We've recently constructed two identical test rigs around USB-6229 BNC units, using most of the available analog I/O channels on them, and a few digital I/O channels. During the commissioning process on the first unit, one of the analog input channels (AI20, specifically) produced very inaccurate data. All other channels worked properly, and we had a couple of spares, so we sort of assumed we'd somehow damaged that channel with an overvoltage signal and just avoided it.
However, during commissioning of the second unity we found one of its AO channels (AO1) was stuck outputting -10V (-10.21V, more precisely), no matter what we did--power cycling the device, leaving it off for a while, running reset and hardware tests in MAX, setting a specific value directly to that output via MAX, writing to it via LabVIEW I/O vis. The other three AO channels are fine, and again we have an extra, but this time we are quite certain that that channel was never shorted, supplied with a voltage as an input, or otherwise abused.
This experience has soured us on the otherwise quite nice USB BNC multifunction units, so I'm wondering if this is behavior anyone else has seen and/or if there are any troubleshooting steps we should be trying to revive these apparently dead channels.
To note, the systems in question are only reading relatively low-voltage signals (nothing over 8V), only outputting voltage to appropriate high-impedance devices, and are supplied by a high-quality, regulated 24V power supply that checks good. Temperature and humidity ranges have always been acceptable, and anti-static measures have been observed, though static is usually a non-issue in this area due to the climate.
