Request: Advice on Building a MIDI Controller

Greetings all.
I’m pondering building, or rather ‘have built for me’ a MIDI Controller for my rig. I’ve been using a Prophet 12 (Local: Off) for ‘buttons, modwheels, sliders, knobs’. I’m wondering if those of you who have designed/built controllers have any suggestions in the following areas: (Note: been ‘designing’ the controller via the Yaeltex website)

  • Arcade buttons vs Those silicon square ones: For the button controls I’d be using the most, I was pondering using arcade buttons because they can take more of a beating. However I’m balancing this with the visual feedback from an LED under a silicon button and not completely sold. Have you used Arcade buttons and then in hindsight thought the silicon ones would have been better? Or the other way around?

  • Fader vs Pots: One specific section of my controller is for some system meta-controls. I’ve been using the Prophet 12 touch sliders, the modwheel, and a foot controller. I’m wondering if faders might be too delicate. (Note: in previous rig incarnations I’ve utilised Akai APC controllers and those faders seem to go out on me constantly, so I’m superstitious (Mostly about Akai, but don’t want it to spread to ‘all faders’. LOL)) Have you used faders and in hindsight thought you should use pots or vice versa?

  • Ergonomics: How TF does someone imagine ergonomics?!? Okay, so buttons/knobs that you want to access together or quickly should be close, or available for different hands. Other than printing/drawing out a diagram of the controller and play-acting performing a song, are there any tricks of the trade or things you’ve done you feel were correct or others you might have done differently? I have an idea to put 16 buttons in not a 4x4 grid - the outer 2 vertical rows would be slid down by 1 so the curve of the hand is sort of represented - and that works fine with visualising the tasks/organisation for each button. I’m unsure this is warranted, but it seems kinda fun.

  • Empty Space: My controller is actually for a system that controls my gear. That is, the MIDI Controller controls a Bome Script that controls sequencers, synths, samplers, FX, DMX lighting, etc. I think a decent analogy is: when you play desktop synths and control each synth from their front panel, that’s like a manual transmission. I have an automatic transmission. That being said, each button or knob has a specific category of tasks. If I have extra space on the controller, and possibly extra $$$: would you recommend/not recommend throwing ‘a few extra buttons’ or ’ a few extra pots’ just in case?

…or if you have any random suggestions for a short busser, I’d love to hear the voices of experience.


1 Like

This is where I’m at currently.
Note: My ‘buttons’ represent logical groupings of Tracks rather than individual Tracks, which is all handled by a Preset within an Event Processor script. So, a button is [Kicks] irrespective of whether that’s a single Track, or multiple Tracks (sounds, samples, synths, fx, etc) to apply to what I categorise as “Kicks” in any given song.

The sort of ‘jogged’ vertical rows are because my hands are kind of screwed up (extensive nerve damage, etc), which is why I play sequencer rather than guitar or keyboard (like I did for decades).

Note: this is from the Yaeltex site - love their system and I’ve been waiting for someone to offer something like this for years!

And yes: Control Freak indeed


So looking at your layouts i’m guessing you’re expecting a two hands simultaneous sort of usage.

I don’t know the nature of the issues with your hands, but i assume designs that minimize that sort of injury are important to you. That is a very specialized area in ergonomic design, and it sounds a very important requirement of your design. Design follows requirements, so make sure you have a very clear vision of your requirements.

Have you considered two smaller controllers one for each hand ? That would cost more but would allow you to place your two hands more naturally, wrists in a more natural position.

In concept i really like the way you ‘jogged’ the buttons.

Your idea of using mock-ups for experiments is a very good one. You even could do things with foam-core, and $20 worth of parts like arcade switches. That might be useful. If you can avoid iterating on yaeltex systems that would cut costs.

There are arcade buttons with LED feedback, just not from yaeltex. The advantage to the LED is the feedback to know state (ON/OFF) of a binary switch. The LED is less important if you are using the arcade button percussively as a momentary switch.

Another difference sliders to pots, besides the positional feedback, is that pots generally require thumb and index finger to manipulate, sliders can be moved with one finger.

How rugged various components (buttons, sliders, etc) are will depend on the specific components that yaeltex uses.

I really like that yaeltex have LED rings for feedback for their encoders, as long as you have ways to get this data, for instance when you change a patch. Alternate to this is sending out initialization values and set the LED ring to that.

Just a few ideas. I like what you are doing.


