Diy Piezo Spring Reverb

[Marcos - Munky]

I was kidding on drawing the schematic, but thanks a lot . Nice and simple. Need to get me some solid wires.

[mac]

Need to get me some solid wires.

I like copper, it's softer than steel springs of the same gauge. But that's my taste.
I was about to try the wire around the core of wound strings too.

I was kidding on drawing the schematic, but thanks a lot . Nice and simple.

I had to draw it by hand since I'm in bed right now
It needs to be polished, a better mixing, feedback, etc.

mac

[Rob Strand]

FYI, I dismantled a spring reverb the other day and started taking measurements of the springs.   Haven't had a chance to get back to it.  I had some unexpected results.   The springs have a degree of pre-tensioning in them ie. you need to apply a minimum force before the turns separate.   I realized that when I got different spring constants at different extensions.   BTW the only reason I found that out is because I removed the springs.  Initially I calculated the spring constants from the dimensions which would have given me incorrect results.

[mac]

http://mae3.eng.ucsd.edu/machine-design/force-torque-power-analysis/ftpe-springs

mac

[Rob Strand]

http://mae3.eng.ucsd.edu/machine-design/force-torque-power-analysis/ftpe-springs

Thanks, at least it confirms the spring model I'm using for pre-tensioning.

I'm using one of these to measure the force,

http://media.rs-online.com/t_large/F0432211-01.jpg

Measuring deflections of 1mm using adhoc methods is a bit dubious.  It's not that important anyway I only want a reasonable ball park.

Anyone know how pre-tensioned springs are manufactured?  I can't imaging them winding them hot and cooling them there has to be a trick.

[mac]

Measuring deflections of 1mm using adhoc methods is a bit dubious.  It's not that important anyway I only want a reasonable ball park.

If the spring follows Hooke's Law you could measure at say 1cm and the costant K should read the same.
Another silly way to get K is to hang a small weight, Mweight>>Mspring, and measure how many times the weight goes up and down in a given period of time, say 30 seconds. Then divide 30sec by N and that's your period. Roughly. Do it many times to reduce the counting error. Statistics
But you need to know Mw 
2.pi.f=sqrt(K/M)

Anyone know how pre-tensioned springs are manufactured?  I can't imaging them winding them hot and cooling them there has to be a trick.

That's what blacksmiths have been doing for centuries, fire, hammer and water... fire, hammer and water...

mac

[Rob Strand]

If the spring follows Hooke's Law you could measure at say 1cm and the costant K should read the same.

Initially I had a few points for large extensions but depending on what points I used I got different numbers.  That's when I knew the springs were pretensioned.   I then did a measurement at small extensions like 1mm to 5mm just to confirm the problem.    I've actually got a least squares line fit now, which essentially does the statistics for me.  It won't get rid of systematic error at small deflections so I'll chuck out those points.

The reason I'm not concerned about the accuracy is that I'm sure even good sounding springs vary a lot.  Moreover, it seems the spring constant isn't an important factor, at least not directly.

But you need to know Mw    2.pi.f=sqrt(K/M)

I remember doing that in physics labs.    The problem is the spring has mass distributed along the length and you need to take that into account somewhere/somehow.   My favorite standard masses are using coins.

hat's what blacksmiths have been doing for centuries, fire, hammer and water... fire, hammer and water...

At 4am this morning I woke and it came to me that tensioning the wire while winding might also work.   That method is probably more controlled.   The only concern is will the wire break to get the pretension you see in the real springs.  And is that what manufacturer's do?  (How to do it is only for interest anyway.)

I'll try to post some more stuff today. 

BTW, what is the diameter and number of turns (or length of wire) of your good spring?  I'm assuming it's 22AWG wire.  I was going to try to compare it against the spring reverb springs.

[mac]

The reason I'm not concerned about the accuracy is that I'm sure even good sounding springs vary a lot.  Moreover, it seems the spring constant isn't an important factor, at least not directly.

Young's modulus and spring constant,
https://www.physicsforums.com/threads/difference-between-youngs-modulus-and-spring-constant.887403/

Lenght, tension, mass, cross sectional area shape, speaker in this case, etc, are likely to be more important than the spring constant.

mac

[mac]

BTW, what is the diameter and number of turns (or length of wire) of your good spring?  I'm assuming it's 22AWG wire.  I was going to try to compare it against the spring reverb springs.

The thicker spring was used in the 2nd sound demo. It was from the secondary of a 220v-6v wall wart. About 240 turns.
Looks like half a mm, so it could be 22 awg or higher.

The shorter and thinner spring was used in the 1st demo.





mac

[Rob Strand]

The thicker spring was used in the 2nd sound demo. It was from the secondary of a 220v-6v wall wart. About 240 turns. Looks like half a mm, so it could be 22 awg or higher.
The shorter and thinner spring was used in the 1st demo.

Awesome mac thanks.  I'll plug in those parameters and see how it goes.

