Agreed on the hot bed. If you’re interested, I posted my devations at ambient, and at 60C, in a 10x10 grid, with a dial indicator. The deviation is small, but significant. On achieving a perfect level.
The hot nozzle makes sense if you’re using a feeler gauge. With a separate probe, it shouldn’t matter except right at the end, and if you measure it once you should be able to leave the nozzle cold and adjust it at the end. At least for my bed, I have had good results with the nozzle cold, and it reduces some risk of melting through the print sheet.
I have a glass sheet in the mail, and am doing a magnetic base over it so i can use the same removable print surface. I don’t think there’s anything wrong with the removable print surface, and using it is not at odds with a glass plate, entirely. If you really prefer printing on glass, then yea, that’s an issue.
Also, I have some other topics on Z probes. I’m working with a few others to improve Z sensing in CNC, and there’s definitely some ways to improve the 3D print also. BLTouch won’t be an option until we can safely upload firmware to it, which at the moment requires a chip flash memory programmer; not very accessible to the normal user. But there are other options, including replacing the inductive probe with something else, that would require manually flipping down, or something - like the allen key levelling method. In the past I have also coated the glass bed of a different printer with aluminum foil and glue stick and used the nozzle as an electrical contact sensor - that worked pretty well, and also should be inductive compatible, though initial results seem like getting the aluminum flat and even enough for the inductive sensor is difficult.
I will say I did check the inductive probe accuracy and repeatability, and it’s damn accurate. +/-0.01mm. I think that’s really good, considering we took about 40 samples, and all samples were in that tolerance.