Spring Reverb: reducing noise

[Gene Pink]

Ive been fooling round with the idea of spring reverbs for years ,always wondered how a spring in vaccum  might sound  ,maybe it could be used as a  time control ,the less air resistance on the springs ,surely the longer the vibrations would carry ,be interesting to mess around with ,bell jar spring reverb anyone?

That's a really interesting idea. Wish I had a bell jar. One of these days.

I'll bet the vast majority of the damping does come from the air. Dick Dale would want one.

Thinking about it, 6" ID sch 80 PVC pipe, one glued end cap, one bolt flange and removable plate, figure out bulkhead connections (are BNC chassis-mount connectors vacuum-rated?) and two NPT ports, one for the pump, one for a vacuum gauge calibrated in 'verb tail-length....

Or different gasses, helium for long tails, xenon for short tails. Under water for really short tails.

Just musing.

Gene

 

[Tubetec]

ahahah nitrous for extra giggles ,helium to shift up an octave ,ahah

 

[Tubetec]

probably a lot more output signal with the air gone too ,make an interesting experiment

 

[JohnRoberts]

That's a really interesting idea. Wish I had a bell jar. One of these days.

I'll bet the vast majority of the damping does come from the air. Dick Dale would want one.

Thinking about it, 6" ID sch 80 PVC pipe, one glued end cap, one bolt flange and removable plate, figure out bulkhead connections (are BNC chassis-mount connectors vacuum-rated?) and two NPT ports, one for the pump, one for a vacuum gauge calibrated in 'verb tail-length....

Or different gasses, helium for long tails, xenon for short tails. Under water for really short tails.

Just musing.

Gene

I wouldn't expect much damping from the surrounding gaseous atmosphere due to relative mass, but according to a google search there were early Hammond reverbs that placed the spring in an oil bath that could increase damping. (not sure why that would be desirable , but it would sound different. )

I vaguely recall (decades ago) discussion of making a physical acoustical reverb chamber into a smaller volume but slowing the speed of sound transmission to simulate a larger space's longer paths. I don't think this was practical because don't recall hearing about it again. 

JR

 

[Matt C]

Ran a couple quick tests on my unit before I start modifying things.  All tests were run with the unit racked normally, no input signal, with other nearby units powered on. 

First I tested the noise of the whole circuit, then just the recovery amp (tank disconnected, recovery amp input shorted), then the tank AND the recovery amp (drive amp disconnected, tank input shorted). 

Here are the results I recorded in my DAW, analyzed with Voxengo SPAN. 

Full circuit: -65dBFS (rms)
Recovery Amp: -102dBFS (rms)
Drive Amp Disconnected: -65dBFS (rms)

In all cases the noise was predominantly 60Hz hum plus odd harmonics.

So this seems to confirm that the noise is in fact getting picked up by the tanks themselves, not the surrounding circuits.  The noise performance of the recovery amp itself is maybe not super impressive but certainly acceptable in this context.

I'll apply some of the ideas you guys mentioned and see if I get better results.

 

[Matt C]

Had some time to experiment with a few different configurations today and got pretty good results.  Here's what I tried:

First I adjusted the tanks' i/o scheme so that the inputs were in phase and the outputs were out of phase (opposite from original configuration), hoping that would create a humbucking effect on the output side. In fact it made no difference, noise was still at -64dBFS.  One thing I did not try was grouding the "center tap" of the out of phase output coils.  Not sure if that was my fatal flaw or not.

Then I physically rearranged the reverb tanks in their chassis.  Originally the outputs coils were on the right side of the chassis, which is where other units commonly have their power transformer mounted.  I flipped the tanks around so the outputs were on the other end of the chassis, hoping the physical distance would help.  This made a very noticeable difference. The noise was now measuring at -67.5dBFS, so not a huge reduction, but doing this drastically reduced the amount of buzzy mid-range harmonics in the signal, so the remaining noise was not nearly as offensive.

Then I tried putting the tanks' chassis in the closet, about 6-8' away from the rack or any other electronics.  Predictably this gave the best performance, noise was now down to -85dBFS.

I also tried putting the tanks back in the rack and modifying the recovery amp to have a more significant hi-pass built in.  This helped too, cutting noise to -71.5dBFS, still with the reduced midrange buzz from reorienting the tanks in their chassis. This configuration sounded very usable to me and I'll probably keep it like this, although it is tempting to mount the tanks in the closet permanently...

 

[Tubetec]

20 db reduction in honk and buzz ,thats alot more usable ,
I wonder what the theoretical maximum reduction achievable would be ,
with a 2250 ohm output coil i have wired balanced into a mic preamp the self noise still swamps the mic amp noise by maybe 20 db or more ,driving a 20k line input with the gain cranked up seemed to just make matters worse.
Probably  a few hundred millivolts  is the max signal you'll get out of the tank at any rate .

 

[Bo Deadly]

There ya go. So -20 dB is pretty clearly going to be worth isolating the tanks. It's very hard to shield from electromagnetic interference because unlike electrostatic interference magnetic fields will go straight through metal. Mu metal has some magnetic shielding properties but I don't think it would give you 20 dB and it might be awkward trying to make some sort of mu metal case around the recovery coil. So isolation is probably going to be the path-of-least-resistance. I think I read somewhere about how the plate reverbs at Abbey Road are in a shack on the roof.

If I were doing this (and I probably will get around to doing this someday), I would make a little box or 1U enclosure with one of those Neutrik RJ-45 sockets for a shielded CAT6 cable run to the remote chassis containing the reverb tanks. CAT6 gives you 4 twisted pairs (plus the shield which can be used to ground the chassis' together). One pair to transmit, two pair for stereo returns and the final pair to power two +40 dB recovery amps. I might also consider using a transformer with an 3.2 ohm secondary (Hammond 109E would be high-end but a cheap one from Ebay would work equally well since a reverb tank isn't high fidelity anyway). That will allow a) the drive signal to be relatively high voltage over the long CAT6 cable and b) you can drive both 8 ohm tanks wired in parallel with ease and low distortion.

Also, it might be worthwhile moving it around to see where noise is the least. Even in a closet there might be wires in the walls that could induce noise. And try acoustically low noise places as well. Any vibration at all could be picked up by the coils and cause rumble. If you can hear someone playing drums in the next room, that could easily be picked up. CAT6 is nice because you can drill a hole in the floor, run it way down into the basement (away from any pumps!) and crimp a new connector on it and now it's in a really quiet place.

Another important thing that should be obvious but I'll mention it anyway is to use something like an AUX send and then mix the wet-only return signals with the origial signal in your mixer / DAW software. Meaning don't mix the dry outside of your mixer / DAW or you will just be adding the noise of that. You want wet-only reverb return and then you can mix in as desired. Usually it's only a little bit which means you're only adding a "little bit" of the associated noise.

 

[abbey road d enfer]

It's very hard to shield from electromagnetic interference because unlike electrostatic interference magnetic fields will go straight through metal.

That is not true. Ferromagnetic materials provide efficient shielding when properly applied. Indeed no magnetic shield can provide 100% efficiency, but there is clear evdence of mu-metal cases providing 20+dB shielding in mic input transformers. And like in acoustics, double insulation makes a very efficient use of materials.