Q:
If I try to run an acquisiton or try to change Medipix settings I receive couple of error messages like:
Transmission failed: Command "Set matrix" failed. (Can't transmit - IO error)
General error: Unhandled background message, Timeout error: Timeout.
No received response to 'M' [0x4d] = "ADC monitor"
A:
The reason of such behavior is most probably bad (low quality) USB cable. The device is taking about 400mA from USB (USB can supply up to 500mA). But if you are using bad (cheap) cable with thin wires the voltage at its end can drop significantly at such current. Power consumption can vary according to operation performed, so drops can occur even after successful initialization of the device. This causes device reset. The most power demanding operation is the equalization - with certain threshold the most of pixels fall to noise and start to count very fast taking lot of power.
Sometimes these problems occur even with good USB cable but poor USB controller inside of PC. So you can fix the problem very easily:
I thing that option 1. will solve the problem and then you will see that the device is rather reliable. Don't be confused at the very beginning.
Some other comments:
Blue light is ON when device has been initialized successfully. But it have to
be ON all the time. Even very short single blink means the reset of the device
and then its malfunction (software doesn't expect resets during command execution).
4.76V is just on the edge, it is sufficient for some commands but definitely not
for all of them. After device reset all parameters are set to their default values
by firmware so e.g. bias voltage value drops to about 4.5V. Once the bias has been
set the software will remember its value and will set this value at next initialization
automatically (just after start). So if you have set bias to e.g. 100V in software
and then you are measuring ~4.5V it is most probably due to unexpected device reset
after drop-out in supply voltage.
Q:
I have good cable but during equalization procedure the interface is reset anyway.
A:
We already encountered such behavior with some MXR chips. Some MXR chips take
significantly more power (even more than 400mA) than Mpx2.1 chips for which the
USB interface has been designed. For most of MXR chips it is OK but sometimes not.
Power consumption of MXR chip strongly depends on actual operation. So during
equalization process the threshold level is changing and at some point most of
pixels falls into noise => most counters are counting at very high frequency
=> at this point the power consumption can be extremely high. Very high power
consumption will cause voltage drop on 5V line (taken from USB) which will reset
CPU inside of interface (indicated by blue light turning off). Error messages after
such reset condition are a mess except of the first one notifying that there was no
response to some command.
With some chips we saw such behavior just with equalized chip (that means during
final scan in threshold equalization process i.e. during acquisition of black peak
in a chart). The reason is that pixels of equalized chip will fall into noise at
single threshold step.
Solution: You have to use larger SPACING to reduce number of pixels
scanned simultaneously. It should be at least 4 (so just 1/16 of pixels is scanned at
the time). Of course the equalization process will last longer with higher values
of spacing - but generated masks are of better quality as there are lesser crosstalks
and chip is in more stable state.