DIY electronic drumkit notes

From Helpful
Revision as of 10:50, 29 July 2021 by Helpful (Talk | contribs)

Jump to: navigation, search

DIY electronic drumkit

look, I was focusing on the electronics, okay?

The idea: build a cheap electronic, velocity-sensitive, MIDI-output drumkit.

There are many simple and cheap DIY ideas out there. Check the usual video sites.

This is neat DIY project, in that it has analog and digital and physical things to figure out, all manageable but some quite educational. Also satisfying result of something that makes sounds.


You may care about a few different things

getting enough energy to the sensor, and consistent amount of energy to it
more makes it easier to reliably measure impact and its velocity
detecting velocity
and getting it reproducably
getting minimal energy to neigbouring pads - i.e. physical isolation from others
avoid false triggers (without having to hard-ignore pads, which would be bad for real drumming)
allows us to be more sensitive
allows some choice between finger drumming and stick drumming (there's a large energy difference)
allowing fast drumming, without false triggers
we'll discuss why this is a thing later

having it be useful for practice
pads that aren't loud
half the reason is fewer complaining neighbours
relevant for stick drumming, not finger drumming
getting the pads to physically bounce roughly like their real-world counterpart
do hihat pedal properly
and potentially things like muting

Physical design

pad bounce experiment: liquid latex formed in a piece of tupperware. The left one looks cruddy because I was impatient and took it out before it dried. The bounce works surprisingly well.

A few of those can be solved largely in physical design, making your life easier once you get to the electronics and code.

if you only want finger drumming, then life is much easier because the forces are much smaller, and so barely carry to neighbouring ones. I'd recommend your first version be this.

Beyond that, the question is how much you want it to be practice for real movement - in which case you need positioning makes sense and doesn't fall apart, a hihat pedal, etc. The cheapest nearly-toy e-drums are not that much more expensive than what you'ld spend making them. Particularly second had

This is a kickdrum (a.k.a. a piece of wood)
...which is nothing more than a piezo (needs better attachment) and a socket

If you want to play as hard as regular drums this is a little harder to design well.

Say, you probably do still want thing to stay in place, but a solid construction will also be better at hard-coupling energy to the other sensors, so the more sensitivity you want (and the more velocity sensitivity you want) the more this becomes a design tradeoff you have to deal with.

If you don't care about the physical loudness of hitting pads, or the physical bounce, then it's enough to put a piezo on the back of a piece of wood

If you do care about the bounce and loudness, and about crosstalk and the false triggers that may cause on closeby pads, then the materials above and below the pad matter.

Good physical isolation is a little more work than you'ld think, if you want an otherwise quite solid thing. (For example, I at one point had a MIDI guitar, which has piezos detecting the picking of each stringalike, which if you pick them hard enough will trigger the next one over even though they're mounted in rubber. Could be lessened with filtering by velocity at MIDI level, but that a commercial product has this issue is an indication of the problem)

These materials also change the strength of the signal and the resonance of the pad, so means your code shouldn't assume too much about the design, and means you may want at least basic calibration in the velocity detection part.

Many designs use a piezo sensor because they're cheap and work pretty well. For finger drumming the constuction can be small, for harder drumming you probably want decent-sized pads for aiming reasons, and the implied large surface is nice for a consistent amount of energy and resonance to measure, which also makes it a little easier to get a more predictable amount of energy.

Keep in mind that on a solid board, the main resonant node is typically in the middle, so that's where you get best response.

Also keep in mind that you want to attach a piezo in a way that transmits vibrations well, doesn't fall apart immediately, or preferably at all. A dollop of wood glue works well enough for testing, but likely to fail sooner rather than later. Something entirely solid holding it in might damage the piezo over time. A holder with something soft may dampening the response a little, for better or worse.

I plan on modifying something like [1] with space for rubber (to connect) and foam (to isolate).

