First post, by CharlieFoxtrot
- Rank
- Member
I've been gradually working with this model 35SX during the last couple of months and I've pretty much finished it during the last week or so. I will post the build log here probably in few parts in short succession within a day or two [edit: ALL PARTS ARE HERE, SO BUILD LOG IS COMPLETE]. So, without further ado, let's begin:
PART I - The Starting Point
I got just the box couple of months ago. It was in very original condition that is, it had the 386SX-20 processor, 4MB RAM (upgraded from 2MB at some point), 40MB IDE HDD and that's about it. It has a Dallas CMOS thingy, but some previous owner had made the CR2032 mod already. Many of these systems had horribly unreliable Magnetek manufactured PSUs, which tended to release magic smoke after couple of years. This probably had one too, but PSU had been replaced probably sometime in late 90s, as the PSU had date codes from 1997 and it was made in Taiwan. It was pretty clean, just some mild dust inside the case and PSU had pretty much none.
Pretty clean system overall and no major scratches or bumps to be seen:
Looks clean inside too! Standard ISA, molex connectors on PSU and Dallas has CR2032 mod on it:
Standard IDE, but IBM proprietary floppy with 34-pin cable and without separate floppy power cable:
It was fully working besides the floppy drive. These systems pretty much require working floppy drives, because changes to the system require using setup floppy. However, drives are proprietary and depending on the system, making an adapter for standard floppy drive is either relatively easy or pretty much impossible. Mine is most likely the latter, because usually these newer PS/2s expect a media sensing drive. Pinouts and signals for a 34pin media sensing IBM floppy drives are completely different from standard, so adapter most likely requires some sort of logic and I couldn’t find any instructions for building one. So project started with repairs.
PS/2 floppy drives are known to fail. Usual suspects are dirt that gets inside because of open floppy slot and even more so, leaking SMD capacitors. My drive was pretty clean and it did rotate the floppy, but it just couldn’t read anything. My computer has a drive manufactured by Alps, and it has practically two circuit boards. The motor board, which has components visible from the underside of the drive and the controller board, which has components towards the top of the drive so they can't be accessed without removing the controller board. So, after removing couple of screws more and flat flex cables connecting the controller board, I had the drive disassembled in a way that I could change all the caps.
Gladly leaked electrolyte hasn't damaged anything, but pretty much all the caps had been leaking. It was evident when I cleaned the solder pads with wick as you can't help but notice that fishy electrolyte smell. Drive had all together 9 caps, five 10uf/16V, one 4.7uf/25V and three 4.7uf/16V of bipolar type. I had everything else except SMD bipolar caps so I replaced those with regular through hole stuff. There are plastic parts near some of the caps, so I was hesitant to use hot air while removing the caps so I went with a twisting method and I managed to rip one pad from one of the bipolar caps on the motor board. Fortunately it was still easy to trace and ripped pad was connected to the leg of one the SMD resistors near it, so I just soldered the leg of the through hole cap to the resistor.
Motor board before recapping, top three caps are bipolar:
Controller board before recapping:
So, I recapped and re-lubricated the drive and after that it worked 100%. Unfortunately I didn't took photos of the end result, but on the other hand, there is not much to see. Now that the floppy was functioning, it was time to start building the rest of the system.
To be continued...
Edit: fixed some typos