I’ll take a look.
The replacement probe, while it looks very similar and labeled “DT02”, is a different build and that one just didn’t work consistently. The aluminum plunger’s slide is very tight and draggy and tends to hang in the plastic bore when extending. The spring has to be light to be safe to work with the very fine-point, fragile bits, so we can’t strengthen the spring.
I tried working it up and down for a few minutes, thinking it might rub off a plastic burr or whatever, but it didn’t change. Lubrication’s not a great option as it’s in a dusty environment and if it mucks I don’t see how it could ever be cleaned where it’s at. I assumed this would be easy to disassemble, somehow loosen the plunger in its bore, and be as excellent as the earlier probe. I don’t recall the details but there was some sort of problem with the way the plunger is retained that made it difficult to take apart in a way that was amenable to being put back together. We can take another stab at it, though.
I don’t know if another Amazon probe will be different. Maybe, this one could just be a bad batch. It wouldn’t make sense for them to keep making probes that can’t extend correctly.
Several cases for damage came from leaving the spindle running with the bit in it. I could add some code to force the spindle off if it’s on, but that won’t guarantee the spindle spins down to 0 RPM before probing. There’s no condition like that in the VFD’s interface into LinuxCNC. Maybe the better condition would be that if the spindle is running, abort the probe cycle entirely, that does make sense.