They’re not difficult to build. The problem is you’ll never stop building. It gets addictive.
I’ve used the yaelteks. They’re “a little” overpriced…

1 Like

never encontered delicate faders so far.
also, never missed them.
when it comes to MIDI controllers, i want to have some pots + some encoders + some pads/buttons, but not faders. in particular, because they take more space doing the same job.

the only faders i really like are the ones on MC-707. they are super convenient for what they’re designed. but again, controlling track volume in live circumstances has nothing to do with fine tuning.

1 Like

Regarding arcade type buttons. Bear in mind (if you haven’t already) the physical size of the button behind the front panel. Some I’ve seen are pretty deep plus you’ll need to get a pcb behind too. Either way big switches may dictate the overall depth of the enclosure.

Likewise unless you’re getting the panel custom laser cut or cnc machined you (or the builder) will have a much easier time putt8ng round holes on the panel rather than square/rectangular. Same goes for slots for faders. You can get swaging tools in a variety of sizes and shapes but they’re not cheap.

As for overall placement and ergonomics you can’t get better than a pad, pencil and eraser. Or maybe front panel express if you prefer to go the software route. When I’m messing I start on paper then once I’m happy with a layout I’ll move to FPE for the final design.


image Front Panel Express
Good suggestion Bunker if someone is going more the DIY route. While yaeltex has definitely got the most options with a turnkey system (like they do joysticks), and their software configuration software gives you lots of flexibility (like it has NRPN built in), nothing can beat the DIY approach for doing the most flexible design, if you can and are willing to do the work.

Yaeltex is no doubt laser cutting their front panels, for their one off designs. And they do provide the printed graphics too. The back of the enclosure they do in wood, which is a very friendly option for DIYers. FPE can do metal backs for the enclosure.

There’s a yaeltex thread at Elektronauts. BTW: I don’t think yaeltex is expensive at all, considering the work involved in quality custom one-off production. You don’t compare price of a one-off product to those being mass produced.

Adafruit is one good place to pick up electronic hardware as a part of a DIY MIDI controller project. Here for instance is a link to their backlit LED arcade buttons:

You can get depth dimensions there. They also have ready made backlit matrix buttons — see Adafruit Neotrellis, and a complete line of other stuff at Adafruit, to do a DIY MIDI controller.


Thank you for the input.
I have to be careful not to blather on, so apologies if (read: when) I do.

Yes and no - the right and left hands do different stuff: Right hand buttons are for Drum and Bass/Rhythm Group and Layer collections of Events, the Left hand are 4xMelodies+Alts in addition to muting option. I’ve been working with a similar layout on a keyboard-as-controller - my Right hand is dominant and Drums/Bass vary more often in these arrangements.

Center stuff tends towards global: Faders are meta-controls, pots are for general purpose (example: I run a MIDI-to-DMX conversion in my rig and my current controllers have pots designated for overal lighting levels for washes)

Yes, I would prefer the LED rings like MIDIFighters have, which is why I’m hemming and hawing between arcade buttons and silicon ones with LED feedback. However, the row of double silicon buttons just above the arcade buttons are specifically for feedback of the play status - and the arcade buttons do have tiny LEDs associated with them.

Also my current system uses a Prophet 12 keyboard for what most of the arcade buttons will do, but I have a generic button box with LED feedback that I use for the visualisation.

This is a very important thing for me to know/remember. Thank you!

The faders are meta-controls for my sequences to control across many Tracks modulating Event data as it’s routed through an Event Processor. I’ve been using a Prophet 12 as controller and I’ve been using the 2 continuous controller pads, mod wheel, and a foot pedal for these 4. In a perfect would I would be able to include those types of controls, and I have the option for including foot pedals and other controllers at a later date.

My biggest worry is that (for example) my “Complexity” meta-control includes being able to Turn On/Off Events via the status of that fader and I might get overzealous. (Picture Ableton in Legato Mode with 3 clips of a bass line in different levels of complexity and bouncing between them - but MIDI - and yeah: I have a system for avoiding stuck notes (it already works, that is))

But I do find I’m adjusting at least 2 of the meta-controls at a time when I ‘get into it’, so thank you again: faders are more likely my best bet until I can do the kind of controls like on the Prophet.

