VOGONS


Reply 40 of 77, by Doornkaat

User metadata
Rank Member
Rank
Member

Hi! just quickly on the measuring: You can try to measure shorts with the multimeter first. The four pins in the middle are ground, all of the other pins should have very high resistance against ground. Pull the connectors off and measure, if one measures low there's a short on the mainboard, often a bad capacitor.
To measure voltages safely while the system is turned on you can jam a bent up paperclip into the AT power connector next to a cable after plugging it into the mainboard and use any ground as reference. Just make sure it makes contact inside the connector.
For help measuring the transistor I'd need a better pic or simply the text on the transistor so I can look up the datasheet.
Measuring on the socket is a little trickier but basically you flip the board over and probe the socket pins sticking through the PCB. If you don't trust the steadiness of your hand the transistor measurement should already be enough to determine wether the CPU is getting power. So we'll leave that fun for later if the other measurements are inconclusive. 😉

Reply 41 of 77, by HandOfFate

User metadata
Rank Newbie
Rank
Newbie
Deunan wrote on 2020-02-06, 17:32:

Just don't short anything and you'll be fine. If you aren't trusting your hands then connect the probes first, then power the system on.
Also, not every PC PSU will start or output correct voltages with no load - and that goes especially for the older ones.

I measured the voltages just now. They're all as they should be, expect for the blue -12V that measured at -9V. Could that be because it wasn't under load?

I wish I saw Doornkaat's paperclip suggestion earlier, because I had the hardest time getting the probes to connect 😜

Doornkaat wrote on 2020-02-06, 19:09:
Hi! just quickly on the measuring: You can try to measure shorts with the multimeter first. The four pins in the middle are grou […]
Show full quote

Hi! just quickly on the measuring: You can try to measure shorts with the multimeter first. The four pins in the middle are ground, all of the other pins should have very high resistance against ground. Pull the connectors off and measure, if one measures low there's a short on the mainboard, often a bad capacitor.
To measure voltages safely while the system is turned on you can jam a bent up paperclip into the AT power connector next to a cable after plugging it into the mainboard and use any ground as reference. Just make sure it makes contact inside the connector.
For help measuring the transistor I'd need a better pic or simply the text on the transistor so I can look up the datasheet.
Measuring on the socket is a little trickier but basically you flip the board over and probe the socket pins sticking through the PCB. If you don't trust the steadiness of your hand the transistor measurement should already be enough to determine wether the CPU is getting power. So we'll leave that fun for later if the other measurements are inconclusive. 😉

When you mean pull the connectors off, do you mean measuring one of the plastic connectors (e.g. P8) while the other one is still on the board, or pulling individual wires from the connector while the rest stay connected?

I will try to look for the transistor and capture it on camera, or at least write down the markings.

Edit: the transistor is a SMP50N06-25

Reply 42 of 77, by Doornkaat

User metadata
Rank Member
Rank
Member

I didn't mean to pull the connectors apart. Just disconnect it from the motherboard to test the board (not the PSU) for shorts.

You tested the voltages with the motherboard connected to the PSU, right? Just checking.😅

-9V instead of -12V sounds a bit low for my taste. This could prevent the power good signal to be sent depending on your PSU. You might be right though and it's simply because there's no load on -12V.

You can measure the output voltage of the transistor at the screw that holds it down on the motherboard, [EDIT: sorry, that was wrong. Deunan has posted the right description.] reference it to a ground point on the motherboard. If it outputs the desired voltage (that you selected with the jumpers) it's fine.

Last edited by Doornkaat on 2020-02-07, 12:25. Edited 2 times in total.

Reply 43 of 77, by Deunan

User metadata
Rank Oldbie
Rank
Oldbie

Pretty much the only thing -12V is used for is some VGA DACs and RS232 ports. And in both cases -9V would still work. That particular line is known to have horrible regulation though granted, something closer to 12 would be preferable.

On SMP50N06-25 tab is connected with drain and I'd expect it at 5V in typical configuration. Source pin should be at the correct output voltage (3V3 for newer 486), and gate somewhere between 1 an 5 volts, depends on partictular transistor.

Reply 44 of 77, by HandOfFate

User metadata
Rank Newbie
Rank
Newbie
Doornkaat wrote on 2020-02-06, 23:04:

I didn't mean to pull the connectors apart. Just disconnect it from the motherboard to test the board (not the PSU) for shorts.

Ah, I completely misunderstood you there.

You tested the voltages with the motherboard connected to the PSU, right? Just checking.😅

Ah, no, I tested the PSU directly.

You can measure the output voltage of the transistor at the screw that holds it down on the motherboard, reference it to a ground point on the motherboard. If it outputs the desired voltage (that you selected with the jumpers) it's fine.

