exxos's Atari, Amiga & retro forum

BennehBoy wrote: 11 Feb 2022 22:34 I sort of want a broken one, I quite enjoy the repairing :lol:

Me too!!

Today's win: Fixing a £3K NAS box with a 5p resistor.

The workaround for the failing output driver on the Intel Atom C2000 series is to stick a hard pull-up on the clock output line, as in ~100 ohm. This is Synology's official fix by soldering the resistor onto the pads on a diag socket. This is the same on the whole suite of NAS boxes from 2015/2016 which use the Atom processor.

stephen_usher wrote: 15 Feb 2022 12:14 Today's win: Fixing a £3K NAS box with a 5p resistor.

The workaround for the failing output driver on the Intel Atom C2000 series is to stick a hard pull-up on the clock output line, as in ~100 ohm. This is Synology's official fix by soldering the resistor onto the pads on a diag socket. This is the same on the whole suite of NAS boxes from 2015/2016 which use the Atom processor.

Wow, that's kind of mad. What is the issue it fixes?

Apparently the output driver on the chip for the two clock pins just dies. Intel's errata for the chip says no fix. The hard pull-up brings the signal up to a level that the other chips can use. In this case to drive the synchronous EPROM output so the machine can boot.

stephen_usher wrote: 15 Feb 2022 14:06 Apparently the output driver on the chip for the two clock pins just dies. Intel's errata for the chip says no fix. The hard pull-up brings the signal up to a level that the other chips can use. In this case to drive the synchronous EPROM output so the machine can boot.

Wow, that's a pretty bad fail!

stephen_usher wrote: 15 Feb 2022 12:14 Today's win: Fixing a £3K NAS box with a 5p resistor.

Oh wow, wish I'd known that a few years ago when my own NAS/router (DIY) with the same bad chip died. Unfortunately I had too much other stuff going on at the time to figure out a redesign, and so reluctantly bought another of the same buggy board. So far it hasn't failed, as far as I know, but due to the nature of this failure, one doesn't know about it until the next reboot. Uptime currently 311 days. :?

The fix is to find a 12 pin unpopulated header location and put a ~100 ohm resistor between pin 1 and pin 6 and bingo!

stephen_usher wrote: 16 Feb 2022 14:51 The fix is to find a 12 pin unpopulated header location and put a ~100 ohm resistor between pin 1 and pin 6 and bingo!

This would probably be a Synology-specific header, though, wouldn't it? Mine is a DIY system, I think with an ASRock board IIRC. I didn't throw out the dead board (do I throw out anything?), so maybe this is a good project, so I can have a fixed board ready for when the current one fails. :D

Yeah, it would be a Synology debug connector probably.

On a different front, the replacement key post for the ZX Spectrum+ has arrived so I should be able to finish the renovation tonight. I can also try the tiny springs on the Return key post to see if they make it less floppy and more likely to spring back up rather than jamming.