assasincz wrote on 2020-06-06, 14:56:
Recapping finished. I also hot glued a few things down and did some cable management. I am glad I did it, since there were indee […]
Recapping finished. I also hot glued a few things down and did some cable management. I am glad I did it, since there were indeed some leaking capacitors, as shown in the picture. I replaced all except one on CRT driver board that I could not find a replacement of proper size.
I removed the 80mm fan that moves air from system board compartment to the monitor compartment, and put in its place a Noctua fan, and wanted to power it from the system board. The trouble is that the power supply board freaks out when there is no fan plugged to its fan header, and does not power on properly. Running the Noctua from the header was not working. So I shorted the fan header pins and now it works nicely.
I also stuck a 40mm Noctua fan to the Overdrive heatsink. You may notice CT3670 and SD-IDE in the system. Also now there is 16MB of RAM.
I am properly satisfied, having the 460 next to me running completely silently, playing Doom demo as I write this post...
Now you're making me want to crack my 433 all the way open and inspect the caps.
I pulled mine out of storage a few days ago and have been slowly going insane. I'm one of the masochists that ran 95 on mine. I was never able to get the onboard sound working properly in 95. But, alas, the hard drive I had in there was dying and I grabbed everything off it I needed and wiped it. Currently the drive has been replaced with a 128MB IDE DiskOnModule to ease testing.
Now to the problem: When I put it away, it had 20MB, working flawlessly.
When I pulled it out of storage, I had to clean up mouse pee and poop off the board (small bits! cleaned up nicely.) and now I can only get it to ever see 12MB no matter what I do. I even pulled out the stash of 72-pin stuff, set aside the EDO, and tested what was left... turns out, not much at all. The two sticks I had in there were IBM branded SIMMs with 16 RAM chips with IBM part numbers on them and two Samsung I can only assume were for parity... ? but these sticks have ALWAYS worked and I think I pulled them from an IBM machine I was playing with at the time 7 years ago. I believe that's also what I acquired the Overdrive DX4-100 from.
At the end of the day, no matter what sticks I use, the machine can't see more than 12MB installed, meaning one of the slots are not registering with the system in the slightest.
Edit: Just confirmed, the dead slot for me is the slot nearest to the CPU. the rear slot if looking at the RAM slots and ISA slots head-on.
Cleaning, resoldering pins, adjusting pins in the slots to make better contact, I've tried it all. I've tried it all so much that the go-handle on the back of the motherboard tray needs to be babied since two of the clips that keep it on the tray are broken now. Contemplating drilling them out and attaching it firmly with nuts and bolts... but, I'm at wit's end with this 433 seeing more RAM. I even tried swapping back to stock with the SX-33 I pulled out of it. No dice.
As far as those that CAN see the full 20MB, the VLSI VL82C486 on these boards DOES support up to 64MB of RAM, so this is something else to look into for getting these machines to see more than 20MB. I have a feeling that the BIOS is what is limiting the RAM ceiling and not the boards.
For convenience, I've attached the datasheet for the chipset in PDF form so those that know more than I can look into this a bit closer.
The only way we're going to know if this is BIOS limitation or board limitation is to get a full dump of the BIOS chip in these. and, while at it, a dump of the modem BIOS too, because why not, it's there and should get dumped. it may help someone out someday. Sadly that person is not me, I don't have the necessary tools to dump the ROMs for analysis. I implore anyone out there with one of these that can, or has a friend who can, get a dump so we can pick it apart and figure this out.