went outside with the #eyemodule today i think it's a little broken https://maple.pet/blog/extracting-eyemodule-pictures-in-2022
definitely something funky going on. could be eyemodtool's YUV decoder though
i accidentally overwrote the original PDB file with the pictures i took yesterday in retrieving the one with this test pattern so those are forever locked in exported funky mode
starting to think eyemodule1 color pictures are not quite YUV (UYVY), and/or that the eyemodule1 color format is different depending on the version of the eyemodule software (since i only tried before on newer visors which needed a newer software than the one inside the eyemodule itself)
the first shot was taken with a visor deluxe on the cartridge version of the eyemodule software, this one on a neo with the updated version of the software (the cartridge version does not support the neo), and it looks about the same, so it's not a version issue
either i'm doing something wrong on conversion or my eyemodule has gotten damaged somehow
for comparison, a picture i took with the first version of eyemodtool that supported the first eyemodule's color pictures (eyemodule2 just spits out straight jpeg)
it looks about right so something was being done right but i'm wondering if it's just off enough that more extreme colors would've always been broken with my converter
sadly i never actually used the eyemodule much to compare
oh, yeah, jeez. didn't think to compare with the generated b/w copy (used for displaying in the visor), it has the same waffle effect. i think the sensor itself is damaged, then. bummer
inside the first gen eyemodule. i was not expecting the sensor to be on a removable header!
oh oof that corrosion
HA! contact cleaner to the rescue!
the corrosion was probably causing some conductivity around where there shouldn't be
there was nothing wrong with eyemodtool after all
well this is a job for tomorrow
well shit. trying to put it back together bent and broke the pins in the connector that goes to the springboard port.
i tried bridging the gaps since i don't have a replacement but it just turned into a horror show...
it's not unsalvageable but i need a break. i'll try more later and hopefully i won't have burnt the pads off the board. removing the entire thing to replace with a new part is probably going to be harder without a hot air gun
however, i do know that the sensor is still working because before i tried to put it back together and broke the springboard connector, i did manage to snap these pictures. there was still a bad connection somewhere and wiggling the camera daughterboard made it do some weird stuff, like the second picture where it went upside-down and contrasted
i have a perfectly working eyemodule2 but the color pictures off that model don't inspire me much. the first-gen eyemodule is much rarer and i'll do my damndest to fix this unit
well.. at least all the pads seem intact. reusing the broken connector was a bust, now this won't be usable until i source a new one...
it's not the same footprint but it is the same pitch... if i can keep all legs straight and not bridge any of them it might work...
unfortunately something is still affecting the quality... i know the eyemodule isn't brilliant, but you saw how the test pattern looked when it worked right
pressing slightly on the corner where the broken solder joint was vs letting go of it. it's clearly a break there, but i've already reflowed all of the pins that go to the sensor... well, at least it's clearly working, i just have to fix a trace somewhere
astonishing how something so promising goes south so quickly
the modular connector i praised earlier in this thread has been nothing but hell the entire day. there's a bad contract gremlin somewhere i can't pinpoint. i've put fresh solder wherever i could. every time i reflow a joint something new happens
i really wish i could pick the brain of the people who developed the first eyemodule for technical information on how the camera works and the file format it outputs