  The ideal would be something like a short 3-spring pan and a long one.  You could blend them or select one or the other for different types of reverberant sounds.

I keep threatening to finish a project I've been working on for a while now, using springs from Home Depot.  I've gotten parts of it working, and it is quite feasible to make a working spring reverb in a fairly small package.  If Danelectro and Little Lanelei can do it, so can we.

I have been using a small speaker from a computer (the beeper speaker) to drive the springs, and a crystal cartridge mic from an old phone handset as the sensor.  The spring is epoxied to the middle of the dome on the speaker at the driver end, and soldered to the piezo disc on the receiving end.  If you get 3 springs (they are a bit over a buck apiece) and configure them in a Y shape so that one speaker branches out to transmit to two piezo sensors, you end up with something that ought to mimic the diversity of multispring arrays.  The 3 springs are soldered together in the middle.  In fact most commercial reverb pans use a pair of actual *physical* springs to form each *functional* spring, with a solder joint in the middle.  The deal with springs is they need to be compliant enough to create diversity in reflection/propogation times, but no so loose that they sag.  The solder joint in the middle helps to provide a little stiffness along the way and prevent sagging.

The speaker, being a low impedance load, needs to be driven by something that is comfortable around such things.  That can be any mini-power-amp chip like a 386 or anything else intended for headphone use, or it can be an LM833 or NE5534/5532 that is able to sink lots of current into a very low-impedance load.  At the sensing end, the piezo mics are very high impedance, so you need either a FET or bifet op-amp stage immediately after them that can handle a high-impedance source.  Using a pair of transducers puts the sensors in parallel which ought to help, and also yield a pretty beefy return signal.  That's good because it means you don't have to crank up the gain that high, so you don't acquire other problems like hiss or other extraneous noise.  If you felt like it, you could also send each transducer to its own output for stereo.  Since there is no reason to expect the signal on each spring to be identical, there'd be a little stereo effect there.

I'm going to build mine into an aluminum chassis that is about 8"L x 4"W x about 2" high (yet one more reason to use computer speakers).

In principle, this ought to yield an acceptable reverb tone for minimal expenditure.  It *won't* likely be a nice tube-based Fender reverb sound, but what the heck, it's DIY and I'll be able to tailor the tone and processing options to my tastes.

Alterantively, you can just go to the Accutronics website, download their schems, and buy a pan or two from them.

Peter Snowberg

For a DIY tank, try these links: :D

http://www.electronicpeasant.com/projects/springs/springs.html
http://www.electronicpeasant.com/projects/thermio/thermio.html

I would just buy a tank if you want classic spring reverb sound and then add a simple driver amp & recovery amp/mixer. Search the forums for "spring and reverb" for a ton of posts about this.

To buy a tank, go here: http://tubesandmore.com/ and then click on "guitar and amp items", and then on reverb tanks. I would steer you towards a 3 spring tank like the 9AB3C1B. I like the tanks with 8 or 10 ohm drive coils for flexability, but be aware that the Stage Center Reverb from Anderton (see GGG) uses a tank with a higher impedance drive coil.

Also see the Accutronics site for more details: http://www.accutronicsreverb.com/

Take care,
-Peter
Eschew paradigm obfuscation

lightningfingers

:shock: fast reply!

thanks guys this really helps :D
U N D E F I N E D