FSR sensor compatible with Megadrum?

[Rubis]

I am testing the special firmwares today, I probably won't get to testing the hihat controller until later this week or next.

[jmcdougall]

Over the past while I have been able to do some research into how FSR's are implemented. looking at other equipment, there is one significant difference in most implementations. Piezo's are treated as an analog signal and so like megadrum are processed through A/D converters to provide the trigger for note on. The digital values are used to provide the threshold control to be able to tune out crosstalk which is a problem with piezos as they will respond to any vibration. On the other hand, FSR's have no need for this so the output can be set so it acts like a direct trigger. Most of teh designs I have seen, feed the FSR output directly to a CPU pin. At rest, there is 0V and at impact either 5V or 3.3 depending on the chip. Because they are FSR's, it should be possible to design a small circuit that would provide power to the FSR as well and convert the signal from a FSR type to a piezo trigger type.

Dimitri, could you explain what you meant by not supported in the firmware? If I understood what you meant, i might be able to come up with a way to support the FSR's with an external mod

[dmitri]

Most of teh designs I have seen, feed the FSR output directly to a CPU pin. At rest, there is 0V and at impact either 5V or 3.3 depending on the chip. Because they are FSR's, it should be possible to design a small circuit that would provide power to the FSR as well and convert the signal from a FSR type to a piezo trigger type.

Will it produce intermediate voltage levels between 0 and 5V?
If it will, will it be sensitive to light hits? If it will be sensitive to light hits, how can it differentiate light hits from crosstalk vibration?

Dimitri, could you explain what you meant by not supported in the firmware?

It means I have not programmed anything FSR specific into the firmware.

If I understood what you meant, i might be able to come up with a way to support the FSR's with an external mod

If this mod will produce variable signals, don't see why it could not be supported by MegaDrum.

[jman 31]

How about using the arduino as an expansion and programming the arduino to send the midi signal to the MD. LIKE THIS but with the FSR.

[dmitri]

As I understand, FSR are looked at as "crosstalk-less" alternatives to piezos. As such they have to produce variable signal and be sensitive. Sensitive and "crosstalk-less" do not look like belonging in one sentence.
How can Arduino help here?

[Rubis]

They are probably the most simple variable hihat design I have found. You can't get much easier than a piece of foam sandwiched between two copper plates. I was planning on making one earlier, but wasn't exactly sure what kind of foam to get. I don't really know how they could replace piezos though.

Here is an instructable on how to build one. http://www.instructables.com/id/DIY-For ... istor-FSR/

[jmcdougall]

FSR act as the inverse to a piezo. At a steady state they represent a constant resistance. As the pressure builds, the material deforms and the deformation results in a change in resistance i.e. typically the resistance lowers. Think of it a a potentiometer. Deformation has the same result as turning the pot. Because they are pressure driven, they are immune to carrier vibration, therefore no x-talk.

There are two approaches that could be used. One would be to feed one side with a positive voltage +5 and then the output would vary between 5 - xV depending on the latent resistance and 5 -yV depending on what the minimum resistance at full deformation. In theory, this could look similar to a piezo. However, from a cabling perspective, this gets messy if you use standard mono 1/4" jacks and cables. Since you have two lines carrying signal, your shield is live and there is no ground. This would mean going to stereo cabling and risks if someone accidentally plugged a mono plug in - instant short.

I think a better method would be to use the FSR as a connection to ground. A limiting resistor would control the current that is being sunk and the voltage could be also fed to an inverting opamp. In other words the input to the opamp would come from the connection point between the resistor and the FSR. The FSR would cause the voltage to go from +5 to +5 - x. By feeding this through an opamp, this could be inverted to 0 to +5-y providing a signal that should be very compatible with a piezo input. This keeps the cabling simple as it uses std 1/4" mono plugs and the ring ground is still ring ground while the tip is the 5V feed.

I have an old junk DrumKat lying around that I might be able to test this out over the weekend. The real trick here is understanding what the current sinking capability of the FSR is so that you can limit the current so you don't fry the FSR. This is the same kind of issue that you deal with with LED's. The dropping resistor is to limit the current so you don't burn it out

[dmitri]

I'll forget about the cost associated with using FSRs for now hoping for zero crosstalk quality.
Let's assume you made an FSR pad which produces a signal similar to a piezo pad. It is very sensitive, just as a piezo pad, i.e. a lightest hit produces a signal. Now, this lightest hit applies very small amount of pressure on the FSR but a vibration will just as well apply the same small amount of pressure on the FSR because we know what vibration (sound) is. So at this point I don't see how it is going to be better than a piezo. Am I missing something?

[jmcdougall]

Piezo's are solid devices and produce voltage by the deformation of the crystal trapped between the two outer surfaces. Crosstalk comes about because vibration is transmitted through the solid surface on one side (mounted) while the other side remains relatively stable because of inertia causing the crystal to twist and produce voltage. FSR's are made with layered polymers so the vibration is muted by the bottom layer. It is possible to induce x-crosstalk but as a triggering device it is less susceptible because of the materials used. Also, because of the way you handle interfacing, it is easier to set the bottom threshold to null out x-talk.

Earlier I suggested using a two wire approach because of simplicity. I checked with the two major suppliers and they both recommend the three wire approach with the resistor on the ground leg side. i.e. voltage to FSR then through resistor to ground. The input voltage is pulled from the joint of the FSR to resistor and fed to an opamp to invert. There is a lot of good information on the Alternate Mode forum under the DIY topic

[dmitri]

Are you going to test this idea and report about your findings?