You need 1 and 3.

derSammler wrote on 2020-01-25, 18:15:

You need 1 and 3.

pretty sure that's wrong - speaker data is referenced to positive voltage so he would need 1 and 4

On this agree with max, nearly all the 4 pin speaker connectors use the two outside pins (1 and 4). Your board being a standard AT 486 board should also be that way.

Thanks for the replies 😀

maxtherabbit wrote on 2020-01-25, 18:23:

derSammler wrote on 2020-01-25, 18:15:

You need 1 and 3.

pretty sure that's wrong - speaker data is referenced to positive voltage so he would need 1 and 4

I've tried connecting a normal PC speaker and didn't get any sound. Of course this could mean the board is just dead, but that's when I found out this connector is (for me) unusual.

I will try to connect 1 and 3. If that also doesn't work I guess it means it's completely dead.

maxtherabbit wrote on 2020-01-25, 18:23:

pretty sure that's wrong - speaker data is referenced to positive voltage so he would need 1 and 4

If speaker data is positive voltage, why would he need that and +5V?

Horun wrote on 2020-01-26, 04:37:

On this agree with max, nearly all the 4 pin speaker connectors use the two outside pins (1 and 4). Your board being a standard AT 486 board should also be that way.

Standard for 4-pin speaker header is:

Pin 1: VCC
Pin 2: NC
Pin 3: NC
Pin 4: Data(-)

VCC should be on pin 1, not 4 as with his pin-out - so it is apparently not standard.

Adapting to the pin-out he posted, you end up with pin 1 and 3, which is GND and Data(+) in this case. But he can try all possible combinations, as nothing bad can happen anyway.

That makes sense. I will try it out (hopefully later today) and report back.

derSammler wrote on 2020-01-26, 10:58:

maxtherabbit wrote on 2020-01-25, 18:23:

pretty sure that's wrong - speaker data is referenced to positive voltage so he would need 1 and 4

If speaker data is positive voltage, why would he need that and +5V?

someone clearly doesn't understand what the phrase "referenced to" means

I've tried both options but without success, unfortunately. I'm getting no sound or signs of life on an oscilloscope, except for the voltage coming from the Dallas battery.

I'm guessing I should give up on this board. Unless I can somehow verify the BIOS chip? (I do have a chip programmer)

well you could try to find the bios and rewrite it, that's always worth a try imho.

I'm having no luck finding a BIOS update for the motherboard, unfortunately. Wim's BIOS site claims that that paid service he links to has it, but it doesn't show up in the search there. Maybe I'll be lucky someday and come across someone who has the same board and revision.

Thanks for the help with my question!

I had two 486 mobos that didn't POST or beep when there was something wrong with the RAM sticks. So having a PC speaker working is not exactly a great help in diagnosing issues. A POST card will serve you much better.
Since you have a scope, probe the ISA slot signals - reset, clocks, address/data lines for any indication of activity. Or the BIOS chip /CE and /OE lines. Could be the CPU is dead or somehow not configured properly.

HandOfFate wrote on 2020-01-29, 20:59:

I'm having no luck finding a BIOS update for the motherboard, unfortunately. Wim's BIOS site claims that that paid service he links to has it, but it doesn't show up in the search there. Maybe I'll be lucky someday and come across someone who has the same board and revision.

Thanks for the help with my question!

Can you take a good picture of the whole board and post it ? Many times a board was used in a few different computers and someone here may have that same board like from Leading Technology, Laser (all Vtech brands) or some other brand. The manual is good for jumpers but bad for identifying specifics...

Probing the Speaker header with an oscilloscope has a couple of gotchas.

First, the data pin is open collector, so it's either in a high-impedance state, or pulled low with a transistor. You will not see a signal on it with an oscilloscope unless you bias it up with a resistor.

The speaker's impedance is 8ohm, so a 10ohm resistor should suffice.

You can bias it using the speaker itself, but back EMF from the speaker will distort the waveform.

Second, the power supply inside a desktop PC is referenced to mains earth, and so are your scope probes. If you you attach probe ground to any voltage rail (in the case you want to reference your readings to, say, +5v, as would be the case if you simply connected the probe across the speaker terminals), you are effectively shorting that rail to ground, and will at best trip the overcurrent protection in the PSU, at worst melt your wires and/or start a fire.

I've tried contacting PCPartner to see if they still have the BIOS (since they still offer downloads for their Socket 7 motherboards), but alas. Props to them for responding very quickly to my question though.

Deunan wrote on 2020-01-29, 22:18:

I had two 486 mobos that didn't POST or beep when there was something wrong with the RAM sticks. So having a PC speaker working is not exactly a great help in diagnosing issues. A POST card will serve you much better.
Since you have a scope, probe the ISA slot signals - reset, clocks, address/data lines for any indication of activity. Or the BIOS chip /CE and /OE lines. Could be the CPU is dead or somehow not configured properly.

The CPU does work in another board. It's a AMD 5x86 and the board I'm trying to revive is newer than the other one, so I think it should support it.

There is a question of the jumper settings: the manual that can be found online for the M486-UPI is missing some pages in the CPU jumper settings chapter. I'm using the one for the Enhanced AMD 486 DX4 now (with a System Clock Speed of 33MHz) but there might be a specific setting for 5x86 on the missing pages.

I ordered a POST BIOS card from China last week but who knows how long it will take to arrive, so I will try to measure the lines you mentioned. What kind of signals should I be seeing?

Horun wrote on 2020-01-30, 02:16:

Can you take a good picture of the whole board and post it ? Many times a board was used in a few different computers and someone here may have that same board like from Leading Technology, Laser (all Vtech brands) or some other brand. The manual is good for jumpers but bad for identifying specifics...

