Bank 1 doesn't record/play incoming sound

Hi, Finally built my WTPA 2, and Bank 0 works/sounds great. Bank 1 looks like it's working great, but no sound is recorded. I do hear the click at the end of the sample as it loops, but is silent otherwise.

Anybody else run into this issue?


  • not me, 

    btw I'm thinking of opening mine up and simply shielding the long audio in wire that is probably the source of a lot of noise

  • Hey!  Sorry to be slow to respond to this.

    BANK 1 not working is usually an issue with the second clock not working.  If you connect BANK 0 and BANK 1 clocks via the rear switch on the jack board, does it change the behavior?
    LMK what you've tried with this and again, sorry to be slow.

    Click at the end of the sample:  That's the ADC jumping from mid-rail to some voltage.  It's most noticeable with a silent sample.
  • No worries on the timing, it took me a couple years to get the time to build, you can have a few years on the support. :-)

    So the switch doesn't seem to affect the recording, but when it's tied to the Bank 0 clock, and I sweep the coarse pitch dial,
    that little click is audible as a buzzing sound that adjusts with the pitch dial.

    When it's not tied to Bank 0, it doesn't do anything.

    I checked the jack board and main board wiring links and they are solid (no shorts or missing), the switch positions correspond
    to the right wires on the main board.

    Any other ideas?

    Thanks a lot!
  • Hey man, I know WTPA is like the college girlfriend you wish would just stop drunk texting. If you can lend your wizardry one last time, I'd appreciate it.

    I left my WTPA2 sit for a while and only used Bank 0, but now I'm wanting to have 2 samples running and not sync them (and use my trusty WTPA1 for yet another sample).I B1 checked it again recently and dinked around and I still can't figure it out. 

    Switch on R5 (VCO): R5 fully counter clockwise, records a portion of the sound clean, but in the middle of the sample. Silence front and back and a hard click at the end of the sample. R5 controls playback speed (turn CW slows it down). The other buttons work fine (stop/overdub/etc), but always has a lag to start.

    Switch on Bank 0: Records no sound, but captures/plays a click at the end of the recording. R6 and 7 both adjust how quickly the sample plays (the click gets faster) but no sound.

    Thanks for any ideas on where to check.


  • One more thing:

    With the sample that does get recorded (which appears to be about 1.5 secs before and after of silence), the switch can be in B0 or VCO position, and control the speed of what was recorded.

    So it just seems the record is slow to start, and lags beyond or cuts off the sound at the end.

    I can try and trim that once it's sampled but that's trickier to do.

  • Man.  That's an odd one.
    If you are dropping bits of sample, you might have some RAM lines not working.  Check the soldering on the RAM chip and on the latches which drive the address lines, which if memory serves, are IC2 and IC3.
    Bank 1 uses a different section of RAM than Bank 0, so that's a possibility.  Not sure anyone has seen this before, though.
    Good luck and as usual sorry to be slow!
  • No worries on timing.

    So today I have:
    - checked, double and triple checked the soldering, no shorts, no bent pins, no missing or odd stuff on the RAM, latches, and CPU chip
    - checked the lines from latches to RAM (against the schem) for continuity and all looks right (no shorts, all seem accurate)
    - both latches have 5v
    - I didn't know how to check from the ATMEGA to the latches, however. I checked a few lines and they seemed OK, but others didn't. I wasn't sure how to read the schem lines from CPU to latches in some places.

    Nothing I did had any impact at all on the behavior of B1, and B0 still rocks as it did.

    Which IC would be suspect? Is it maybe a bad latch, or more likely a bad RAM chip? Is anything with the clock suspect, or does it seem to you like the ram write is getting botched somewhere (latch or ram chip) because some of it writes OK.

    I do have a Diligent Discovery 2 scope thing that has digital features, (I've yet to use them but maybe now is the chance). If you had a scope or probe on any of this where would you look first?

    Thanks for the help!
  • Also just FYI the continuity tests I did were from the top, so at the chip pin on latch to the chip pin on RAM (confirming the solder and trace are in tact).

  • Oy.  Yes, this is a good one.

    Best guess is still that something isn't going into RAM, although I don't know why that would be.  You aren't wrong, there's an off chance it could be the clock (see below)

    If you can use a scope, you can check and see if your data bits are toggling on the RAM chip during record (during the part of the sample that comes back silent).
    D0, pin 13 on IC5 is a good place to check.

    If data is toggling:
    See if the address bits are toggling too.  A0 is good to check.  This should be fine, since Bank 0 is working.
    You should also check A16,17,18 -- these probably only affect Bank 1 on short samples (not Bank 0).  They will not toggle as often as A0, but you should be able to watch them change if you wait long enough.

    Data not toggling mean:
    Busted connections on RAM or MCU, firmware messed up, or a fried latch somewhere on the board pulling those lines down (could be a lot).

    Address low lines (A0-15) not toggling:
    Check the other side of the data line going into the latch.  IE, for A0, check pin 3 (1D) on IC2.
    (again, this should work if Bank0 works)
    If you see data into the latch but not out, the latch is probably at fault.

    Address high lines (A16-17) not toggling (like, ever -- they will go slow) mean something screwy with MCU, firmware, or the RAM IC.

    Check /WE and /OE on the RAM also.  I believe they should change when you start and stop sampling.

    If all this is OK:
    Make sure you see audio going into your ADC input on the MCU (pin 40).

    The second bank (Bank 1 or Bank B) triggers slightly differently than Bank0.
    This seems unlikely but it is possible that your pulse shaper is messed up and you are double triggering, or not triggering sometimes.  Make sure the signal out of "BCLK" looks believable.
    The signal at the MCU should be a short, high going pulse with either clock source selected for Bank B.

    If it's a square wave, or mostly high, or the amplitude is low, or there's no signal -- then this is a good place to start.  Check if D3 is backwards or C28 is the wrong value (like it's an 0.1uF cap instead of an 820pF).

    Good luck and let me know how it goes!
    Thanks for sticking with this,
Sign In or Register to comment.