Diy Piezo Spring Reverb

[mac]

Rob,
your last picture will go viral

As soon as I finish playing with the speaker-piezo combo, I'll try a coil-magnet-spring-piezo with torsion force to get more delay with a shorter spring.

mac

[mac]

I replaced the speaker with a coil and a flat cylinder neodymium magnet.
I placed the magnet perpendicular to the coil, as seen in the picture below.
Again, the coil is driven with a LM386.
A speaker cone can also be used, and the impedance matching will be better.
But I used the coil to get more torque.

I am having a visible torque. I can feel it in my fingers too.
And when the magnet flat side is horizontal, the torque is gone.

I need to modify the circuit a bit, but the good thing is... it works



mac

[mac]

I uploaded a video showing standing waves in the speaker-piezo reverb. Just for fun

The spring in this case acts like a 1D gas in a tube, which is closed in one side and has a moving piston in the other side. In equilibrium, gas and air pressure are equal.
But if the gas inside is compressed a bit by the piston, and then the piston is released, there will be a net force acting on the piston, because the gas pressure in now bigger than the air pressure. The piston will move back to the equilibrium position.
The force exerted by the gas on the piston surface is PA or using ideal gas equation, nRTA/V=nRT/x
Assuming small displacements from equilibrium 1/x can be aproximated to A+Bx... and boila! harmonic oscillator!
(the demonstration is a bit more complicated, I tried to make it more DIY friendly -> https://en.wikipedia.org/wiki/R%C3%BCchardt_experiment )

This reverb is less complicated for a DIYer. A decent speaker, a LM386 or any small power amp, and a piezo.
But you a pay a price.
Delay time is shorter. You need a larger spring to get the same delay of a standard reverb can.
I'm getting short delay with long decay and lots of lows and mids.
Is it bad? Well, it's a matter of taste.


Rob attached some papers from Parker & Bilbao.
The delay time of a spring under torsion is R/r longer.
How is that?
Think the compression spring as kids in a row playing a silly game. When a kid is touched by the kid before, she/he has to touch the next kid. Kid 1 touches kid 2, immediately kid 2 touches kid 3 and so on.
In the torsion case, kids must describe a full circle with their arms before touching the next kid in the row. Kid 1 makes a circle then touches kid 2, and so on.
In the last case the speed is slower.

This is the best analogy I can think of to remark the differences without the use of Lagrangian or Hamiltonian Eqs

I'm still experimenting.

mac

[Rob Strand]

Looks cool.  I like your coin-flipper motor.  You will get a lot more force for a given drive current with a closed iron path similar to the reverb tanks.  With the coil you have it will get quite large, however, with the iron in place you can shrink down the scale of things somewhat.

[mac]

I did not put an iron because the magnet is so strong that it sticks to the iron.
Maybe a non magnetic material?

I also thought of a divergent field outside the coil, but it can generate transversal waves adding a more complex pattern.

I will continue experimenting when I return home. Too cold in my town, in NE Brazil right now 

mac

[Rob Strand]

I did not put an iron because the magnet is so strong that it sticks to the iron.
Maybe a non magnetic material?

Hmmm, that *is* a problem with those magnets.
It's probably a delicate balance of choosing the right size magnet, gap height and spring tension.
Keep in mind with the close magnetic path you will probably get away with a much smaller/weaker magnet.
You could spend days tuning that in!

The magnets on the reverb tanks are tiny and very weak.  That gives you an idea of the lower limit of a working system.

Some types of sticking, like sticking when you hit the assembly, could be solved with plastic sheet on the face.  This limits the minimum gap and the maximum force when the magnet is displaced.
If the magnet is so strong it sticks through with plastic present then the it's obviously too strong for the gap or spring.

I also thought of a divergent field outside the coil, but it can generate transversal waves adding a more complex pattern.

Maybe that would sound better?

