Bug: Acquisition server crash/freeze
Using Brain Products BrainAmp HW, we regularly get acquisition problems.
Here's a description of the issue:
-
we are using a 64 electrodes setup, and BrainAmp amplifier, plugged in USB to a laptop with Windows 10 & OpenViBE.
-
the ACQ server is launched, set to driver Brain Products BrainAmp Series. "Connect" & "Play" are pressed. First remark : the "drift" is around 750ms, which seems very high. This displayed metric should be linked to TCP tagging, but is this relevant?
-
the designer is launched, and an MI protocol using an ACQ client is run. Signals are correctly transmitted from the ACQ server.
-
as long as the scenario is run in the designer, everything is fine.
-
Randomly, when the scenario is stopped (with the ACQ server still in "play" mode), we can reach a situation where the ACQ server freezes. One can notice the "drift" value starting to increase erratically, from 750ms to 900, 1000... Sometimes getting back to 750, then going up again, towards 1400-2000. Around 2500, the ACQ server usually freezes. Obviously, no data can be received by the ACQ client once the ACQ server is frozen. The only solution is to kill the ACQ server, and to launch it again. At that point, the server is usually unstable again (with the drift going around, and eventually freezing), or can be stable.
-
if the "drift" starts moving erratically, one can "stabilize" the behavior by running a scenario (using the ACQ client) in the designer. Even a simple "ACQ client + plot" can help. Whenever such a scenario is run, the "drift" seems to stabilize, and no freezing/crashing has been observed.
-
this behavior was observed on 2 different Windows 10 machines, using the same BrainAmp drivers (as far as we know). One of the two machines has a bug occurrence far larger than the other, but the bug appears on both nonetheless.