Newbie after a few answers (that I couldn't find elsewhere)

[escapemcp]

I am thinking of building my own midi controller, and have come across the midibox. From reading all sorts of articles, it sounds like what I am after. I am trying to replicate the 4midiloop controller (52 Pots (incl. faders), 9 Encs, 152 Leds & 166 buttons), and £1400+ is just too much for me!:

I also like the idea of it being my own creation (the satisfaction I will get from it when built will be immense), and as such, I can alter & modify it as needs must (i.e. whenever they bring out a new feature in Traktor).

My questions are as follows (simple one word answers will do... I know you guys are busy building your own boxes!):

1) In order to get an idea of the architecture (h/w & s/w) of a midibox, I was thinking of building one (of the four) channels (with just fader, eq & a few buttons), then building a whole channel & one whole FX slot. Only once I have this fully operational (and more importantly, understood) would I then think about building the whole shebang. Do you think that this is a good strategy?

2) I am planning on using the PIC controller (for now) as midibox on a PIC is much more mature than on STM32. If I decide later on to swap the PIC for the STM32 (when the STM32 has been developed further), would this involve a major rewiring of the whole mixer (I am assuming not, but just want to check). I assume that the only bits that would need to be rewired are the DIN, DOUT & AIN modules to the core. Y/N??

3) Has anybody used push POTS (not encs) to replace the buttons on things like filter & EQ? I take it that I would just need to wire the analogue part to an AIN module, and then the push switch of the pot to a DIN module... but as I think that they share the same ground, would this cause any interference?

4) This is more for info at this stage. but in terms of parts, I *think* I need a core module, 2x AIN modules (giving me 64 pots/faders), but seeing as I want 128+ buttons & LED's, I was looking at running 2 cores (with seperate DIN & DOUT modules) & using an IIC Midi module to connect them. As I cannot program for toffee (although looking at Thorsten's files, a lot of it doesn't seem to be actual programming, but more like assigning and configuring the boards in c), will this be impossible to achieve with my level of programming knowledge (I work as a computer tech, so I have a very basic idea of programming, but cannot really write code myself)? Can you think of any other ways of supporting around 166 buttons, 9 encoders (so that's 184 DINs), and 152 LEDs? I like the idea of splitting DINs and DOUTS of the mixer into two (channels & fx 1&2 on one core & 3&4 on the other), so that it easier to follow what is going on, but is having 2 x 3/4 filled modules worse or better than having 1 full module and one half full one? Will the performance of the mixer be slowed down due the use of my half&half method? You can kind of ignore question 4, as I haven't read up on it *fully* yet (as it takes quite a while!), but any thoughts on how to support this would be appreciated. I will not be building this until I am confident that I can build and understand a much simpler setup (like my one channel mixer I mentioned earlier), but I am just trying to get an idea of what is involved (i.e. a LOT of soldering & loose wires! :)), but what else should I be considering (if it is something nightmarish, then it may kill this project before I even start, which is why I am trying to get an idea now).

Many, many thanks in advance for any ideas or help that you may have to offer. Sorry for asking Q4 before I have properly read up about it, but I am just trying to get a basic idea on how you would go about it.

Regards,

Aidan

[nILS]

1) Yes

2) You'll only need to rewire the modules<->core connections, not the entire thing, yes.

3) It shouldn't as long as your grounding is decent :) Usually the switches in the pots are not electrically connected to the taper. Which means you can still separate analog/digital ground.

4) There's different approaches. You could use multiple cores, or wire the inputs/outputs in a matrix, or...

No worries about not fully having understood *everything* before asking, that's next to impossible. It's quite obvious you've done your homework :flowers:

One thing to keep in mind though - DIY is in no way cheaper than buying the ready-made box. It's definitely more satisfying, more personal, more customizeable and fun, but in no way cheaper ;)

[escapemcp]

Many thanks for your (very) prompt reply.

In response to 4) it sounds like there are several ways of doing it, which is reassuring. If I used 2 cores, and connected them through the midi bus (via a midi lead), would this be slower (in terms of latency) than using the I2C bus? (Although I would probably move to the I2C midi board if I can manage it, as it takes care of some of the midi processing on the I2C board, which sounds like a good idea if I do have as many buttons, leds etc that I am dreaming of!)

I have also noticed that using 2 cores gives me 2 LCDs (which I would use, seeing as they seem to be invaluable for debugging). That's very useful, and just what I always wanted :) (1 screen for 2 channels allows me to display the info that I am after... I have yet to check out the LCDs fully, but I think I can manage it :P)

Thanks again.

Aidan

PS. Oh, you mentioned the cost thing... do you really think that a project like this is going to cost £1400?! (I don't count my time, as it will be something that I will enjoy doing). I was thinking I would save *some* money, surely?. I also won't have to blow all the cash in one go, which makes it more manageable (e.g. buy the lectrics in one go then buy front cover (cut by Front Cover Express (I think that was their name)) at someother point in time etc)

P.P.S. Talking of front cover faceplate things & grounding, I was thinking of a wooden box with a metal faceplate (so that I can easily round the corners, and also cause finding a metal box the size I want will be difficult), but how likely is this to introduce interference (due to lack of a Faraday cage?). Could I line the wooden box with tinfoil & connect to earth, or is that just a fire waiting to happen due to the resistance of the tinfoil (as its so thin)? :D

If you don't to answer cause you have loads of other posts to answer to, then don't worry... I am writing this off the top of my head in response to your answers. I need to read some more (arrgh! more reading!!! :D) anyways, so, in fact, don't answer. I'll read on and then ask some more (informed/relevant) questions. Sorry for wasting space on the boards hard drive with this response :D

Edited February 23, 2011 by escapemcp

[nILS]

If you use multiple cores and simply merge the MIDI outs the added latency will be tiny enough to ignore. I haven't done much work with multiple cores, so I can't really give you an exact answer though.