Yes, I’ve considered that - butI’m also torn between flexibility (that is, designing a system that follows the standards I’ve already defined in addition to being open to future varations of those standards) and a system that’s completely bespoke to this approach. Also, some of the upper controls are more ‘global’ and I fear that would push me towards THREE controls (Right, Left, & Global) because: autism. LOL

Yeah. I have that data available.
I run the controller into a BomeBox, which controls a Pyramid, and that data goes back through the BomeBox for additional processing and routing. So, all data is basically BomeBox’d and manageable.

Thanks again for the feedback. I really appreciate it.
I dont’ really know what I’m doing, but somehow I’m finding ways to do it. LOL

1 Like

Perhaps, but I really suck at hardware.
And honestly: I spent 2 years working on the scripting and I’m getting…impatient. I would take way too long to do it myself - and it would still look little better than cardboard with duct tape. LOL

I thought Event Processor scripting was addictive - hey, dream it up, then write the code…then debug and try again!

1 Like

Good points, thank you.
My faders aren’t for volume - they’re for 4 ‘meta-controls’ that affect the processing/routing of all the MIDI Events.

Good points, thank you!
I"m using a website for design from a company (Yaeltex) which builds them. I think the reason they can’t use LED rings on the arcade buttons is because their controllers are so thin.

Also, excellent insight into designing panels - I wanted to respond because that means I’d be more likely to remember if I ended up finding a local builder to build one to my specs.

1 Like

Tangent: The funny bit about building my own controller from scratch is that then I’d want to incorporate the BomeBox tasks (the processing of controller and MIDI Events from the sequencer) and a sequencer into an rPi or something within the controller…

…and then the project gets overwhelming.

To be honest I’d not heard of Yaeltex until this thread. Every day is a school day!

Regarding FPE, you can download their software and use it as a design/layout aid. A bit like using sketch up or similar. I’ve never used it and then had them make me something but it’s cool that they give the software away. However I guess you’re already using something similar from Yaeltex?

I’ve used FPE for laying stuff out for years however I didn’t use it for a good while. When I opened it I saw there was an OS update. It’s cool; you can mage control scales etc now. Pretty cool IMO

1 Like

This is probably the best place to learn about gear that I’ve found so far.

I remember bookmarking them years ago when I had a different focus on my workflow, but they weren’t really ‘set up’ yet. I like what they’ve become and they seem like a great group of people - passionate, positive, creative, and brilliant.

Yaeltex has a ‘factory’ on their page to drag and drop elements. Fewer options than the FPE software, I believe - I actually d/l’d that years ago too. :slight_smile: (again, I suck at hardware - or at least that’s what the piles of soldered and partially built projects in the basement tell me)

1 Like

I’m not bad at electronics and hardware (been dabbling since my teens and I work as an electronic service engineer so I should know one end of an iron from the other by now!) it’s the software side I really struggle with. I started to teach myself basic assembly coding for PIC chips years ago. It’s been in fits and starts tho.

1 Like

I’ve been considering this too, (my post in the Hydrasynth thread), but on a lot more limited scale.

The code on the yaeltex is open-source, i believe, so i was open to the idea of moding that code to enhance functionality, perhaps a simple sequencer or something. I understand your hesitancy, but i’ve got a life long background in software and hardware development. Scripting the Bomebox though would drive me nuts. You’re a lot further along with this.

1 Like

Well, that’s the thing - I’d need slightly better than a ‘simple sequencer’.
Advantage is that my concept is a MIDI player, rather than a recorder.
But still: I have no idea how to get from here to there, and that probably keeps humanity safe.

Regarding the BomeBox: I just flow charted and pseudo-coded everything - the rest is just syntax. My first BomeBox was a pre-production model though and I’d already been using MTPro, but I’m not an expert - just a hack.

I think my approach is innovative, but it’s more likely I’m ignorant and I’m not terribly smart except I can write MTPro scripts to cover my mistakes. LOL

If I win the lottery, can I hire you to write my sequencer if I flow chart it? LOL

1 Like

I built my first Heathkit in 3rd grade and my ‘first synth’ was from PAiA!

But then I got into software.
Then I got old.

Picking up a soldering iron again is fine for little things, but without a ‘kit’, I wouldn’t trust myself - or rather, it sounds like something I’d be better off hiring someone to do


I know this is like a child asking an adult if they need help paying the mortgage, but if you want some help with an MTPro script for a mockup, I’m pretty good with getting around in it.

Thanks, I’ve been mucking about with a challenging FPGA project. That’s where i could use the help, and where i should focus my time. :grin:

I want to hear more about this project of yours as it develops, i like your approach to things so far. And you’ve shown a determination that stands out.