Yes, I will take one tonight. There is a good chance it came from a Laser machine since those were heavily advertised here in the 90s.

SquallStrife wrote on 2020-01-30, 02:18:

Probing the Speaker header with an oscilloscope has a couple of gotchas. […]
Show full quote

Probing the Speaker header with an oscilloscope has a couple of gotchas.

First, the data pin is open collector, so it's either in a high-impedance state, or pulled low with a transistor. You will not see a signal on it with an oscilloscope unless you bias it up with a resistor.

The speaker's impedance is 8ohm, so a 10ohm resistor should suffice.

You can bias it using the speaker itself, but back EMF from the speaker will distort the waveform.

Second, the power supply inside a desktop PC is referenced to mains earth, and so are your scope probes. If you you attach probe ground to any voltage rail (in the case you want to reference your readings to, say, +5v, as would be the case if you simply connected the probe across the speaker terminals), you are effectively shorting that rail to ground, and will at best trip the overcurrent protection in the PSU, at worst melt your wires and/or start a fire.

For some reason I have trouble grasping grounding (I'm very new to electronics). I don't have the board in a case at the moment, so can I attach the probe to one of the screw holes in the corners of the board instead?

I have a bunch of resistors so I'm sure I have a 10ohm one, so I will try it tonight as well.

HandOfFate wrote on 2020-01-30, 10:33:

I ordered a POST BIOS card from China last week but who knows how long it will take to arrive, so I will try to measure the lines you mentioned. What kind of signals should I be seeing?

Start with BIOS chip. One of the /CE (/CS) and /OE lines might be connected directly to ground so if that's the case, ignore it, look at the other one. It's an active low (negated) signal so it'll be at close to 5V when idle and close to GND when active. You should see it toggling a lot, if perhaps briefly, shortly after power is applied. Might need to set the scope to normal (rather than auto) triggering to catch that.
No activity - CPU is not running (misconfigured or reset signal is not letting go for some reason). Or the mobo chipset is toast but that's usually not the case.

HandOfFate wrote on 2020-01-30, 10:33:

For some reason I have trouble grasping grounding (I'm very new to electronics). I don't have the board in a case at the moment, so can I attach the probe to one of the screw holes in the corners of the board instead?

You can, just make sure first the plating around the hole is actually connected to GND pins on the PSU connector. In general for now always stick to the rule that the scope ground must be connected to target ground, and do not attempt to measure anything in switching power supplies unless you know exactly what you are doing. Do not break these rules until you are confident you know your stuff (and even then you'll probably know better than to stick scope probes in PSUs unless there is no other choice).

Horun wrote on 2020-01-30, 02:16:

HandOfFate wrote on 2020-01-29, 20:59:

I'm having no luck finding a BIOS update for the motherboard, unfortunately. Wim's BIOS site claims that that paid service he links to has it, but it doesn't show up in the search there. Maybe I'll be lucky someday and come across someone who has the same board and revision.

Thanks for the help with my question!

Can you take a good picture of the whole board and post it ? Many times a board was used in a few different computers and someone here may have that same board like from Leading Technology, Laser (all Vtech brands) or some other brand. The manual is good for jumpers but bad for identifying specifics...

Here are the photos: one overview and two detail shots of some markings.

U30 socket (left bottom side) is empty - did you remove that chip for some reason before taking the photo?
Typically this is where keyboard controller is. If it's missing the mobo might not boot at all, these chips provide (as a secondary function) control over the CPU reset line. Might very well be why it appears dead.

Good eye Deunan ! Was going to mention that. Some of the UMC chipsets has on chip KB controllers but they also had on chip CMOS, which this one does not since it has a Dallas RTC. One other thing that caught my eye is the damage at the PSU connector. The Yellow lead is +12 and the blue lead is the -12v, both are required on old XT and AT boards. I would check to make sure the board is actually getting those voltages. EDIT: +12v is required on all boards 🤣

Horun wrote on 2020-01-31, 02:12:

Good eye Deunan ! Was going to mention that. Some of the UMC chipsets has on chip KB controllers but they also had on chip CMOS, which this one does not since it has a Dallas RTC. One other thing that caught my eye is the damage at the PSU connector. The Yellow lead is +12 and the blue lead is the -12v, both are required on old XT and AT boards. I would check to make sure the board is actually getting those voltages. EDIT: +12v is required on all boards 🤣

That isn't damage, that is bit of carton to keep the metal clip of the EDO RAM in place. It's a bit weak and couldn't always keep the RAM stick in.

Deunan wrote on 2020-01-30, 23:49:

U30 socket (left bottom side) is empty - did you remove that chip for some reason before taking the photo?
Typically this is where keyboard controller is. If it's missing the mobo might not boot at all, these chips provide (as a secondary function) control over the CPU reset line. Might very well be why it appears dead.

It was always missing. But there are some jumpers for setting the keyboard controller to "internal" and "external". As far as I can tell they're set to "internal". JP31 and JP16, which you can see just top-left of the BIOS chip, are open which according to the manual mean the internal controller should be used. My other 486 motherboard also has an empty socket there, so I didn't think anything about it.

There are indeed lines going from the Dallas chip to that socket. Should I be looking for a replacement chip?

I didn't succeed in measuring the BIOS chip btw. I'm trying to take the shiny Award label off without damaging it too much so I can check the type of chip (for the datasheet). Because AFAIK there wasn't a single chip that was always used for the (presumably 128kb) BIOSes?