• Please review our updated Terms and Rules here

Debugging a 486 DX-33 VLB motherboard (and VESA video card)

R

RickNel

Veteran Member
Joined
Apr 24, 2009
Messages
641
Location
Canberra, Australia
I've not had a working 486 mobo for a while, and just picked up a generic one that looks in near-original condition. It's branded "US Technologies" (sure sign it wasn't made in USA) and model TK 82C480 - with UMC 82C480 chipset. There's a description and photo of it in the collection hosted at www.redhill.net.au, and I found a perfect visual match for it under the "Informtech" brand in the TULARC archive of ISA/VLB motherboard jumper pinouts

BUT

I can't get any boot video signal out of it, having tried several VGA and a VLD video cards in it.

I don't have a compatible board to swap socketed components - ie Intel DX 33 CPU, 8 x 4Mb RAM, 256k cache RAM.

Here's what I've already done trying to isolate problem:

1. Found visual match for the board on TULARC and checked all jumpers are correct for identified components
2. Tried adjusting mobo VLB wait-state on/off
3. Checked keyboard flashes LEDs as expected on power-up and system power/turbo LEDs and switches function correctly.
4. Checked all three oscillators (system, bus and I guess keyboard) with oscilloscope for clean signal.
5. Verified 33MHz present on ISA bus pin 30 (OSC)
6. Found no clock signal present on ISA bus pin 20 (CLK) with no boards in bus.
7. Touch-tested CPU and chipset chips all about as warm as expected.
8. Pulled AMIBIOS ColorBIOS 27C512 from socket and read it - seems OK so far as all ASCII data in the BIOS can be seen without errors.

I'm an 8-bit kind of guy and not too familiar with ISA hardware. What tests can I make to locate faults?

The motherboard has no jumper to enable/disable VLB video. Should I expect to see a VGA video BIOS message when trying a VGA card in the ISA bus?

I'm trying a Trident VLB video card based on TGUI9420, but with no docs found yet . Jumpers as follows:

1. Select STMB (two options of ARP/BRP configurations - are these on-board memory configs or upgrade option configs - there are a couple of unpopulated 40-pin SMC locations on the PCB?)

2. Select I or NI (is this interrupt/non-interrupt mode?)

3. Select +RAS on/off (what is this?)

4. Select Fast clock on/off

It would be nice to get a VLB video card working after all these years.

As to the motherboard, I'll only write it off if there are faults in the surface-mount UM chipset. Anthing socketed or minor discrete components I will try to replace, if I have a working board chipset.

Is there a standard procedure for this sort of trouble-shooting?

Rick
 
Last edited:
Just a shot in the dark. You said you checked all the mobo jumper settings. See if there is a jumper for 'Color/Mono'. On some boards that jumper may sandwiched between the card slots or otherwise 'hidden'. Although there may not be a specific jumper to enable the VLB video card, but there may be a BIOS selection to enable VLB. I threw that in here because one of my 486 VLB boards has it. Another long shot would be to take a magnifying glass and go over all of the card slots, especially the VLB slot area, and check for shorts in the slots. I assume your P/S is okay and all of the rails check good under load. And while you're poking around, don't leave out video plug/cable on your monitor.
 
As a general rule, try to add the fewest components possible. So fill only one bank of RAM with the smallest capacity of RAM you can find. That probably means a 16MB system unless some 1 MB SIMMs are lying around. Make sure all the RAM is of similar type. I have found several 486 motherboards that won't boot if RAM is placed in the second bank because they cheaped on the design to make all the RAM sockets work.

The VESA motherboard should work if you only have an ISA video card and no VLB card. Sometimes, it takes several tries to find out what VLB settings work for both motherboard and VLB card(s). ISA only video can provide evidence as to whether the motherboard works at all.

What CPU are you using in this motherboard, just to confirm it isn't a wrong voltage style issue? Are you sure the power supply is good and large enough?
 
Thanks guys for both sets of suggestions -

I rechecked the color/mono jumper, checked/cleaned all slots. CPU is Intel 80486-DX 33. Checked all 8 x 30-pin SIMMs - all correct specs but smaller capacities than I had guessed. Each SIMM has 2 x 4400 module chips, so I calculate 1Mb per SIMM, total 8Mb in 2 banks, all 70ns, no parity-check.