Also keep in mind that the wires on the piezo also carry shock directly to the piezo, so you probably want to isolate at least the last part of wire (Bit of foam, hot snot, or whatnot) from the part that might move.


The piezo in a circuit

Piezos are essentially stress sensors, so respond well to force/weight, impact, and vibration including that caused by impact.

Due to their design, they can be seen as stress-to-voltage sensors, but also have a decent amount of capacitance, on the order of nanoFarads capacitor.

You can model a piezo as three things in parallel

  • a stress sensor in parallel
  • a few-nF capacitor
  • a GOhm-scale(verify) resistor(verify)

That combination means that, without any further components, it will hold the stress-generated voltage for a relatively long time, to the point that if you hook up a piezo directly to an ADC, you get what looks like a fairly steady voltage after each hit, that is at a different level each time.

Resting a weight on the pad will give a proportional high value.

You may even get small changes over time and with temperature as the piezo and/or the thing it's attached to expands or shifts.

You can write code to deal with such shifty input values (largely subtracting recent average), but it's messy, and there's more edge cases where you can still get wild triggering, e.g. the voltage going out of range, which you can't really control.

So many designs add a resistor across the piezo.

This is essentially a discharge resistor for the capacitance, which means the voltage output is now primarily the changes in stress, and will go to 0V the rest of the time.

This setup resembles an RC lowpass filter, so you can quantify its impulse response.

Since a piezo disc's capacitance will probably be in the ballpark of 500pF to few dozen nF (depending on size), and you probably want dropoff within a few dozen milliseconds, you want a resistance on the order of 100kΩ.

(More widely something in the range of 10k to 1M can make sense, varying with specific components and further design choices. If you wanted to make a drum interface separate from the impact sensors and make it "plug in any sort of piezo", you could to add a ~1M trimpot, in series with ~10k as a minimum, so you can easily calibrate this later)

If you are multiplexing a single ADC and find the signal is cross-bleeding between the channels.

This is probably not because of flaws in muxing, or sampling too fast. It's probably because ADCs like the Arduino's, and its muxing, are designed to sense things with lowish output impedance (<10kOhm for AVRs, see the "Analog Input Circuitry" section in the datasheet). If the impedance of what you're measuring is higher than that, then the interaction with the ADC's sample-and-hold capacitor starts to noticeably become part of the sampling.

One thing you could try is lower the discharge resistor. While lowering it to ~10kOhm lessens this crossbleed, it comes at the cost of having a weaker signal (making it less sensitive) as well as changing the RC circuit with the piezo to dampen faster (which makes velocity sensitity harder), and it only lessens and doesn't solve the ADC interaction issue.

Looks messy but is mostly just wires, and everything for eight channels:
- Arduino Nano (muxes 8 analog pins),
- two LM324s (quad op-amps) for eight buffers,
- discharge resistors,
- a serial MIDI socket (out of frame).

Those op amps are configured as voltage followers, because the large piezos I have happened to peak at a reasonable voltage as-is, but in general you want to control the gain e.g. with a non-inverting amplifier.

Yes, I should add diodes to this.
And yes, those are different-valued discharge resistors, I was experimenting.

A better fix is presenting a low impedance to the ADC, via a buffer (which also implies the discharge resistor now only affects the dropoff), such as with a FET or an op amp.

My personal preference is op amps, mostly because those will also let you control the gain (and it's slightly easier to breadboard, with quad op amp ICs).

That said, op amps come with further considerations, like that the output voltage swing is usually lower than the rail, so try to keep under that effective clipping level.


  • You can only multiplex so much before you can't sample fast enough to get well defined oscillations per drum channel, start aliasing, which would mean velocity starts being less robust.
You may want to run the ADC faster, as the time resolution is more important than the amplitude resolution or noise.

Protecting your input

Diodes you see in a bunch of circuits seem to typically be ~5V zeners from ground to the input socket's signal, to clip peaks and protecting the ADC/buffer from larger piezo voltages, which could be a few dozen volts if hard-coupled and hit hard enough.

