WTPA3 Advice

So I began work on WTPA3 last week. It will be a eurorack sampler, sold complete. I am in the middle of schematic capture right now and have some sketches of how I want it to work and look.

Thing is, I don't have Eurorack gear so it's possible I could commit some sort of faux pas in the way I'm thinking about it.

Right now I know that WTPA3 will have more memory than WTPA2 and separate VCOs per bank, probably calibrated to 1v/octave. It won't have a menu/button system, but rather a handful of knobs and cv inputs which control effects and windows. It will not natively have MIDI although it may have an add on which enables it (since MIDI code is already in place In WTPA2). And I'd like it to look "Miami Vice", kind of like this:


If any of you have suggestions I would love to hear them. I have one specific question for now:

How wide is too wide for a eurorack module? It seems like people want skinny modules but that means restricting the number of controls and jacks. If it were up to me I'd err on the side of "controllable" rather than small.

I will let you know the tentative release date in a month or so.

Thanks in advance,


  • edited February 2015
    great news. it is a bummer for me that is is not gonna be diy though. (i only have one non-diy module in my system) i hope you make it open source.

    i'll try to post some stuff that might be helpful to avoid some faupaxes, i think you might have to get some eurorack gear though tbh.
    probably you know a lot of this stuff already. i also hope i get corrected if something is false.

    eurorack mechanical details: http://www.doepfer.de/a100_man/a100m_e.htm
    eurorack technical details: http://www.doepfer.de/a100_man/a100t_e.htm (stuff about the power headers etc.)

    there are a lot of things that are not really standardized in euroland, but a kind of "good practice" way of doing it has been established.

    i think the 10pin power headers without 5v rail are used more often than the 16pin ones. probably because some people do not have a 5v rail in their system. digital module then make 5v with their own regulator.

    for some crazy reason the "standart" for the power header is to NOT have boxed headers. (so that it is easy to put the power cable the wrong way around and fry some ics) a lot of diy pcbs are layed out so that boxed headers can be used to avoid that.
    i think it is also good practice to put diodes to protect the module.
    also it is the modules responsibility to filter the power input. (a lot of times ferrit beads are used)
    in a lot of modules there are also fuses or small resistors that are supposed to act as fuses and burn out before anything else in case too much current is pulled. you probably have already looked at the mutable instruments schematics to get an idea about how the power input could be done..

    i love the idea of a miami vice look!!
    be aware that some people do not like a very crazy look on their modules though.
    http://grayscale.info/ actually sells replacement panels for people that want to replace their "mad professor"- / "pink madness"- etc. style panels with "boring but sane & clear" stye panels.



    about the width:
    of course people want as many features as possible within a minimum width.
    this is sometimes kind of contradictory. if you ask users they probably want "every feature possible with a pot and cv in (that does audio-rate and has an attenuverter) for every parameter". i think the hard thing is to decide what is really usefull and makes sense and sometimes this has to be determined by building a prototyle and testing it in a patch rather than by thinking about it. i think this might be where you might want to build/buy/borrow a eurorack system.

    there is not really a maximum width. people also like huge modules if they have a crazy amount of features:

    people often look for a small depth in modules aswell, because they want to build them into shallow "skiff" cases like this:
    thats why in a lot of modules the pcbs are parallel with the panel - which also makes assembly easier because there is no wiring. if only one pcb is used and there are a lot of pots, jacks and knobs the layout is often quite dense which makes troubleshouting harder for diy modules sometimes.

    i think some effort should probably be put into deciding which frontpanel components like pots,buttons and jacks to use. i imagine a lot of manufacturers order a lot of sample parts before releasing their first module. i think quality jacks are a must. (something like PJ301BM)

  • @loderblast-
    Thanks! This sounds like good advice all around. I hit up Olivier about advice too and as usual he ruled with part suggestions.

    I'm shooting for 24hp width right now. Jacks, hardware and power considerations I've got a decent handle on for now, it's more aesthetic stuff I'm trying to sound out.

    24hp too big?
    0-5v cvs?
    input / output impedance?
    1v / octave?
    attenuverters: why?
    gate thresholds ok at transistor VBE?