I'm testing this mobo totally bare on a benchtop. It has no onboard ports other than keyboard and speaker. All other IO requires ISA cards and I haven't tried because I figure if I can't see a BIOS or video power-on screen, I'm going nowhere anyway.

I rigged a speaker to the SPKR pins. On each boot or reset I get the same beep pattern: 3 x 1-second beeps at 5-second intervals, followed by 10-second interval, then repeat. This happens whether memory SIMMs are present in ANY combination or none at all.

AMIBIOS beep code standard suggests 3 x "short" beeps indicates failure in lowest 64K of RAM.
I get the same beep result with no SIMM in slot or any combination of the 8, so I think this means the failure is in the RAM addressing rather than any RAM module.

So new questions:

Does RAM address failure indicate motherboard chipset failure, or could it be somewhere else?

What does the consistent beep output say about whether the CPU is OK, or about the motherboard chipset?

Are there particular ISA bus signals I can check to point toward what is blocking RAM access?

Is the 256k cache RAM (20ns) a possible problem at boot, or does it only function after BIOS load?

Rick
 
I am not sure but it could be a system that only works with parity enabled memory. That is what TULARC lists for the Informtech systems shows. And would match the symptoms. Is there any chance you have 4 SIMMS of parity memory available to test?
 
I see what you mean - the Informtech 486VL is shown in TULARC with [8] 1Mb x 9 modules for an 8Mb RAM setup.

This would really surprise me, since the non-parity modules that came installed on this board carried the same workshop warranty stickers as the CPU and chipset in the board - and someone had gone to the trouble of filling all slots and cache sockets. It's possible that my board, under the different "US Technologies brand", was set up for non-parity RAM. I suppose its also possible that someone tried to stick together a bunch of left-over components from a workshop cleanup, and then tossed the lot when it didn't work.

Mean time I'd really like to know what else can be deduced from the fact that the BIOS beep signal is reliably generated. If I have some assurance that main chipset is probably OK, I'll keep trying, but if main chipset is blown, I won't take this further...

Rick
 
Last edited:
OK, I found some 1mb parity RAM sticks in a working 386 machine of mine, and tried them on this 486 board. Same beeps, so problem seems to be upstream of RAM array, and non-parity RAM is not the problem.

I did some research on POST cards that give a HEX readout of the progress byte that BIOS POST sends to port 80. Experienced techs will have used these but as an amateur I did not know they existed, though I have made ROM devices with similar function (serial output) for my S100 system.

All but the most ancient have both ISA and PCI edge connectors, for use in both bus types. I found one advertised at a "clearance" price of $50 + $20 postage from a US supplier, and I found several new products in China for $6 including postage. Guess which one I am waiting for as it comes down the Pearl River in a sampan. I'll shelve the 486 board now until I get that POST card, but I think the chances are rising that the mobo is faulty.

While searching online, I saw 486 motherboards offered for $50-200 + expensive freight to Australia. I remember the day, about five years ago, I dumped my last working one in the recycle bin. In the IT product cycle, rarity comes quickly!
 
So what happened?

I just ran into a very similar problem today - except it was on a working 486 system. :( It was working - came back to it and no display and the system was locked up. I thought it was the DPMS mode just fluked. Gave it a hard reset and... no display and no boot.

Mine is a BEK 486 mobo. Can't recall the exact model number right now but I suspect it isn't going to be relevant, since it was a functional system at the time it crapped. Unlike the OP I am getting no POST beeps at all. The only thing I can see working is: Power on, all three keyboard LED's come on in unison, go off... and that's it. I have to assume that means the POST is getting at least past the keyboard BIOS.

I have stripped it down to a bare system - only video card, cpu and memory. Reseated the BIOS, cache RAM CPU & memory - no joy. Power supply voltages are testing ok at the motherboard connectors. Also, I do have one of those China $6 diagnostic cards. It confirms voltages are getting to the mobo, but it's not displaying any code - error or otherwise - on the 2 digit display. Is the chip set stuffed? (Or the China diagnostic card could just be crap too.)

