Can you put the eprom in reader to see if it still opereates?

One step ahead of you. Here is the BIOS. I can read the PLD if needed...

I'm thinking it could be the HM6818 RTC/SRAM that may have gone bad. As I'm fairly certain that BIOS settings would be set there. I've ordered a few from China, but they'll take a few months to get here I'm sure.

jdgabard wrote on 2020-12-20, 23:36:

So I'm attempting to repair a 386sx MB that I received a little while back. The board is identical to this board on Stason.org, […]
Show full quote

So I'm attempting to repair a 386sx MB that I received a little while back. The board is identical to this board on Stason.org, although without markings it's hard to say for sure: https://stason.org/TULARC/pc/motherboards/M/M … 386-A-B307.html

This board came with a Varta battery installed, and there was a decent amount of corrosion around the keyboard connector. I cleaned and neutralized it, and then looked for damage. While there was some corrosion, no traces were broken, at least as far as I can see... The only other damage I could see around the board was a small amount of corrosion around the VL82C311 IC (far opposite side of the board strange enough). Several of the IC legs had come unsoldered, and I repaired these and tested for connectivity to the nearby vias. The repair appears to have been successful.

The motherboard is not completely dead. Upon boot I get 10 short beeps, AMI beep code: CMOS shutdown register read/write error.

But I'm not entirely sure what this beep code is referring to. Now, knowing quite a bit about early computer design, I am making an assumption that this is either a hardware register within the 82C311, or they're using one of the 74LS373 octal latches as a register and one or the other is not functioning properly. But before I decided to start desoldering ICs, I wanted to get some input from others.

Anyone got any suggestions?

Yes MAGITRONICS-386 looks like the same board, also Deksor has a variant (same chipset, and BIOS referring to SCAMPSX, just slightly diff layout):
Re: 80386 BIOS image collection
Looking at the VL82C311 near the silk screened pin number 70 something does not look good. Can you take a close up of that area ?

That picture is about the best I can do with my camera. But I've tested it, and it is soldered securely, and shows continuity between it and the 386.

jdgabard wrote on 2020-12-21, 01:39:

That picture is about the best I can do with my camera. But I've tested it, and it is soldered securely, and shows continuity between it and the 386.

Ok, pin 65 right off the IC looks odd, is not like a SMT pin and is very dull. It could be broken inside the edge of the plastic IC casing.
I actually thought if you backed your camera away a bit it may focus on that area a little better...

It's a tiny bit of solder that bonded to the leg while reworking it. I was doing it by hand, so it wasn't the prettiest job. But it worked.

jdgabard wrote on 2020-12-20, 23:36:

So I'm attempting to repair a 386sx MB that I received a little while back. The board is identical to this board on Stason.org, […]
Show full quote

So I'm attempting to repair a 386sx MB that I received a little while back. The board is identical to this board on Stason.org, although without markings it's hard to say for sure: https://stason.org/TULARC/pc/motherboards/M/M … 386-A-B307.html

This board came with a Varta battery installed, and there was a decent amount of corrosion around the keyboard connector. I cleaned and neutralized it, and then looked for damage. While there was some corrosion, no traces were broken, at least as far as I can see... The only other damage I could see around the board was a small amount of corrosion around the VL82C311 IC (far opposite side of the board strange enough). Several of the IC legs had come unsoldered, and I repaired these and tested for connectivity to the nearby vias. The repair appears to have been successful.

The motherboard is not completely dead. Upon boot I get 10 short beeps, AMI beep code: CMOS shutdown register read/write error.

But I'm not entirely sure what this beep code is referring to. Now, knowing quite a bit about early computer design, I am making an assumption that this is either a hardware register within the 82C311, or they're using one of the 74LS373 octal latches as a register and one or the other is not functioning properly. But before I decided to start desoldering ICs, I wanted to get some input from others.

Anyone got any suggestions?

Hi, I have same issue. It is the first time I searched for something like that, i.d. the *.BIN file!!!
I have a AWARD BIOS 4.50G for my 486 motherboard. I have to erase then to re-program the EEPROM, but...
WHERE could I get the AWARD *.BIN file???
🙁
I thought it was easy to find, but I cannot find a download site... It was a very common BIOS...

Thank you in advance,
cheers from Italy

I got stuck on the same error with a 486 board a while ago, still stuck there. On the board I was working on the cmos memory is embedded into a qfc chip. Naturally I assumed it went corrupt and replaced the qfc chip but still same issue.

