• Please review our updated Terms and Rules here

Troubleshooting OPTi 486 VLB motherboard

Denniske1976

Experienced Member
Joined
Jan 17, 2011
Messages
480
Location
The Netherlands
Hi guys... well, after a week of loosing tape after tape and also 3 tape drives (QIC80 with all my old data, tapes didn't even survive a retension and I had about 50 tapes). Now also my 486 seems to have passed :-(

The motherboard (or is it mainboard?) looks totally OK, I can't see anything that even looks to have cracked or burnt at all. What happened is that I acually was able to power it up (it's been standing in the loft for about 3 years now), it booted normally... then just powered off. And since then when I try turning it on, the PSU does it's thing about 1 second, LED display comes up (you know, the cool 66MHz LED) and then it immediately shuts down again. No beeps no nothing. So the mainboard shorts the PSU I guess. Tried another PSU and same problem.

I'd really like to try and fix the board, since it's been with me for the last 30 years or so and it's a very nice board: 3x VLB, 8x 30pin SIMM, OPTi 495SLC chipset and 256K cache, oh and there's no battery corrosion since that I took that out a long time ago and replaced it with a CR2032 on the external battery connector. What would be the correct approach as to try and repair this? (I had this happen to 550/5160 motherboards before and you just cut the C56 or whatever blew up and all was cool). I do have a multimeter, so my guess is start with a PSU attached and see what the pins of P8 and P9 read?

Next weird thing: being a hoarder of some kind, I found another 486 motherboard so I put that in the case (the original board is looking pretty in my display stand now). Also an OPTi board (just not as nice as the other one). So all seemed to be allright with the world again. I installed the board, replaced the CPU (Intel 486DX2-66), 4 sticks of 30pin 4MB SIMM in it, Gravis UltraSound MAX back in place, disk controller (VLB) back in place and connected to my NEC 4speed CD-ROM and Quantum Maverick 540MB harddisk, 1.44MB floppy. Put the Diamond Stealth 32 back in place (2MB VLB). Powered on and all seemed OK except that there was a low battery message on screen and I had to run SETUP (because I just connected a new CR2032 to the external battery connector). So I went into SETUP, set the date and time, A-drive floppy and autodetect the harddisk (and that found the right settings for my harddisk too). EXIT and SAVE, reboot and all seemed OK. EXCEPT: I can't get it to read floppy drives (and that's weird as with the other board yesterday I was able to boot from a DOS 6.21 floppy and use other floppies, and all of a sudden now it won't read that DOS floppy nor any other and when I try to format a new one all I get is error reading disk ABORT/FAIL/CANCEL or whatever). The other weird thing (I guess not related) is that I set the date correctly and when I reboot is back to the same month/day but ALWAYS 2094... dafuque is up with that? So when I disconnect the battery, it resets to 01-01-1989 I believe and that's normal behaviour. Then I set the date in SETUP for 02-08-2021, I exit (and save of course) SETUP, system reboots, all is fine: 16000KB memory counts, P24D-S at 66MHz (for some reason it won't list the CPU as 486DX2, even the jumpers are set for it but in the manual jumpers for 486DX2/DX4 and P24D are the same)... I end up in DOS and when I type in the DATE command it comes up with "Current date: 02-08-2094" and when I CTRL+ALT+DEL and go back into BIOS SETUP the date is also 02-08-2094. Every time again after I cange it back to 2021 it's going back to 2094 (and not 01-01-2094 or something, no the month/day I set but just in 2094).

So now it seems I'm without a working 486 :-D :-( But anyways, if any of you guys would have some tips on that defective motherboard? It kinda looks like brand new and like I said: it was my first VLB 486 board and always worked great. I can't imagine this would be dead now. So any help would be greatly appreciated :bigups:
 
If two different PSUs go into protection and shut off, there is definitely a short on the board somewhere.

Get out your multimeter and start checking from power rails to ground. If you find any that are a few dozen ohms or less, you probably found out which rail is shorted.

If the board has any tantalum capacitors on it, you'll want to start checking those. Either remove them and test them, or just replace all of them.
 