Cost: There's lots of hidden cost in DIY. Typically the electronics are cheap, the UI isn't; Pots, encoders, buttons and (most obviously) the panel add quite a bit of cost. But, as you already noticed, it's not really as bad since you can spend manageable amounts of money at a time over a fairly long period.

I wouldn't worry about possible noise on the faders/pots until you actually run into problems. Tinfoil may help but shouldn't be necessary.

[lylehaze]

If multiple cores sounds attractive, there's an idea that might help:

If you pick up a GM5 board, even just the little one (not necessarily the 5X5 big board)

you could connect up to 5 individual cores to that, and still have just a single USB cable to link them all to the host computer.

(downside) is that the cores would not be talking to each other, just to the host computer.

Also, I would NOT try to power everything from one USB cable.. (might work, but I wouldn't bet reliability on it)

Still, it's not a bad place to start, maybe.

Have Fun,

LyleHaze

[stuartm]

I guess there is a lot of bi-directional things going on in the 4midiLoop, i.e. the software (i guess Traktor) sending messages back to the controller for lighting up LEDs and such.

This is not impossible with a Midibox, but do take into account that this might require some more or less programming work to get it running.

/edit: I re-read your post and saw your comment on the LCDs - note that those are basically for displaying "internal" information of the box. If you want it to show for example track names from Traktor, this again would need some clever modification of the firmware.

My personal advice: think about what YOU really need, or what would fit your workflow best.

A 1:1 replica is not impossible, but fairly complicated.

But I think you already did well in understanding the midibox concept - keep reading, researching, thinking and sketching things out, we're always open to comment on concepts.

Edited February 24, 2011 by stuartm

[escapemcp]

Many thanks for both of your answers. I am going to be ordering a PIC core, along with a DIN, AIN & a DOUT module in the next couple of days, and have a go at building a single channel (or more like a few pots & buttons initially). This should give me an idea of what exactly is involved in building a midibox (as when you imagine the procedure in your head, you miss out 3/4 of what is involved!). In terms of the PIC core, I noticed that there are 3 variants.. the basic, original one, the one with extra memory, and the one for code intensive projects. I don't *think* I need much code or memory, so should I buy the basic PIC18F452 for max compatibility?

Stu: As I use almost all features in Traktor, I am after a (close-to) 1:1 mapping, as with the 4midiloop. I need my AINs, DINs and some DOUTs, (don't care on which note or CC they are on as they can be mapped in Traktor... admittedly, I don't want to be putting a Pot on F#-1!). The only thing that I want the LCD display to, well, display, is the size of the loop and the size of the move function (both ranging from 1/16-16 beats) - one for each of the 4 decks (this could be done on a single LCD, although I am going to use 2 - one attached to each core, so that I can use it for debugging as well). I expect that I will have to get my hands dirty in the code somewhere to allow this, but that's all part of the fun! You said keep sketching things out... my 2 latest sketches are attached (although 001.jpg is already outdated as I got confused and read that the 64E could only take digital ins, but this is only it's default state and not the case after a bit of firmware twiddling).

Thanks for the midi merging ideas. As I said before, it's nice to see that there are several different ways of going about linking the cores. It's always good when there is more than one way to skin a cat (meeeow!).

I am going to try to build the mixer as modular as possible, in order that I can get one section working, and then move on. This way, I can work slowly and methodically around the mixer getting all bits working independently of the other. For example, I need 31 buttons for each channel (maybe 32 - but no more :)). I will then use a DIN unit for each channel, and hook all of these up to a MIDIO128 programmed core. The pots on the FX section are on their own AIN module also. (See bottom 2 diagrams on 002.jpg if you want to see what I plan to have attached to what). By breaking it down like this, the project (to me) looks & feels far more manageable, but I will find out when I get my parts through and I let rip with a soldering iron :D)

Again, many thanks for your ideas and suggestions.

Aidan

P.S. I keep coming across the thought .o0O(do I need a bankstick)? As I will only need to use the controller in one mode (as everything will be close to a 1:1 mapping) and don't need any of the controllers pots, encs, buttons or leds to change their mappings within the midibox, is a bankstick needed?

Edited February 24, 2011 by escapemcp

[nILS]

Haven't had a closer look at the schems, but picked up one question in them: Yes you can have encoders and buttons on the same DIN module. As you correctly noticed, encoders need 2 pins, buttons need 1 (hence 14 + 9 * 2 = 32). Encoders HAVE to start on an even pin (if you count from zero) ie, 0, 2, 4, 6, ...

Bankstick: Unless you want to store some kind of patches, you don't need one. Since a bankstick doesn't require anything but the IC itself, it's easy to add later on if you feel like it.

[escapemcp]

Yup, got all that. I think I've read the site dry (even for stuff that I am not using (in case I do need to use ;))

Oh, one last thing, are SmashTV's & Mike's PCBs identical?

[nILS]

Absolutely not. Mike sells home-etched single sided PCBs that need some wire-bridges, while Smash sells pro-made double-sided PCBs with silkscreen that don't need any bridges.

The circuits used are the same, so both will work, Smash's PCB are a lot easier to work with though.