It's not too late to educate... That might be a fred original, but no matter.
I've removed the questionable glass tube from my machine after performing another test, ensuring I had good connectivity on the ends. Your suggestion that I ensure the insulators on the anode and cathode raises a question. The wiring is securely attached and the failure came halfway through a test job of perhaps 30 minutes total. The first portion was working well, the second ended with lightning bolts at the hv end.
I've had discussions in the past regarding high voltage, unrelated to laser cutters, but certainly relevant here. The voltages were somewhat lower but the currents many magnitudes higher. This particular team informed me, when I ignorantly questioned them about insulation, that the best insulator is dry air and distance.
Allowing for the fact that one can not easily change distance within the cabinet and at these currently levels, don't really have to worry about that, the dry air part comes into play. At the levels of these cutters, adding a silicone plug in a piece of tubing is not really going to reduce the insulation level appreciably either. Would you agree? If that's the case, then not having a silicone plug in a piece of tubing isn't going to increase the insulation level either. I suspect the silicone plug inside the tubing is more a safety feature and minor mechanical securing method than any insulation factor.
Onward from there, the problem probably originated from the excess condensation on the tube, due to strong cooling and low, infrequent power use. I might have been at 30 percent and running for too short of a time to "burn off" the condensation. I'm going to strongly suggest that burning off condensation is a poor practice as well. It wasn't in my operating philosophy and won't be.
I've added the high volume high speed muffin fan to the cabinet and during the one test I performed, a half hour of cold water produced zero condensation, while almost immediately after shutting down the anti-evap fan, I had condensation haze on the tube.
Wouldn't you allow that having a nice thick layer of water over the hv end of the tube is bad practice, and that it would permit these wonderful lightning bolts to travel to ground? I suspect it is/was the water, as the spark began at the top of the tube, near the hv contact point and traveled along the outside to ground. This explanation fits in with the perceived curvature of the spark I saw as it was zapping away during the scan process.
Of course, when the replacement tube arrives (or is it "if"?), I will take additional care to ensure that there are zero bubbles in the tube. Draining and removing the existing tube shows me that it is a simple enough process.
Redsail clone 60w 700 x 500
home upper right
Emblaser A3 4w diode laser
BCN3D Sigma Dual Extruder 3d printer
Cube 3rd gen 3d printer (works great, for sale)
Under construction: OX CNC 1000x750 router - out of the box, still in baggies!