    That kind of thing.
  • - I think the trend seems to be swinging back away from "let's see how narrow and feature-crammed we can make this module", towards more feature-rich and/or ergonomic (i.e. knobs with a little bit of finger space between them) modules. Probably due to the fact that there are more and cheaper case/power supply options out there now.

    - 0 to 5v CV's are fairly standard, although being able to respond to negative voltages (i.e. -5 to +5v) will give more modulation options.

    - might be good to have two different inputs. One for modular-level signals (for sampling other osc's, etc. within your system), and one for sampling external stuff.

    - 1v/octave, obviously desirable for pitch CV's. But also useful for other parameters (i.e. filter cutoff) where you want modulation to track your playing. Can't immediately think what, aside pitch, that you'd want to track 1v/Oct. on a sampler, but I'm sure others probably have some fun suggestions.

    - attenuverters: give you more modulation options, via ability to invert CV. Most typically they'd be on a CV source (i.e. something you were using to modulate a parameter on WTPA), rather than on a sound-generating module like WTPA.

    - gate thresholds? Dunno, this one's over my head;)

    Best advice, sign up over at https://www.muffwiggler.com/forum/ and ask all your questions there. Pretty much your entire potential customer base is already there ready to focus-group it up for you, not to mention lots of friendly and tech-savvy people who can give you good feedback/advice on the nuts and bolts stuff.
  • I/O impedance on the euro stuff seems to be around 100k input and 10k output. Usually the higher input impedance is desired so folks can use unbuffered multiples easily.

    I'd definitely start a thread on MuffWiggler but that might be a rabbit hole you don't want to dive into.

    Bipolar cv is cool - think of a lot of LFOs/VCOs with bipolar waves that you might want for cv control. At the very least protection on the cv inputs is super important so if for some reason someone plugs in something that end up being -10v it doesn't blow.
  • edited February 2015
    Yay! My dream come true. Is this going to be for audio stuff or DC based stuff as well?

    Skiff friendly wont hurt it's appeal and 24HP is reasonable.  Have you considered half of a WTPA (i.e one bank in a smaller form factor?) to get the size down?  Sometimes it's better to trim things down and make it smaller than jam pack everything onto one module
  • @rosch --
    The DIY version came out first!  You have one on your desk right now, right?
    WTPA3 really won't be feasible for your average person to build without a hot air rework station or soldering oven and solder paste; they'd need tweezers and magnifiers, and heaven forbid they need to replace a part.

    The differences between WTPA2 and WTPA3 (the DIY and manufactured versions) are based mostly on available parts.  It's not really feasible to ask most kit-builders to solder fine pitch SMT parts (I already tried easy SMT parts on WTPA1) and once you're committed to SMT parts the options open up a lot.  MCUs get more pins so you don't need latches, MCUs get more advanced and have DMA and memory controllers, so you can use SDRAM easily.  SDRAM is always surface mount.  Mutlichannel ADCs become available so you can read pots instead of encoders.  The list goes on.

    I believe in helping people sweat to learn.  But I also believe in fully using all the skills _I've_ developed.  In this case, I tried to do both, with two releases.  That's about the best I can do.

    I think it will handle unipolar CVs as well as audio.  It will be a bit of a PITA but I think it will be worth it.  Do you?
    Also, the "multi-channel" part of the design is small right now.  I'll post a sketch of the control panel when I have it done.

    The input should handle everything you throw at it, it just might not be meaningful under 0v.  You can always attenuate a higher voltage CV, so that won't be a problem.
    Hardest part will be dealing with the coupling differences between CVs and audio.