How necessary this is depends on a little the choice of piezo and resistor, and specs of ADC/buffer,

In a generalized design it's a good idea that it can help lifetime, and costs cents. They should be in my circuit.

Detecting empty sockets


As the above electronic construction means self-dampening, the voltage output is amount of recent change.

That makes basic velocity-less code easy to write. Do analogRead, Is it above threshold, and have we not triggered this in the last 100ms or so? Then trigger MIDI note" goes a long way.

When you want velocity, and smartness to help isolation, then things become a little more interesting.

Physical isolation and sensitivity

If the physical part of isolation wasn't perfect, you'll still see a little of the force from neigbouring pads.

If you don't care about velocity, then the energy loss along the way may let you choose a threshold that is fine. A little threshold tweaking can help (and you can make it adaptive).

Still, your probably want you physical design to lessens this a bunch.

Particularly when you do care about velocity, because the better the physical isolation is, the lower the 'played at all?' threshold can be, and the better we support softer playing without false triggering, even while you're hitting something nearby harder.

Yes, you also have the option of 'if two pads seem to hit less than a few ms apart, and one has rather lower intensity, then then maybe just trigger the strongest' - but

  • that means you have to delay all hits a few ms.
  • this will sometimes ignore actual playing, and may punish good timing improving physical isolation is the better solution.


Velocity makes things more interesting.

There are various methods. Some are simple, some are complex, some are slightly faster (latencywise), some are more consistent, some are handier for softer playing, etc.

For example, we could watch channels for the initial hit. Once we decide it has been triggered, we keep sampling it as long as the values keep increasing, trigger once it starts falling, with the peak value as the velocity

This is probably the lowest-latency you can get, and it's simple code in that it doesn't need to store much. However, high frequency content makes this sometimes stop too early/low, and using just the first peak may be a little less robust between similar hits depending e.g. on the resonance of your design. If/when (just) that first peak clips, it will always report the same value

Another way is to, once we decide to trigger, keep sampling for another 2-15 milliseconds, and average the values (or e.g. over-threshold time), and use that as the velocity. This fairly robust to any clipping of the initial peak.

Because tiny peaks that were only just enough to trigger the "could this be a hit?" threshold, but the average would be very low, it's a little easier to reject it as probably-crosstalk or noise, while still allowing soft hits.

has a fundamental tradeoff between longer (more consistent velocity) and shorter sampling time (lower latency)
can be done without storing samples in RAM (keep adding magnitudes to a large-typed counter, and divide by sampling time)

Apparently there's some that try to find the envelope of the signal. Probably more accurate, but slower, and harder to do

Since you don't have direct control over what pad voltage should correspond to the softest hardest hits (depends on the size of the piezo and the size of the pad), you may want to calibrate.

A 'play a few sample hits on everything, and store that into EEPROM' is a simple and effective idea.
This also makes it more a lot more modular/portable around varying physical designs.

That also makes it easier to change the curve away from linear by applying a function (e.g. a power).

Fast playing (flams, drumrolls), hard playing, and false re-triggers

Playing harder means the signal is stronger, but also a slightly longer time it takes for the physical vibrations to taper off - may be up to 40ms or so.

So if you're reading fast enough, then you may see the tail end of one hit's peak, and when hit hard enough it might still be high enough to trigger again.

A conceptually simple fix is to add the condition "AND did I not trigger in the last 100ms?" in your note-triggering code.

A 100ms backoff on a channel is enough for basic playing, because that's 10beats/second, = 600bpm and most playing is far slower.

...except for flams and drumrolls, which are hits on the order of milliseconds apart, so if you care about these, you now do want to distinguish between a falloff and a new hit.

This is harder to get right. It's certainly possible, but involves some informed tradeoffs.

