I have several joysticks on a board running though op amps to scale the voltage to safe levels for Bela. When I check the voltages at the pins going into the multiplexer, there is no crosstalk, but there is crosstalk in the console after the inputs get sampled. The unused pins are grounded. It seems like the multiplexer could be sampling too quickly.

The multiplexer is set to 64 channels. Is there a way to slow the the sample rate down beyond the 2.75kHz? Or is there another issue that might be causing the crosstalk to only show up after getting sampled? Thanks.

Can you share details of your wiring? How are the joysticks connected to the board, what are the resistor values in use? The multiplexer switches well ahead of sampling, at least 44us if I get it right. I think channels 0 and 4 are the ones at highest risk of crosstalk (as they are the first to be sampled after a multiplexer switch), and channels 3 and 7 the most immune as they are the latest. See if you can observe any difference between those.

    sorry for the delay. The joysticks are on their own PCB with X and Y pots receiving -12V through a 1k resistor and 12V through a 1k resistor. The resulting voltage on that board is -10V to 10V. The joystick signals are sent to another board where each X and Y signal go through their own op amp to attenuate and shift the voltage between 0V and 4V. That signal is then sent to the multiplexer. I've attached two schematic screenshots. One shows a single joystick, the other shows the op amp schematic for one of the X signal. Each of the other X and Y op amp configurations are the same as the one in this screenshot.
    alt text

    alt text


    Are the various 1k resistors on the joystick board one per each of the A1, A3, B1, B3 lines of each joystick? If they were shared across the joysticks, that would explain some crosstalk.

    Is the multiplexer capelet's input connected after R163? Can you try bypassing that resistor and see if the crosstalk is reduced?

    giuliomoro I think channels 0 and 4 are the ones at highest risk of crosstalk (as they are the first to be sampled after a multiplexer switch), and channels 3 and 7 the most immune as they are the latest. See if you can observe any difference between those.

    did you try this?