I finally got my power supply adapter in the mail at the end of April 2018 after an absurdly long wait. That was the missing link in the chain and with it I could finally start testing components. This will be a shorter update- believe it or not the project is a few days and an entry ahead at the time of writing.
I tested both my original Gigabyte board, the GA-586 ATM/P, and the substitute Azza PT-5IVH. I tested both with a bare minimum of components,
The GA-586 ATM/P has been through hell. I ran the first test with the Pentium 133, 16MB of what I’m pretty sure is its original RAM (many of the “original” parts are 1990s upgrades), and the Mach64 graphics card. To my complete and utter surprise, it POSTed and entered setup just fine!
Emboldened, I got more ambitious. I want 64MB of RAM in this machine, but I had to try several configurations to get it to recognize the full amount. I think this is a mechanical issue. The RAM slots are missing clips and several of the modules barely stayed in. I’ll probably zip tie them in later or something.
The Pentium MMX 233 was recognized as a Pentium-S 133. I thought this was odd because I had set the jumpers for a 100MHz processor. This is the correct configuration for an MMX, and I figured this motherboard would recognize it as a Pentium 100 if it didn’t recognize it properly. Thinking back, the Pentium-S 133 misdetection does make some sense. When this motherboard was made (or at least when the BIOS was written), I think that was the newest and fastest processor. It certainly couldn’t be “remembering” the last good CPU, not with a dead battery.
I also tried the Kingspec DOM and that was an adventure. A known sometimes-issue with these is that they tie the key pin to the 5V line. On most motherboards this is fine because it’s missing or provides 5V, but on mine it’s (apparently) grounded. I tried plugging the DOM in without power and the board refused to POST at all. I plugged in the power cable, and the system refused to even turn on. I ended up tearing open the DOM and clipping two of the pins (I missed the first time). Now it boots and gets recognized fine.
While that board POSTed and I was able to get it to recognize the Kingspec DOM, it refused to boot from any of the floppies I tried and just rebooted when I tried to continue after the clock battery warning. It could simply be that this board won’t boot without a working Dallas RTC module. My floppy drive might be dead or connected backwards. Or it might really not like the MMX 233.
I was very worried about the PT-5IVH because it hadn’t really been adequately packed for its trip around the world. However, it appeared to be in good physical condition and booted up with no problems. The Pentium MMX 233 worked fine as well, although swapping the neat yellow jumpers was quite tedious. Even the coin cell battery seemed to still have charge, though I’ll probably replace it at some point.
The Next Step
I’m pretty sure the Gigabyte board won’t fully boot without a working Dallas RTC module, and I really need to replace it anyway. Fortunately, I have some on order, though I might get impatient and hack a new battery into the existing one.
After that? Start building the system, installing software and trying hardware! With the critical components tested, it’s time to start playing with the other pieces that are going to make it awesome. The open-air test rig is really awkward, though, so I’d prefer to have a case to put it in.
Since both motherboards work and I have multiples of several other components, a “leftovers” build might be in order.