    Thanks everybody!
  • edited February 2015
    Sure I understand. So WTPA3 is basically the same thing, no additional features?
    I have to come up with a panel solution for '2' soon. I've decided to do it Caveman style for hands operation Euro format (fatass not skinny), but with as many features as possible vor CV and GATE control, signal levels etc, probably without MIDI IN for starters (although soldered). If I make a controller from my heap of good ole MIDIbox PIC cores I'll just add another panel, not all in one (If I go for all in one, as I know myself, I would end up with none-at-all, backlog)
  • edited February 2015
    @todd cool.  I keep having visions of sampling a couple cycles of a LFO or a sequence of notes then mangling them but if that is in the realm of possibility is a bit above my pay grade if you know what I mean.  I'd be just as happy to have a euro formatted WTPA that I can do horrible things to oscilator samples.  It would be nice to have boards available (or board files that we could take to OSHpark) for some MSOP masochism though.  Anything but QFN or BGA is doable by a (steady) human hand and to be honest, I'll take a big board with 0603 size parts over a big through hole one any day..
  • @Altitude --
    I totally agree on the 0603 front.
    Well, maybe I can make boards available.  I def can't kit these out, but boards maybe.  There will probably be custom pots again, although you can probably fudge them (they'll be similar footprints to Alphas).

    Re: CVs --
    Fuck it, why not.  It will probably just mean a button on the front panel which toggles the input coupling.  Like a scope "coupling" switch.  Just one more thing on the panel, but admittedly cool.

    WTPA3 will have more features (and a lot more knobs) than WTPA2, but it may lose a few as well.  WTPA2 will be more easily hardware customizable.  WTPA3 will be more powerful and optimized for modular use.  They will be different animals to interact with, but the general idea of the things will be similar -- multiple banks of independent samples with a lot of un-subtle effects.
  • well, so if I'm lucky I may have reason to add 0603 to my standard parts here too then :D
    Would be awesome.
  • Just after finishing putting together the WTPA2, I jumped into my first modular building... I've just got a Music Thing Modular Radio Music, which I plan to have a folder full of WTPA2-mangled samples on for further mangling with the modular.

    A WTPA3 module would be... absolutely great. Can you get mirror-backed neon acrylic for panels, or something... Your case has ruined me for neon acrylic - I need to build more stuff in that.

    I'd love to DIY it, but totally get the SMT thing. Though I'd probably take a board as well, just to have a go...

    Layout wise, I agree with @loderbast - that sort of balance is ideal. I'd rather be able to use the thing, than have a bajillion knobs that I can't get my fingers in to turn!
  • There are also several euro modules that have optional "expander" modules that break out additional functions to CV jacks and such. Might be an option for WPTA3 if it starts to seem like there are too many control options to comfortably fit on one panel.
  • edited February 2015
    As of now, MIDI will be a breakout. SD card functions will be their own module, sort of like a drum sampler, with a dedicated interface to WTPA3 for saving and loading samples. But you'll be able to stream multiple samples from an SD card even without a WTPA3. And it will accept FAT formatted cards. I hate FAT, but the people have spoken. The people hate packers.
  • Did you just imply WTPA2 will be able to handle FAT formatted cards with samples dragged onto them?
  • Wtpa3 will :-)
  • Stuck with packers huh...
  • WTPA2 is too cool for that ;p
  • The wtpa3 sd card module will have its own MCU so it will be able to handle FAT addressing and buffering and etc.

    -I- don't have any trouble with the wtpa2 packers :-)
  • Oh oh oh separate module for FAT SD. Got it. I just got stuck in the wishful loop reading "you'll be able to stream multiple samples from an SD card even without a WTPA3. And it will accept FAT formatted cards." My bad. That was just the stress talkin', man.
  • Those statements are still true, I think.  So the plan is:

    WTPA3 and the as-yet unnamed SD card module.

    Standalone, WTPA3 does live sampling kind of like WTPA2.
    Standalone the SD card module can stream loaded samples from a FAT formatted card.
    Together, you can save/load live samples to/from the card.

    If you read "even without a WTPA3" as "with a WTPA2" then my bad.
    It is probably possible to shoehorn FAT into WTPA2, but probably without live streaming (only sample loading and unloading).