What board is it? Sounds like it could be a PX486P3 http://www.hyakushiki.net/misc/px486p3.jpg

Did you try powering it on with cards removed and drives disconnected? Maybe that would reveal that the short is not on the board itself.
The other weird thing (I guess not related) is that I set the date correctly and when I reboot is back to the same month/day but ALWAYS 2094...
Likely the BIOS is not Y2K ready. I fixed that on one of mine by adding a small ISA card with an option ROM that replaces some code. I'm not sure what else you can do if there isn't a BIOS update from the board vendor.
 
What board is it? Sounds like it could be a PX486P3 http://www.hyakushiki.net/misc/px486p3.jpg

Did you try powering it on with cards removed and drives disconnected? Maybe that would reveal that the short is not on the board itself.

Holy cr*p, that is EXACTLY what it is... how'd you guess? I never knew the board brand in all the years I have it. Here it is (sitting pretty behind a SuperSPARC and some Itanium and Pentium CPU's:

486opti.jpg

Guess I'll start off with checking the power rails then. From what I can see there isn't any damage to the tantalums, board has a few. I'd be happy to replace them, justhave to find out what the values are of course. At least with those IBM boards they just crack open or really blow up... not hard to troubleshoot that. So uhm, I just attach a PSU, apply power and then measure the Ohms with my meter (black on GND and red testing the other pins one by one), right? I know a bit of electronics and some soldering but I'm not an engineer by any means, so I don't wanna f*ck up (for reference: I can manage to replaxe the X2 RIFA caps in an exploded 5154 monitor or change the cap that keeps the RPM on my Rega turntable consistant, that's about as far as my electronics goes... OK and build a NIXIE Tube clock... with directions and a schematic that is) ;-)

And yes, I tried the board with just a PSU attached and same problem... power for just about a second and then deadly silence (drives won't even power up when the board is connected, and power up just fine with the board disconnected).
 
Well, from what I just measured with my multimeter set to Ohm:

All three +5V rails show nothing (wel from -168 to then 0) or max 10 Ohms.
-5V rail actually "charges" (so goes from 200 to 450 to 730 etc).
+12V also "charges"
-12V does the same as all three +5V rails do, minus something to about 10 Ohms and nothing else.

And all Tantalums sit there being pretty, see picture (sorry, phone pic and I'm not allowed larger than 256K):

486opti_01.jpg

So uhm... any thoughts? Do I check the tantalums themselves for Ohm values?
 
You can remove them from the board one by one until the short goes away.

Another method would be to use a current limited power supply set to like 0.5-1.0v and put it across the shorted power rails and start feeling around for hot components. Don't go up too high current wise though or you'll start blowing traces.
 
I'll try that later when I get home... I did check the tantalums and from the four near the P8/P9 connector I have three actually charging when measured with the Ohms meter but one goes from -168 to 0 and then nothing. Maybe that's the one, but I'll snip 'em 1 by 1 until something changes or it works again.

I guess I'll have to replace them afterwards, right?
 
Holy cr*p, that is EXACTLY what it is... how'd you guess? I never knew the board brand in all the years I have it. Here it is (sitting pretty behind a SuperSPARC and some Itanium and Pentium CPU's:
No, it's not. It's not even close. Just look at the position of the BIOS chip and the keyboard controller. Or simply the fact that your board has *two* 8-bit ISA slots instead of just one... These are different boards.
 
YAY!!!!! It's working again! Sooo happy... So, I desoldered the tantalums @ PSU area. Although they hadn't exploded, two of them were "kinda black-faced" and with all 4 gone the board fired right up again. So I cleaned the soldering holes with some copper weave and looked in my stash to see that I had a bag of 16/10 tantalums (I think I had those left from when I've been tinkering with a PET 3032 that wouldn't work). Anyhoo, replaced all 4 tantalums and the board still worked.

486mobo01.jpg

486mobo02.jpg

So there she is, all up and running (and formatting a QIC tape):
486mobo03.jpg

Weird thing is: this one does keep the time/date correctly even though it's an older mainboard than the other one that keeps going to 2094... and also, using the same floppy controller, cable and floppy drive, floppies are being read and formatted without problems and the tape drive seems to work without a problem too. So I guess that other mainboard isn't reliable.

But, thanks for pointing out where to start. I couldn't be happier right now!
 
Back
Top