Deunan wrote on 2020-02-07, 11:41:

On SMP50N06-25 tab is connected with drain and I'd expect it at 5V in typical configuration. Source pin should be at the correct output voltage (3V3 for newer 486), and gate somewhere between 1 an 5 volts, depends on partictular transistor.

I will measure all three things (resistance, voltage when connected to board and transistor output voltage) tonight 😀

Reply 45 of 77, by Doornkaat

User metadata
Rank Member
Rank
Member
Deunan wrote on 2020-02-07, 11:41:

On SMP50N06-25 tab is connected with drain and I'd expect it at 5V in typical configuration. Source pin should be at the correct output voltage (3V3 for newer 486), and gate somewhere between 1 an 5 volts, depends on partictular transistor.

Absolutely. Thanks for pointing out my mistake and thank you as well for not pointing fingers!👍

Reply 46 of 77, by HandOfFate

User metadata
Rank Newbie
Rank
Newbie

Alright, so "tonight" became "this morning", sorry for that. But I measured everything:

  1. The source pin measures at 3.46V.
  2. The voltages from the PSU were all the same, except for the -12V that now measured at -10.9V
  3. The resistance on all the 5V pins was about 330 Ohms. The rest were "very high". 330 doesn't sound high to me. Could this be something?

Reply 47 of 77, by Horun

User metadata
Rank Oldbie
Rank
Oldbie
HandOfFate wrote on 2020-02-08, 11:32:

The resistance on all the 5V pins was about 330 Ohms. The rest were "very high". 330 doesn't sound high to me. Could this be something?

That sounds about right for some boards, specially if you have ram installed. Did you try 5v to ground with Ohm meter connected both ways ? One way it should read somewhat low, the other should start fairly high and slowly drop as the caps get reverse charged (or whatever you want to call it 😁)

Hate posting a reply and have to edit it because it made no sense 😁 First computer was an IBM 3270 workstation with CGA monitor. 🤣

Reply 48 of 77, by Deunan

User metadata
Rank Oldbie
Rank
Oldbie

Voltages are good. I take it that 330 ohm is mobo resistance from 5V to GND? That's reasonable as well, in fact a bit high but you can't really measure it properly with a meter like that - it's only good enough to test for obvious shorts. And you don't have any. If the PSU is able to maintain 5V on it's output and nothing is glowing hot/smoking, it's good 😀

If you have nothing better to do, inspect the mobo again around the battery, see if there isn't any leak damage that might have gone unnoticed. Re-seat the BIOS ROM and CPU. Try different memory sticks / in different slots. I've seen quite a few 486 mobos that would appear dead just because they didn't like the memory - there were no beeps like you'd normally expect in that case.

Reply 49 of 77, by HandOfFate

User metadata
Rank Newbie
Rank
Newbie
Deunan wrote on 2020-02-08, 18:27:

Voltages are good. I take it that 330 ohm is mobo resistance from 5V to GND?

Ha, you're right of course. I should've measured it the other way around. I've got some ways to go with learning about this electronics stuff.

If you have nothing better to do, inspect the mobo again around the battery, see if there isn't any leak damage that might have gone unnoticed. Re-seat the BIOS ROM and CPU. Try different memory sticks / in different slots. I've seen quite a few 486 mobos that would appear dead just because they didn't like the memory - there were no beeps like you'd normally expect in that case.

There is no battery damage as far as I can see. The Dallas chip that I removed also wasn't damaged on the outside.

There are some scratch marks near the Dallas that I caused during the removal process, but I still measure continuity on all those lines (btw, these are the lines that are going from the Dallas chip to the missing external keyboard controller chip)

I reseated the CPU and BIOS (which wasn't easy, as if I was the first one to do it in 25 years). I also tried all the EDO RAM that I have. I don't have that many compared to my SDRAM collection, but I don't come across them too often.

No improvement, unfortunately.

Aside: Next to the missing floppy drive connector I also noticed another weird thing about this board: there is no power LED. Maybe the turbo LED doubled as a power indicator?

Edit: I've contacted BiosAgentPlus (the company that sells BIOS updates, which Wim's BIOS site sends you to) to confirm if they actually have the BIOS for this board. This page says "yes", their search engine says "no". I don't like the idea of having to pay for it at all but it seems to be the only way of getting it (barring any lucky find of another contemporary ISA/PCI 486 board)

Reply 50 of 77, by Deunan

User metadata
Rank Oldbie
Rank
Oldbie
HandOfFate wrote on 2020-02-09, 13:05:

I also tried all the EDO RAM that I have.

I'd suggest sticking to FPM RAM, aka "normal" non-EDO. Frankly not that many 486 mobos support EDO and those that do are more often than not very picky about it anyway.

EDIT: Seems like your mobo can accept a single 4MB stick in SIMM1 slot as a bare minimum, and that's what I'd try to get working first.

