Note that there wonky crystal is the same on both boards, and seems to be some kind of weird hack to put a thru hole component on an SMD only board.

Oh yeah, I've tried a different power supply but with no change.

There's also a chip on the board that seems to get kind of uncomfortably warm. So perhaps it really is just entirely dead.

@hp Sticker says the board is 2001 vintage, which would be in scope for the capacitor plague. If able, I think my first stop would be swapping out those caps and seeing if that gets it to boot more. Symptom would be decreased capacitance and increased ESR, which if those are the bulk caps for the power rail, could result in unstable Vcc. That could certainly get everything else to malfunction.

I don't know if those form factor caps were prone to the plague tho.

@danderson The caps *look* okay, but that doesn't necessarily mean anything. I'll try to replace them on one of the boards.

As for airflow etc, nah, these are just in a plastic shell. And it all feels incredibly cheap. Poor solder job, cheap plastics...

It might just be utter shit.

@danderson @hp I think I'd go straight for the 2 caps on the north west peninsula. Sounds like power supply rail may be unstable, I'd guess they are filtering the noise usb 5V input.
Other consideration may be USB, like its USB 1.1 maybe even 1.0. Dunno if what u are plugging into may be >2 and may not support or kills the supply. 🤔 Out of my depth and possibly talking through my hoop now.

@Doomstrike @danderson it's not USB powered, it has a separate power supply. The symptoms I got were all without even plugging in USB.

@hp @danderson ah OK so. I'd go for the caps then. Two failing boards says to me it is something time related, and everything else there should be largely unaffected, unless any obvious corrosion etc.

@Doomstrike @danderson Well, ordered a set of caps of various values. I'll let you know if it does anything!

@hp @Doomstrike Good luck! Fingers crossed.

A long time ago, I helped someone fix an HDMI receiver that was intermittently failing to power on, which involved replacing half a dozen caps in that form factor. None of them had visibly failed, but some combination of them had degraded enough that successful boot had become probabilistic and heavily temperature dependent. Similar symptoms, some LEDs would light and chips warm up a bit, but no activity. After cap swap, good as new, <1s bootup.

@hp @Doomstrike re that IC being 40-50C, I'm not a magic smoke doctor but that seems like a tolerable amount of heat, if it settles at that level. It's toasty, but I don't _think_ so much that it'd be awful junction temperature on the inside. Plastic enclosure might get a bit warm as it dissipates that to atmosphere, but not hot enough to be a safety concern, and 🤷 cheaper to not bother with fancier cooling if the chip's not in danger, is the usual mass production ethos.

@hp @Doomstrike It could also cool down more once it boots successfully. If there's not enough supply rail capacitance, one possible failure mode is the microcontroller initializes, tries to power up some hungrier parts of itself or the board, that extra current draw causes the supply rail voltage to droop, triggers brownout protection, microcontroller resets, rinse, repeat. Effectively running at 100% utilization in a boot loop, instead of initializing then being mostly in low power sleep.

@danderson @Doomstrike That's a good point! We shall see!

I'll replace the caps, I should get them on Sunday. Should be an easy job.

@danderson @Doomstrike I have two more zip drives of a slightly different vintage that are also broken.

I'm beginning to think these devices really, really suck.

@hp @Doomstrike Skimming the history of zip drives, I'm not surprised. They had a 4-year reign between data growth making floppies impractical and CD-R taking over, and only shipped a single design in that time. Probably different manufacturing spins, but that's not a ton of time to iterate on reliability. And then in 1998 they got a class action lawsuit for high failure rates, apparently?

Certainly doesn't paint a picture of being a paragon of reliability.

@danderson @Doomstrike Well, I definitely have two models that are very, very different from each other.

All broken though :)

@danderson @hp @Doomstrike I used to own an internal 100Mb SCSI version, later on upgrading to 750Mb USB.
They were very unreliable - see "click of death".

@hp @Doomstrike Huh, all 100MB? They sold them with a bunch of different bus interfaces, so shows what I know I guess.

@danderson @Doomstrike the PCBs are completely different, as is the actual drive mechanism.

The more you know! 😄

@hp Yeah could just be generally dead, just trying to figure out what would cause runaway heat rather than none heat. I guess power surge making the IC fail in a dead short, but a pair both dying the exact same way... I dunno it's possible but makes me look for common factors in manufacturing.

I guess you could probe out the power rails if you can find them, see if they look stable? And maybe measure current draw and see if it's worryingly high?

@danderson I don't really have a very good bench power supply, so measuring current draw is going to be a bit challenging.

It isn't getting "dead short" hot, just like... way hotter than I'd think would be reasonable given the device. Probably like... 40-50C? based on finger feel.