[moid]

Without wishing to derail this thread,  I have a query about the springs in use in a spring reverb (something I am sort of working on at the moment, so this thread is very interesting). I recently bought a set of extension and compression springs, but in the great tradition of buying online bought some rather short springs instead of long ones. Can I connect a few springs together to make a longer spring, or would doing this destroy the reverb effect? If I can connect them together would using a hot glue gun be a good idea to reduce any damping effect on the sound to a minimum? Thanks.

[moid]

While I think of it, I presume I could make my own spring by just coiling a length of bare copper wire around a cylinder of some sort - would you know what thickness of wire would be best? I assume thin wire for more 'springiness', I can get hold of 0.25mm thick copper wire easily. The springs I bought are quite stiff and are 0.5mm thick.

Also I will be attaching the spring to a speaker at one end and a piezo disc at the other, would you recommend using hot glue for this or something else? Thanks for any advice you can give.

[mac]

Martin,
I am using solid Cu core wires, those used in alarm systems.
5m to 10m and a pencil will do it.

Is Cu better than steel? Cant say.

Hot glue something with a hole and a screw to the speaker. This way you can try different springs.
I am using an alligator on the piezo side.
Leave the piezo floating.
See my pics.

mac

[mac]

Rob,
If the magnet surface vector, blue line in the picture, is now 45° respect to the spring axis, transversal waves  be generated.

mac

[moid]

Martin,
I am using solid Cu core wires, those used in alarm systems.
5m to 10m and a pencil will do it.

Is Cu better than steel? Cant say.

Hot glue something with a hole and a screw to the speaker. This way you can try different springs.
I am using an alligator on the piezo side.
Leave the piezo floating.
See my pics.

mac

Thanks Mac

I thought I could get solid copper wire but it turned out to be braided I assume that would be worthless so did not buy any. However while doing some DIY for my mother this weekend I found a shop near her that sold piano wire at a very low price (bare steel wire, very springy, and a website told me that it is used to make piano strings AND musical springs!). So I bought 30 meters of it... and have now wasted an entire afternoon trying to make a spring from it . It will coil into a loose spring shape but it is so resistant to being bent into shape that it took me the whole afternoon to wind 2 meters worth of it round a pencil, and it refuses to be any less than 1m long now! So it won't fit in any enclosure and is not long enough for a reverb sound. I thought I'd let you know that whatever metal spring reverbs are made from; it isn't piano wire, so don't ever buy any. I have ordered some solid copper wire (0.6mm thick) online, hopefully it will arrive in a couple of days and I can continue with the project.

[PRR]

It IS piano wire, but their winder is much stronger than you can hack-up in a woodshop.

[Rob Strand]

Can I connect a few springs together to make a longer spring, or would doing this destroy the reverb effect? If I can connect them together would using a hot glue gun be a good idea to reduce any damping effect on the sound to a minimum? Thanks.

Many reverb units joined springs.  I noticed the join points on the spring are rigid and lossless they use small metal crimps.  You don't want any movement in the join as it could cause "creaks".
Maybe the best home friendly method is to epoxy or super-glue the joint but let it dry with the two springs under tension.

[mac]

mac

[moid]

It IS piano wire, but their winder is much stronger than you can hack-up in a woodshop.

Or even standing in my garden using a G clamp and brute force I did plug the spring in eventually and discovered it does not need to be as long as I thought - the 2 meter's worth of spring I have made does indeed create laser sounds and some hint of signal (mostly fuzzy noise, but it is there) however placing a piezo disc about 50 cm along the spring (from the speaker) does indeed give a reverby sound, although one that is far too loud on the bass strings of a guitar and too quiet on the high strings. I will work on making a high pass filter to calm the bass response down... although maybe that is the material? Mac, have you noticed if your copper springs make certain frequencies louder than others?

[mac]

...placing a piezo disc about 50 cm along the spring (from the speaker) does indeed give a reverby sound, although one that is far too loud on the bass strings of a guitar and too quiet on the high strings. I will work on making a high pass filter to calm the bass response down... although maybe that is the material?

If you use a LM386 to drive the speaker, lower the output cap at pin 5, that's your high pass filter.

Mac, have you noticed if your copper springs make certain frequencies louder than others?

I get a similar freq response.

mac

[PRR]

All the Fender reverb paths have heavy bass-cut.