  • So you are saying you'll add FAT for WTPA2 for those of us that pony up for a second kit so we can have both streaming and sample load/unload :)

    I think I need to take a rest, i still to plug mine in, build my mudibox seq v4 (and find a case for it), my monome, and the modular stuff I just ordered (to get started in that world).  Too much stuff.  I need to lock myself in a room with nothing but my sound stuff and just play.
  • In reality, Batsly should just "port" his DPCM sample packer GUI to the regular WTPA2 AIFF SD card format. Or make it so his packer can do either... He's totally got tons of free time so I don't see why this can't happen... pfffft!
  • We gotta get all you DIY types into Linux. And the command line. You really want to fly, kid...
  • Another one bits the (modular) dust!  First Oliver (mutable-instruments), now Todd!  It was a foregone conclusion.  The lure of Eurorack is too strong to resist!  
    If I recall, Todd, I emailed you a few times about a WTPA-based drum-module with ROM sounds, but this is even better.

    The first, obvious piece of advice is to talk to Oliver about his experiences.  He's got this game down to a science.  

    Anyway, are you saying that WTPA3 would be modularized?  That is, something like: module #1 -- WTPA core (sampling, controls, audio inputs/outputs); module #2 -- (SD card and supporting I/O); module #3 -- (MIDI, misc.)?
    That would be really great.  
    Personally, I view a modular as an instrument with ephemeral qualities; that is, loading/saving isn't important to me.  I just want to turn it on, do my thing, and start fresh the next time I turn it on, so I'd probably only be interested in the core module.

    MIDI: I don't think MIDI is a necessity for this module.  It's assumed that most Euro-heads will have a MIDI-CV converter and since you're planning to make most parameters accessible via front-panel controls, MIDI seems superfluous.  

    Power concerns: what's the target for current-draw of the WTPA3?  Keeping it under 200mA seems like a good idea, considering that some Euro systems can have a capacity of as little as 1.0 A (+12v/-12v).

    UI considerations: LEDs are great, but buttons/pots with integrated LEDs (or receptacles for LEDs) are the new-new.  They look cool and are space-saving, but can be more expensive.

    Audio/CV sampling: nah.  I'd keep the WTPA3 a sleek and simple module that does one really awesome thing: record and mangle audio samples.  There's enough stuff out there that can generate weird CV signals.
    Plus, when you sample and play back sampled CV, don't you have to consider aliasing/stepping and other stuff?  Maybe that only comes into play when, say, a fast LFO is recorded and played back at a slow rate.

    Anyway, Todd, please do keep asking questions and posting ideas.  We got your back.
  • This game is heating up:

  • Whoa, look at that.  That "windowed playback" looks a whole lot like the one from WTPA1.

    Do people like that "fake pitch shift" thing where you roll short sections of the sample?  I guess I wouldn't have called that "granular" although it uses small sample sections.
    The code is in WTPA2 to do that, even.

    Guess I better get my ass in gear.  How much does that thing cost?

  • Dunno.  His stuff is usually expensive and hard to get though.. here is the thread at muffs: https://www.muffwiggler.com/forum/viewtopic.php?p=1820058
  • holy shit, they are talking 600 clams for that thing.
  • plan on $300 a unit with some slack for retail,m ask muffwiggler store and analog heaven and noisebug.net and thonk and schneifdder in germany about markups, then produce at 1/2 the competoitors is the business school advice.
    Mine is be happy with what you do and hold back one trick from the wtpa 2.0 for the hell of it.

    The big thing you have is memory card, but can you push record and go straight to the card memory and then play back is the mega question.  1st one to do that wins, or usb socket, not an sd card.

    all the atari work you've done will have no utility to the modular crowd, so loose the code for that in the wtpa 3.0  Buy knobs that match Oliver and it will piggyback nicely.

    the question is do you want to a NAMM style retailer or a loft-livin crazy free style longboard sound masher, you know.  iT s the classic builder's dilemma.  Oliver decided to be a real company business.  Props to him.
    his stuff is polished and doing well.

    modular people want to make it with the other units, tape record it and then f-it up, for them its an efx, for you its the whole damn ballgame?
    borrow someones rig that is way crowded, cable the crap out of it and see what your fingers think of spacing.
    there is a machine getting nice reviews made by nice people that I think is too crowded to work once cabled.

    me own 1 module (an aries oscillator) cause its an addiction man...

    or what would biggie do?
Sign In or Register to comment.