First Prototype Hardware Design Done!

All right, ladies.
As promised, I finished the first WTPA2 hardware design by the end of June. This does not mean it is in my hand, but it does mean the gerbers are done!
Check it out:

Through-hole SRAM, 8 Buttons, extra MCU IOs, Flash Memory Expansion Port, Rotary Encoder, New VCO Based Clock, Updated Ground Plane and Signal Routing, Updated Analog Section.
And the return of the prototyping area! Make yr own damn analog filter :-)

New dimensions are 7.75" x 5.00".
It's passed ERC and DRCs on my machine and it just got the clean bill of health from Advanced Circuits DFM! I'm going to sleep on it and order it tomorrow after another look with fresh eyes to make sure I didn't do something totally stupid.

After that I'll make the BOM, order parts for the proto, and update the firmware to handle all the new goodies.

I should have a new sampler by the end of next week!
Xo and Onward,


  • True to your word ;)

    I better stoke up my old faithful soldering iron huh..
  • Droolicious!
  • Excited about WTPA2...but even more excited now that I've really incorporated WTPA1 into my workflow.
    It's such a creative device for how simple it is.
    I start all my tracks by putting sounds through it, then mangling it, etc.

    Theme music to this post: "I Can't Wait", by Nu Shooz.
  • yeessssss!!! wooohooo!!!
  • Friggin awesome!
  • Muwahhaha! Let chaos rule! Nighty night Todd
  • ...and i'll be there! Great news Todd!
  • ...And here's the daughterboard proto. 64mBit Flash memory, a 3.3v regulator, and a fancy level translator. Associated bypass caps.

    That order goes out today also.
  • this kicks all kinds of ass :)
  • Lookin forward to it all!
  • mmm, storage.

    So are we talking patches here or what?
  • Nope, sample storage. No reason you couldn't store patches, I guess.
  • So sample storage with edit points? Something like that would be real useful..
  • Meaning, you can store the entire sample, and also the sample start, end and window data as you'd entered it?
    Such that when you loaded the sample, it would play in the edited configuration?
  • edited July 2010
    @ todd, yes exactly. I dont think the functionality of the flash has been discussed but I just assumed it was a place to store a sample(s?). Having several slots to permanently store samples with the start/stop edited points already saved would be really nice.


    I guess that would also entail a sample "format" which may be opening up a can of worms that u dont want to open..
  • @altitude -- Exactly. I'm thinking another way to get what you want would be to allow you to "permanently trim" samples to their window. Then you can save the trimmed one in one slot and the un-trimmed one to another and load whichever one you want as audio data.

    Would that work for you?
  • ^^ Me likey. That's how we used to do it, back in the day.
  • edited July 2010
    @ todd: I dont even see a point to saving the untrimmed sample if you can actually edit the file. A simple button combo with a confirmation step would be fine (i.e. hit combo once to save the edit, some flash pattern comes up indicating it is armed to write, then hit again to confirm). What are the plans for the FLASH? Does it only hold 2 samples for the two banks or will it hold more?
  • Dope. Sounds good. That's how it'll be.
    If you want the original file you can save it to flash. Then, while it's still in RAM, you can edit it and save it to another spot in flash. Or you can just record it, edit it and save it.
    That's what I was thinking, anyway.

    So here's how I was thinking about FLASH:

    *The flash is 8MB. That's 16x the memory space as the SRAM. I am not keen about fucking with defragmentation or a memory map on this, so that means you can store 16 "things". Or maybe better to say there are 16 "slots".

    *A slot can either store a sample from BANK A, a sample from bank B, or both samples. If there is one sample in a slot already, WTPA will look to see if there is room for the sample you are trying to save, and if there is, it will save that sample. Otherwise it won't let you unless you overwrite the other sample.

    *WTPA will remember what's in a given slot. If it's a sample, when you load it, you can pick the bank. If it is two samples, you load them into bank A and B respectively.

    *There are a couple things about this worth considering. Although with this scheme the max number of samples you can store is 32, you can only guarantee 16 in flash, since each sample CAN be the entire SRAM.

    *The other thing to consider is how to display this in a way that doesn't suck on the LEDs. I'm thinking 4 LEDs will tell you sample slot, and the other four will tell you what's stored there: nothing, the first sample, the second sample, or both samples. That's two LEDs for that. The leftover ones will handle telling you when there isn't room or asking you whether you really want to overwrite, etc.

    This is a little goofy but does make sense, and guarantees that there is maximal storage with minimal memory management.
    A simpler way would just let you save 16 individual samples, and that's it. That's theoretically half the samples of the other method, but thinking about it is easier.

    What do you think?
  • You guys ARE watching this come together, right?
  • That's fucking awesome.

    I always forget to check the blog, I just jump straight to here. I'm more forumer than blog reader...
  • edited July 2010

    Dope, everything I had hoped for. Going back to my original question/suggestion: being able to load said slot via midi program change (simple bank msg cc0)


    who does your protos?
  • edited July 2010

    Edit: Profound sadness...
  • LED 4 + 4 comment, whirds!? sparkfun lcd? :- )
  • And shouldn't show sample geeks toys they can't buy instantly, v cruel Todd. sniffle whine, when when! oh and, yahoo too.
  • @sealion: Strong objection to the sparkfun serial LCD. In retrospect, this is the biggest mistake I did on the Shruti-1. There's definitely something wrong with the firmware of this thing - settings occasionally reset themselves to strange values, which causes at best gibberish on screen, at worst a "bricked" display requiring a set of specific commands to be sent to it to revert it back to its factory settings. The worst part is that some settings can only be accessed by "toggle" commands, so if you don't know in which state the display is in, you don't know which command to send - so it's not even possible to reset the thing to a predefined state at startup. Even if I bought 50+ of those for my kits, Sparkfun hasn't provided me with any support ("fix the firmware yourself, we provide (outdated) code"), and, worst of all, this part is only available from one supplier at an unreasonable price, with no decent discount for bulk orders. In batches of 100, a LCD module is in the $3 - $5 range.

    I'm not sure Todd wants a LCD on the WTPA2 (I think this has been discussed elsewhere...), but if he does, using the same shared 8-bit bus with an extra parallel latch + a standard HD44780-compatible thing would be a cheaper, less troublesome way to go.
  • Nicenicenice.

    Regarding the display. IF there is ever anything like that I would go retro all the way:
  • edited July 2010
    Yep, all flash commands will be doable by MIDI.

    Olivier is right on all counts with the LCDs. Sparkfun stuff is weird; I've used their LCDs which are no exception.
    And ain't no LCD happening, so stop yer whining.

    The board above looks cool. What's it for?

    I get protos from Advanced Circuits, usually. For production (usually) I use PCBCart.

    Bricked is a state of mind. Well, it's actually a state of fuse bits or latch lines. But it's reversible. But not today.
    I'm running away to the woods with a bunch of people and my longbow and getting all extra analog this weekend.
Sign In or Register to comment.