The only other thing I can try is swapping out the cpu. I should have another DX2-66 processor kicking around. But the complete lack of any POST beep, or message on that diagnostic card, doesn't sound like it's a cpu problem.
 
Congratulations on finding a thread this old!

Sadly, I was forced to conclude I had a failure in the chipset, beyond my skill and scope to repair.

Got a tested Asus 486 board with DX2-50 CPU locally (Australia) for $50. It's configurable for 16-75Mhz 486 clock speeds, but the DX-33 CPU from my original found board won't run, so I now assume dead CPU. Neither will the found board run with the known good DX-2 50Mhz CPU. So probably there is damage to both CPU and chipset (big power spike?) and recovery unlikely.

A further lesson on buyer beware - the $6 POST card from China was a failure in this context. In either ISA or PCI bus slots, it showed power rail status, and in PCI it showed clock status. But tried in several boards, ISA and PCI, it never showed BIOS POST progress codes, and also locks up the bus. I'm assuming bad code programmed into the ROM. Only consolation is that the 4-digit 7-segment LED display is worth more than I paid for the card, so I'll keep the thing to raid for components.

If anyone knows of any tricks with these POST cards, I'd be grateful to hear. This one is poorly documented. It has an instantaneous switch that I guessed was a reset button, but didn't seem to do anything.

Rick
 
RickNel said:
Congratulations on finding a thread this old!

Sadly, I was forced to conclude I had a failure in the chipset, beyond my skill and scope to repair.
Click to expand...

Oh, desperate people can find anything with google.

Grrrr - the chipset issue. Is that the conclusion we have to come to when nothing else will get the mobo to work? Funny(?) the subtle difference I ran into between two failed 486 VL mobo's: Same scenario, both were dead with no POST and no video. On the first one when power was applied all three keyboard LED's went on and stayed on. On my current fail, they come on and go off.

Got a tested Asus 486 board with DX2-50 CPU locally (Australia) for $50. It's configurable for 16-75Mhz 486 clock speeds, but the DX-33 CPU from my original found board won't run, so I now assume dead CPU. Neither will the found board run with the known good DX-2 50Mhz CPU. So probably there is damage to both CPU and chipset (big power spike?) and recovery unlikely.
Click to expand...

I trust you changed the clock divider jumpers when swapping the cpu's between the two boards? That can cause seizing up depending on how far off the clock is running versus the cpu's speed rating.

A further lesson on buyer beware - the $6 POST card from China was a failure in this context. In either ISA or PCI bus slots, it showed power rail status, and in PCI it showed clock status. But tried in several boards, ISA and PCI, it never showed BIOS POST progress codes, and also locks up the bus. I'm assuming bad code programmed into the ROM. Only consolation is that the 4-digit 7-segment LED display is worth more than I paid for the card, so I'll keep the thing to raid for components.

If anyone knows of any tricks with these POST cards, I'd be grateful to hear. This one is poorly documented. It has an instantaneous switch that I guessed was a reset button, but didn't seem to do anything.

Rick
Click to expand...

Basically, that's all a "me too" regarding the diagnostic card. It did show power status, but that was it.

The only - really important - thing the lousy documentation skips over is the card orientation in an ISA slot. The card has to be inserted "components up", with the display facing the same direction as the components would on any other ISA card inserted in a slot. My documentation tries to make a pass at telling you "no harms come by your card" if you put it in the wrong way, but I wouldn't count on that assurance.

Us technical people would have appreciated a simple pin 1 notation on the card silkscreen itself...

Now for the hunt to find a replacement mobo.

Rick (Not that Rick, the other Rick.)
 
Rickajho said:
I trust you changed the clock divider jumpers when swapping the cpu's between the two boards? That can cause seizing up depending on how far off the clock is running versus the cpu's speed rating.
Click to expand...

Oh yes, I was careful with the jumper settings, reluctant to decide it was time to turn off life support for the mobo...
 
You must log in or register to reply here.
Back
Top