I present here the results of my investigation into Gunnar SM5EIE's failure mode (a step increase suddenly from 6V to 12V supply); and also concerns of any potential vulnerability to intermittent?supplies common in field scenarios. See? for the full article... the executive summary is:
I constructed a test jig to produce a sudden step input voltage from 6V to a selectable level. I was able to reproduce a failure when going from 6V to 12V in a sudden step.
I was able to substantially improve the control loop in QMX to make the response to large input voltage steps very much faster (approx 100x faster), which will greatly improve its resilience to step input voltages and other similar forms of abuse, from firmware version 1_00_008 onward.
Despite this improvement, I was still able to damage QMX by a sudden step from 6V to 12V. Therefore the recommendation must be to AVOID this type of extreme scenario where a large step in input voltage occurs. I do not consider it unreasonable to avoid such an extreme and cruel scenario.?
As a further experiment, I tried intermittent power connections; in normal use this causes no damage and just causes the QMX to power-down. With a great deal of effort and using my supply set to 12.5V, and keeping the left "ON" button permanently held down, I was able to eventually damage QMX. Again, a rather unrealistic scenario.
In ALL cases of "QMX damage", QMX proved remarkably resilient; the only damage was to the 5.6V zener diode D108 which took the hit EVERY time; replacing it restored the QMX to normal operational condition, all functionalities without performance degradation.?
