Show more

The tip isn't too bad. It's within a factor of 4 or so of where it should be. Since I'm taking these photos using an emitter I made (that I couldn't inspect), I know I'm on the right track.

The etch took far too long and ended up making a long thin needle.

The second failed tip was a tragedy. I must have damaged it putting it on the stub. There's no way it was that blunt (and covered in carbon) when I installed it.

First few hours with the new emitter went well. Alignment was easy and the spot is very bright. Looking at previous failed tips is interesting. This one is way too round. I think it melted from overcurrent and overflashing. The spot weld looks nice and clean though.

After months of effort reverse engineering an FEG emitter that actually works, I may be laughing a bit more maniacally than usual.

"It has been my experience that new tests often uncover things that were unexpected and not easily understood." Corollary to "what you do not measure you cannot understand".

My only glass infill base. Etched, aligned, installed, and ready to bake!

Yeah, that is pretty nasty. I guess I'll have to go with glass infill after all.

Well that is unfortunate. The UHV epoxy apparently did not withstand the heat of the bake, and decomposed. 😑

I'm amused and annoyed by this. Annoyed because the motivation seems to be the same kind of gatekeeping that uses new terms, acronyms, greek letters, and mathematical notation to obfuscate rather than illuminate. And amused because the gate hit them on the ass.

TIL that some ML researchers had an idea (batch normalization) that worked, then tacked on a post-hoc mathematical analysis to claim that it reduced "internal covariate shift". The technique works, but not for that reason.

Show more

Where the 1337 meet to federate. Home of the most interesting local feed on the Fediverse!