Analog spikes

[Firejo]

I would approach the diagnostics by looking at what’s common at each event. I’d start by looking at the values just before and just after the spike. Is the before value always the same or is it different every time? Maybe the actual sensor has a fault where at 0.25ppm it reports 8mA instead of 4.5mA (guessing at numbers here).
Instead of thinking of what could be causing the spike and then seeing if that’s happening, look at what is happening before, during and after the event and start looking for patterns. If you start seeing patterns then you can start looking at what would cause the pattern (rather than just the event).

 

[westom]

I would approach the diagnostics by looking at what’s common at each event.

His recommendation is spot on. A difference between shotgunning and approaching a problem as defined in the show CSI. "Follow the evidence".

But remember, every voltage must always be observed between two points. For example, both incoming wires could carry a major spike through a conneced box. But measuring between both wires shows no voltage - no spike. Every measurement must be made between two point.

Of course, this is easy (almost trivial) if using an oscilloscope in differential mode.