[Rob Strand]

Finally got a chance to get back to this.  I found this paper on calculating reverb delay times,

http://research.spa.aalto.fi/publications/papers/dafx09-sr/
http://dafx09.como.polimi.it/proceedings/papers/paper_84.pdf

Then there's this paper which discusses loading (ie. stretching or compressing) the springs,
http://newt.phys.unsw.edu.au/music/people/publications/Fletcheretal2001.pdf

From what I can work out so far.    The behaviour of the coil doesn't depend so much on the tension in the coil at the operating position.  It depends on the length of the coil at the operating position (H)  (and not the free length L0).   These papers calculate Delay Time not Decay Time.  Decay time relates to damping.  Also note R is the average radius.  In the table below I used Dav, the average diameter.

The tension in the coil prevents it from sagging.

The first paper gives some example of springs.

I disassembled two reverb tanks and extracted the spring data.

Old reverb from an organ.  Probably early 70's.
No brand the only marking is "Mod 003.930.861" which I assume is the model


Accutronics 1AB_D1_  from an old Washburn amplifier (probably an SX60C).
I don't know the full number because the top of the reverb tank is corroded.  I found it on the side of the road.  It was only the box and the reverb.  The speaker and amp chassis were removed.


BTW all the reverb springs have pretensioning.  I suspect this is done to keep the length of the spring down?

I plugged data for the various springs in the formulas from the first paper.   So there's:
- the examples in the paper
- the springs for the two tanks I measured.
- your two handmade reverb springs.
- an example where I tweaked the dimensions to increase the frequency response



The key results are Td and Fc.  Td is the delay time and Fc is the upper cut-off frequency.
For the split spring I added the Td's for each spring calculated individually.  For Fc I'm not sure
how to combine it correctly.  It will be between the values in the table and the values I annotate at the bottom.

By playing with the coil diameter it is possible to tweak the frequency response (Fc) and then you tune the length to give you the desired delay.   I've added an example where I took your second coil and reduced the diameter then tweaked the delay to be in the order of the commercial units.

According to Accutronics they give this guide:
LONG   41 ms   (guitar)
MEDIUM   37 ms   (organ)
SHORT   33 ms   (voice)

[dschwartz]

Just an observation..
You are feeding the piezo transducer into a really low input impedance transistor stage.. i would use a jfet or an opamp with at least 10meg input impedance if you want decent frequency response from it..

[Rob Strand]

Just an observation..

There's also the current drive vs voltage drive issue.


Interestingly I found a reference to a commercially available piezo based Reverb Tank from a 1976 Electronics magazine:

[mac]

Just an observation..
You are feeding the piezo transducer into a really low input impedance transistor stage.. i would use a jfet or an opamp with at least 10meg input impedance if you want decent frequency response from it..

Yeap.
A TL072 can do it, one ic as the input buffer, the other to mix the signal.

mac

[thermionix]

Interestingly I found a reference to a commercially available piezo based Reverb Tank from a 1976 Electronics magazine

The 60's Danelectro/Silvertone amps (1484 etc) used a piezo tank, didn't they?  I think those were made in-house though.

[Rob Strand]

The 60's Danelectro/Silvertone amps (1484 etc) used a piezo tank, didn't they?  I think those were made in-house though.

I don't know.   Getting out of my knowledge zone.    Probably a good way to avoid stepping on Hammond's toes.

I remember some DIY spring reverbs in 60's electronics magazines that used a ceramic cartridge (from a stereo) for the sense end.
--------------
Edit
Apparently 1960's Danelectro 9100 tube spring reverb unit.
Also saw VOX used a commercial ceramic cartridge.

[thermionix]

I've got a wacky one here.  I recently tore apart a dead '67 Silvertone solid state amp (scored several nice Si and Ge transistors, most are in pedals now).  The reverb tank is in a cylindrical cardboard tube.  It has a shielded wire (output?) and two unshielded red wires coming out.  I get about 60 ohms across the red wires, but the shielded wire reads open.  Maybe it was driven with a transducer, but has a piezo pickup?  Who knows!  I haven't tried to open it because I might list it on Ebay.

[mac]

Thanks Rob, a good reading while I'm still in bed. (thanks Yellow Fever vaccine )

I'd like to take a look at the Love-Kirchhoff theory.

Parker & Bilbao say: "The magnetic bead is driven torsionally by passing a signal into
a nearby electromagnetic coil. Torsional vibration of the magnet/wire
system translates to vibration tangential to the path of the
wire at a point within the helix."

The ratio R/r on the Fc eq. looks like the rotation of one closed turn.

It makes me wonder,
-the speaker gives a nice punch to the spring along the axis. The movement of the turns is visible, so the spring must be expanded enough to avoid turns hitting each other.
-and the edges of the coil are fixed at both sides, no rotational force at all.

This gave the idea to use a 2D sawtooth like Fig 1 in the Parker & Bilbao paper.

mac

[diffeq]

It makes me wonder,
-the speaker gives a nice punch to the spring along the axis. The movement of the turns is visible, so the spring must be expanded enough to avoid turns hitting each other.
-and the edges of the coil are fixed at both sides, no rotational force at all.

This gave the idea to use a 2D sawtooth like Fig 1 in the Parker & Bilbao paper.

mac

What if you fix the spring with one of those swivels, allowing it to rotate?

[mac]

What if you fix the spring with one of those swivels, allowing it to rotate?

Good idea.

Those in particular can introduce noise with a loose spring like the ones I used.
With a tighter spring, a plastic/metal swivel can work without introducing noise or damping.

mac