ESD-protective encapsulation sounds really interesting. At what airspeed does triboelectric charge become significant?
We would need something in liquid thermoset form that doesn't darken under UV (sunlight).
Rich, to break this down to a finer level,
1) Do any of your tags work reliably time after time or do they all fail intermittently? Design issue?
2) If voltage spikes are suspect, inserting a zener diode in the circuit might be an interesting design revision.
3) Going back to electrostatic discharge, if your tags are encapsulated in any non-conductive plastic package like polycarbonate, they will develop charges on the order of 3-5 KV just from air friction leading to ESD events. Using PEEK or similar conductive plastic will dissipate charge collection and ESD events.
Since the tags fail during deployment could we be seeing SEU problems due to solar radiation? What altitudes are you logging among similar tags that survive?
It also is possible that your stripped-down ultra-light design allows more exposure to radiation than ordinary enclosures.
The cell is loaded during transit. With X-Ray this may "perhaps" lead to permanent traps that "may affect" the light to electricity conversion efficiency. Also the 0 instead of 1 logic you mensionned may be stuck bits that can be also due to ESD damage internally in addition to the I/O.
Side question since the Voc is 4V, I assume the lithium battery is at least 2 diodes voltage-drop below the cell Voc?
"these failures are observed after they've been in use"
Possibly moisture getting into them? How are they sealed?
Amendment to previous test - immerse some devices in salt water while at the beach. Your downhill ski idea is also a very valid test for simulating the shock and temperature extremes of a bird in flight.
The PV cell Voc is 4V, Isc is 50 microamps. The cell is loaded during transit. A diode on the output of the LDO prevents reverse current flow through it. Checking output shorts and opens postmortem is a good idea. Thanks!
You may have several problems that occured durang the cicuit build-up! ESD may be one of the culprits. You maybe able to see the I/Os quality with a voltmeter. ESD may create a short (metal melt from excessive zap). If that the case your data output maybe shorted after the ESD event. As for the X-ray, silicon based PV may survive X-ray pulses if not loaded. I am not sure what you have!. I assume the PV is not loaded (to generate electricity) during transit from FedEx for instance. Finally what is the output voltage spec of the PV (Open circuit voltage Voc, and other!). That diode may act as back to back scheme limiting current flow to the LDO!