Reply 51 of 77, by HandOfFate

User metadata
Rank Newbie
Rank
Newbie

Interesting, it looks like I don't have any FPM RAM at all. My other 486 apparently runs on EDO RAM (or 30-pin SIMMs)

I will have to look around and see where I can get some.

Edit: I found someone close by who is selling two sticks of 4MB. They seem to be FPM.

Edit II: I will pick them up tonight 😀

Edit III: Too bad, it doesn't boot with this memory either. They are working in my other 486.

Reply 52 of 77, by HandOfFate

User metadata
Rank Newbie
Rank
Newbie

I just found this 486 BIOS archive topic here on Vogons.

One of the boards in it looks very similar to mine (slots, layout, Dallas battery, identical UMC chipset) except that it used a AmiBIOS instead of Award. But the chip looks to be compatible (that board: M27C1001, mine: L27C010X), because it has the same pinout (or is that comparison too simple?)

Would it be worth to get a programmer that can flash (E)EPROMs and try to flash it on a chip?

Reply 53 of 77, by Deunan

User metadata
Rank Oldbie
Rank
Oldbie

Worth? Not really, except maybe the TL866 if you can get it cheap. Kinda depends on how much are you willing to spend on this mobo and/or if the programmer can be useful to you in future projects.
In general repairing PC motherboards is difficult, it can be done with early XTs using just basic tools but 286/386 ones require scope and some experience. PCI-based 486 ones... not saying it can't be done, but that's a "hobby" on it's own 😀

In other words, are you interested in building/using retro PCs or repairing electronics - make your choice based on that. Also, I'd wait for that ISA-based POST card before making any more purchases.

Reply 54 of 77, by HandOfFate

User metadata
Rank Newbie
Rank
Newbie

But you're saying there is a chance this could work? 😜

I'm sure a good programmer (I've heard the TL866 mentioned before) will come in handy. The current one I have now is "okay" (but cheap) for programming small ICs but a better one couldn't hurt.

You're right about waiting for that ISA card. No update in it's shipping status, so still have to be patient.

Reply 56 of 77, by Deunan

User metadata
Rank Oldbie
Rank
Oldbie
HandOfFate wrote on 2020-02-12, 20:10:

I'm sure a good programmer (I've heard the TL866 mentioned before) will come in handy.

I wouldn't call TL866 "good" but if the chip you need is supported then it will do it's job and you can't argue with the price. And being able to at least read the exising ROM to make sure it's not corrupted is nice too, one less thing to worry about.

Reply 57 of 77, by HandOfFate

User metadata
Rank Newbie
Rank
Newbie
Zuupertronic wrote on 2020-02-12, 21:27:

Could always try and see if you could get another post card off of ebay of that one from china is taking too long.

That's true but this "no manual" thing left me a bit cautious of Chinese stuff from eBay.

At least with AliExpress you have product reviews (sometimes with photos) so you have some an idea what's included. With eBay you only have the reviews of the seller as a whole and the product description. In this case, the description implied there was a manual: "watch for the visual and beep code and reference the error code with the manual."

At least the eBay seller offered for me to send it back and reimburse the postage. Not going to do it as it's too much hassle for a few Euro, but it's a nice offer.

Reply 58 of 77, by HandOfFate

User metadata
Rank Newbie
Rank
Newbie

Well, the card from AliExpress finally arrived. PCI + ISA and the manual was included. Amazing!

However, no good news on the part of diagnosis this motherboard.

Both when plugged into a PCI or ISA socket the POST card shows '----' (4 dashes) and a few LEDs come on, the LEDs for the +/- 5V and 12V power and the one for the clock. The clock LED stays on instead of blinking.

I've tried to get the card to show something else. But even when the CPU is removed it shows '----'. I will check (again) if my CPU jumpers are set correctly.

Reply 59 of 77, by CoffeeOne

User metadata
Rank Member
Rank
Member
HandOfFate wrote on 2020-02-22, 16:15:
Well, the card from AliExpress finally arrived. PCI + ISA and the manual was included. Amazing! […]
Show full quote

Well, the card from AliExpress finally arrived. PCI + ISA and the manual was included. Amazing!

However, no good news on the part of diagnosis this motherboard.

Both when plugged into a PCI or ISA socket the POST card shows '----' (4 dashes) and a few LEDs come on, the LEDs for the +/- 5V and 12V power and the one for the clock. The clock LED stays on instead of blinking.

I've tried to get the card to show something else. But even when the CPU is removed it shows '----'. I will check (again) if my CPU jumpers are set correctly.

Hello,

Sorry to repeat myself, maybe I missed something?
Did you test always with the same CPU?
If yes, then test at least one time with a 5volts cpu (DX33, overdrive, early intel dx2/66 or whatever), that should not be that hard to get.