Spring Reverb - hows it work?

[yano]

I saw that post about the crate with Spring Reverb, which made me realize my own amp (a TOA KD-3) had spring reverb. When I first saw the box, I thought maybe it was a shielded transformer (my best guess, but I was still confused), because it had Main and Secondary stamped on the box.

Anyway, my question is, how do these units work? Could a Spring Reverb stompbox be made?

[aron]

Yes, Craig Anderton posted a circuit long ago. I believe there are many on the web too.

Here's one:

http://sound.westhost.com/project34.htm

[Mark Hammer]

The thing about springs is that any energy directed towards them can take a long time to completely dissipate.  Especially if they have the right properties.  Next time you take a bath, wait until the watr is perfectly still  and just watch how long it takes for the ripples to stop after you smack the water's surface.  Same thing with springs.

The tricky part is turning that continuing motion into an audio signal.  Reverb "pans" do that by sticking electrical tranducers at each end of the springs.  One end makes the springs wiggle when you feed voltage/current into it, and the transducer at the other end produces a small voltage when the spring wiggles.  What is fortunate for us is that the "wiggles" (no, not those Australian kids entertainers!!) contain a representation of the original input signal (although without the same bandwidth as the original).

Although well-designed springs have the desirable properties that they mimic what multiple unsynchronized reflections sound like (including the reduced bandwdith that come from bouncing off walls, ceilings and floors), they have a few shortcomings.  One is that they are pretty inefficient systems.  The signal you get out of the receiving end is puny and the amount of boosting you have to do to make it usable can inject more noise into the system.  There are ways around that though.  

The other shortcoming is that as a mechanical system, ANY mechanical stimulus can produce wiggles in the spring and an audio signal at the other end.  I suppose one way around that is to have a noise gate at the receiving end which is actuated by the incoming audio signal.  Since any kicks and bumps that produce wiggles in the springs have not passed through the input stage, they should be gated out.  Come to think of it, that's a nice feature.  Wonder why it hasn't been implemented in commercial units?

[SeanCostello]

I always thought that kicking a spring reverb was one of the coolest features about them. You can't do that with a Lexicon 480l!

I suppose the noise gate would have to have a release time equal to the T60 of the reverb, in order to avoid cutting off any signals that might still be decaying.

Sean COstello

[Mark Hammer]

Well, kicking the springs IS a cool effect, except for when it's not you doing the kicking and banging.  One of the difficulties of many stage situations is that there are a host of unwanted vibrations that make the springs jiggle.  You may not notice them when you're playing, but stay quiet for a moment and they can be irritating.  Again, it would depend on one's playing context and setup.  In a studio situation, vibrating kickdrums won't be a problem for what you have in your control room rack.  In many guitar amps, the reverb pan is installed in a manner that tries to reduce extraneous sources of vibration.  For example, the chassis often has the spring unit itself suspended on a kind of "hammock" by another set of springs that are not part of the audio path.  Many amps have the reverb springs enclosed in a vinyl bag, or shock-mounted to the amp chassis.  Companies that manufacture reverb pans also provide instructions/advice on how to mount them to minimize such extraneous noise.

[Hal]

i started making chainmaille (armour) not too long ago, and as part of this, I make springs (to cut into rings)....

I was wondering for a while how possible it is to make your own reverb _tank_.

First, would aluminum springs work, second, how close should the things be (they're ususally touching), and last, how would I measure the impedance or whatever tanks are measured in....

[Mark Hammer]

Geez I gotta finish that project so I can stick up a URL instead of writing it out every time, but here goes...

The springs need to be:

a) Long and articulated (lotsa coils) enough that it takes a while for the mechanical signal to travel from here to there.

b) Light enough that it doesn't take much energy to create vibrations.

c) Stiff enough that the wave created at one end  doesn't lose momentum and fail to reach the other end...especially given how subtly the transducer will shake the springs.

d) Compliant/soft enough that it will vibrate over a wide range of frequencies.

If the springs can meet those criteria, then I don't think it matters what they're made of.  Having said that, it may be harder for some materials to meet those criteria than other materials.  For instance, can aluminum easily adapt to making springs with 20 or so coils/turns per inch?

One of the things you'll see in many reverb spring pans is the use of two springs end to end (and soldered together in the middle) as a functional spring.  So, if it is a 3-spring reverb, it is actually a 6-spring unit, but there are 3 sets of two springs in series.  Don't confuse the number of separate parallel transmission lines with the number of physical springs it takes to produce those transmission lines.  My guess is that since you need them soft AND long, they can easily sag too much in the middle, which would impair the physical wave travelling the full distance.  Consequently, the solder joint in the middle (which connects two end loops such that there is maybe 3/4" in the middle that is stiff and not coiled) is likely there to recover some stiffness and step around the problem of how to support a long spring without damping the vibrations.

The history of spring reverb is pretty varied.  Not all spring reverbs have been like what we're accustomed to seeing in the back of a Fender amp.  There have been plenty of short spring pans (many budget amps still use them), single spring mini-reverbs were used in car radios, believe it or not, and the springs don't HAVE to be in a straight line (Peter Snow can tell you about his boyhood project that involved a Y-shaped transmission line.)

I've been working on a homebrew reverb project (that has a little too much dust on it now for comfort :cry:  ) for a while that uses some springs I bought from Home Depot.  Compressed, they are about maybe 4", and stretched just a bit (with lots of compliance left), they are about maybe 6-8".  The plan is to connect 3 together in a Y-formation and install them in a small chassis.  Conventional reverb pans use a direct transformer/magnet coupling.  A slug is situated in the middle of a transformer-like device with a coil around it and current passing through the coil makes the slug move around.  At the other end, a second slug is made to wiggle inside the gap of another coil and that produces a voltage.  It works, more or less, like a loudspeaker voice-coil at the incoming end, and a moving magnet phono cartridge at the outgoing end.

You don't have to do it that way, though.  As long as you can make the spring wiggle to music, and produce an AC electronic output, you're fine.  So, I have the "speaker" end of the coil epoxied to the middle of a voice coil dome on a small (2") computer speaker, and the outgoing end soldered to the middle of a piezo disc in a telephone.  The minispeaker is fed via either a 386 power amp or an NE5534 op amp.  The speaker has a fair amount of the cone surgically removed so that the speaker *behaves* like a speaker but moves much less air so you don't hear it.  The crystal mic cartridge end up putting out about 60mv, which is a mice level, requiring very little additional gain for further processing.

My intention is to use the Y-coupling so that the one speaker drives a single spring then 2 springs feeding separate mic cartridges for a more diffuse sound (less obvious boing).  In principle, if you mount the transducers on each end on a sort of bracket, one should be able to adjust the tension of the spring by moving the ends further apart or closer together, to yield different reverb sounds.  The whole thing fits in a box that is about 4" x 9" x 2", with power supplied via a wallwart.