One decent approach is to

  • normally: leave a low threshold for triggers, because
they come up from nothing
a lower threshold can help velocity sentitivity
a lower threshold can help lower latency a little
  • for some time after a note trigger: temporarily set a higher threshold
based on the last triggered note's actual strength if possible
one issue is determining that new threshold reliably
you probably want it lower than the first peak
but higher than the average or you risk retriggers

This means a balance between

general threshold low for lower-latency triggers,
evade most dampened peaks from the same hit,
still catch most similar-velocity hits.

If you do velocities, then you may like the ability to tweak the curve on the fly.

You may want some parameters to bias this towards one behaviour or the other.

Consider also that some channels have their own logic

You're not going to do few-ms drumrolls on a kickdrum, so can reject this harder
fast playing on a crash cymbal amounts to a sustain so the speed doesn't matter as much - while on a ride drum or tom this is more important.

And it's generally useful to have distinct inputs be specific drumkit parts anyway, though, if only because then you don't have to think about configurable MIDI mapping.

If you vary trigger designs, e.g. have kickdrum just be a block of wood, consider physical design having a some way to tweak thresholds and curves per channel.

Hihat logic

If you care to have physical practice for real hihats, you want to include the pedal.

Since the pedal should allow staying closed, a piezo isn't enough.

A basic footswitch works well enough, though making a sturdy one that feels something like a real one is a project of its own, and you may prefer buying a cheap sustain pedal.

In theory you could combine it with a piezo to also get velocity of the hihat close, but that's probably overkill, and I'm not sure a lot of drumkits even use that(verify)

It should emit

open-hihat when hit while open, note number 46 in GM(name: A#1)
pedal-hihat when pedal closes, note number 44 in GM(name: G#1)
closed hihat when hit while closed, note number 42 in GM (name: F#1)

Those notes are from the General MIDI percussion set, and some drumkits are known to deviate somewhat. {{comment|(Remember the octave numbering is not well set within MIDI, so if set by note name, the octave may be +- 1(verify))

Two-sensor logic

Muting logic

Cymbals could be muted. I've seen kits with a strategically placed pushbutton, but I'd suggest something capacitive, even if that's more work.

Probably via a NoteOff.

I think I've seen aftertouch for this use, but whether that works seems to depend on whether your drum software listens for this at all.

MIDI practicalities

(see also MIDI_notes#Percussion)

Most drum related sound makers only care for the NoteOn, but DAWs that visualize notes will get confused if NoteOffs never come, so my code sends a NoteOff ~10ms later (unless the next trigger is sooner than that).

MIDI Velocity

MIDI velocity may need to be a little stronger than other things. That is, regular playing may be around 100, and compared to that, velocities under 30 or so are barely heard, if at all. Even ghost notes would probably be at 40-60.

Much lower and it'd sound like you're playing part of your kit from the next room.

(I spent a day trying to figure out why hits went missing when it was just that I'd spread them over the MIDI range, and velocity ~15 doesn't actually trigger a lot of drumkits)

As such, you should think about tradeoffs, between more expressiveness (sounds more natural), better repeatability, and and allow ghost notes.

It helps to have

good physical isolation (lets you do softer hits without triggering from neighbouring pads)
habit in playing
may help to put a transform that makes the mapping from physical to MIDI velocities nonlinear.

A little tweakability to that can't hurt.

PC side

Drum sounds

This article/section is a stub — probably a pile of half-sorted notes, is not well-checked so may have incorrect bits. (Feel free to ignore, or tell me)

If you want to get decent sounds for free, try options like

  • a decent free drum VST (like MT-PowerDrumkit), in a free VST host (like Cantabile) or a fuller DAW if you have one anyway.
that VST has natural drumkit sounds, velocity senitivity, and is free (though has a click-once-at-startup annoyance, that you can disappear by contribute something once)
tell cantabile to use the WASAPI output, 48kHz (or higher) if you can, and you can usually get away with 384-sample buffer on generic modern hardware, for ~6ms latency which is decent for drumming.

Linux has a steeper learning curve when it comes to MIDI routing, and low latency audio output. It's on my list of things to figure out and summarize.