There are MANY things going on. But the primary thought is that mechanical things we can pick-up with fingers do NOT like to shake 20,000/second. They are too heavy (massy).

Cone loudspeakers all roll-off above 200Hz. They are flat to >2KHz only because typical cone-size will not "grab air" below 1KHz. So a high-cut due to mass, a low-cut due to too-small cone, works out pretty flat over half the audio band. You are indeed likely to want a rise from bass to treble.

[moid]

Can I connect a few springs together to make a longer spring, or would doing this destroy the reverb effect? If I can connect them together would using a hot glue gun be a good idea to reduce any damping effect on the sound to a minimum? Thanks.

Many reverb units joined springs.  I noticed the join points on the spring are rigid and lossless they use small metal crimps.  You don't want any movement in the join as it could cause "creaks".
Maybe the best home friendly method is to epoxy or super-glue the joint but let it dry with the two springs under tension.

Thanks Rob, I solved the issue using the video Mac posted - using a drill to wind the spring creates a (with copper anyway) tighter coil that contains a lot of metal in terms of length, so I don't think I need to join multiple springs now (I hope).

[moid]

...placing a piezo disc about 50 cm along the spring (from the speaker) does indeed give a reverby sound, although one that is far too loud on the bass strings of a guitar and too quiet on the high strings. I will work on making a high pass filter to calm the bass response down... although maybe that is the material?

If you use a LM386 to drive the speaker, lower the output cap at pin 5, that's your high pass filter.

Mac, have you noticed if your copper springs make certain frequencies louder than others?

I get a similar freq response.
mac

Thanks for those videos Mac They enabled me to try the below:


These are two steel springs from piano wire and two copper springs. The shorter ones have 1.5m in length of wire, the longer ones contain 3m of wire. I haven't tested the longer wires yet (everyone here has gone to bed, so it will have to wait until tomorrow). The shorter wires do have a sonic difference to them. Admittedly it's not the fairest of tests; the steel spring is not a great spring, but it was made in exactly the same way as the copper spring - it demonstrates that doing this in steel without a special machine for it as PRR noted is a semi waste of time - the spring just uncoils all over the place. The copper holds its form well and I would recommend to anyone else making their own springs to use copper in future. The two metals sound different. The steel spring sounds more like spring reverb sounds - more metallic and hollow, although the 1.5m length isn't long enough to create the drippy spring reverb sound (which I don't mind). The copper spring has a warmer, softer reverb sound. I prefer the copper sound, but the steel is interesting;  might put both in the tin box I have ready for the springs and use a pot as a mixer to blend between the two sounds as they are reasonably distinct. Both require low frequencies to be removed, but the copper is not effected as badly by low frequencies as the steel spring is.

I decided not to change the cap on the amplifier (thanks for suggesting it though), instead I will add two high pass RC filters after the audio has left the amplifier circuit. The reason for this is I want to blend the dry signal from the amplifier back into the audio after the springs so I can have full undistorted clean sound (all frequencies) plus the reverb with reduced bass frequencies.

[moid]

All the Fender reverb paths have heavy bass-cut.

There are MANY things going on. But the primary thought is that mechanical things we can pick-up with fingers do NOT like to shake 20,000/second. They are too heavy (massy).

Cone loudspeakers all roll-off above 200Hz. They are flat to >2KHz only because typical cone-size will not "grab air" below 1KHz. So a high-cut due to mass, a low-cut due to too-small cone, works out pretty flat over half the audio band. You are indeed likely to want a rise from bass to treble.

Thanks PRR - I didn't know that about Fender reverbs, but it's nice to hear that it's not something I've done wrong but an actual physical issue that professionals have had to deal with. I suspect the low cut is required severely in my circuit because the speaker cone I am using is only about an inch across. It sounds OK when I listen to sound coming through it (it's obviously not bass heavy or particularly loud) but it reproduces higher pitches best. A larger speaker would be better for bass, but I don't have any of those spare at the moment and there isn't enough space in the enclosure I'm using to put a larger speaker in it.