There is not much info about this error but form my research the cpu can't be out into the correct mode to boot because the reading is corrupt or saved wrong inside the cmos.

So cmos memory could be corrupt, or the signal is not getting through to the cpu.

Think the logic goes like this:

1. Power on
2. Bios does startup checks
3. Bios attempts to switch cpu into protected mode
4. Bios - > northbridge - > cpu
5. Bios reads the cpu register bit from cmos as not being set into protected mode.

So Bios, cmos, all the lines going to northbridge from cmos, bios and cpu are in question.

I never fixed mine most likely I have a crack in one of the many lines somewhere

Here is my thread may help you

https://www.vogons.org/viewtopic.php?t=73084

https://www.vogons.org/viewtopic.php?f=46&t=74620

treeman wrote on 2020-12-21, 02:50:

I got stuck on the same error with a 486 board a while ago, still stuck there.
cmos shutdown register read/write error

Thanks ! I remember that time. My FIC board still has an error also, it sits in the "waiting to repair pile". Am thinking one of the hidden via's is corroded or just not conducting proper.
Have fixed a few others since then but not that one 🙁

Well, I did consider the CMOS memory. The only thing I can think of that is battery backed is the 6818 RTC/SRAM. I’m pretty sure these are compatible with the 146818, so I ordered a few to see if I could get it working that way. I suppose it could be the 82C311 is just dead. Which would be a shame... But that doesn’t make sense logically since it can read enough of the bios to know there is an error and throw the appropriate beep code.

The only other thing I can think of is possibly the 6818 isn’t getting a clock signal. Guess it’s time to break out the oscilloscope and check...

Well, interesting results from the oscilloscope....

Pin 1 - MOT: 32khz square wave at 50mv (way out of spec for TTL logic)
Pin 2 & 3 - Oscillator: 32khz square wave at 5.5v
Pin 23 - SQW: No activity, logic low.
Pin 22 - PS: No activity, logic high.
Pin 21 - CKOUT: 32khz square wave at 5.5v
Pin 20 - CKFS: No activity, logic high.
Pin 19 - /IRQ: No activity, logic high
Pin 18 - /RESET: No activity, logic high
Pin 17 - DS: No activity, logic high
Pin 16 - STBY: No activity, logic high
Pin 15 - R/W: No activity, logic high
Pin 14 - AS: Open, just ambient noise.
Pin 13 - /CS: Open, just ambient noise.

With this in mind, my immediate suspect was the /CS, which is the chip select line. Following the trace back, it goes underneath the Keyboard controller to a TC4069 (equivilent of the CD4069, which is a 6x NOT-gate off the shelf logic chip). However, this is where it is interesting. It connects to pins 1 and 2 of the TC4069. Now, there is a reason I mention this, and it's because those are both the input and output of the NOT-gate. I can think of absolutely zero reasons to connect it up like this. As a logic 1 on this line would short the output to ground, and either burn out the TC4069, and a logic 0 would flip the bit turning output high and negating the input. I'm seriously wondering if this motherboard EVER WORKED given what I just discovered. It's almost as if they installed the incorrect IC from the factory....

jdgabard wrote on 2020-12-21, 05:00:

Well, interesting results from the oscilloscope.... […]
Show full quote

Well, interesting results from the oscilloscope....

Pin 1 - MOT: 32khz square wave at 50mv (way out of spec for TTL logic)
Pin 2 & 3 - Oscillator: 32khz square wave at 5.5v
Pin 23 - SQW: No activity, logic low.
Pin 22 - PS: No activity, logic high.
Pin 21 - CKOUT: 32khz square wave at 5.5v
Pin 20 - CKFS: No activity, logic high.
Pin 19 - /IRQ: No activity, logic high
Pin 18 - /RESET: No activity, logic high
Pin 17 - DS: No activity, logic high
Pin 16 - STBY: No activity, logic high
Pin 15 - R/W: No activity, logic high
Pin 14 - AS: Open, just ambient noise.
Pin 13 - /CS: Open, just ambient noise.

With this in mind, my immediate suspect was the /CS, which is the chip select line. Following the trace back, it goes underneath the Keyboard controller to a TC4069 (equivilent of the CD4069, which is a 6x NOT-gate off the shelf logic chip). However, this is where it is interesting. It connects to pins 1 and 2 of the TC4069. Now, there is a reason I mention this, and it's because those are both the input and output of the NOT-gate. I can think of absolutely zero reasons to connect it up like this. As a logic 1 on this line would short the output to ground, and either burn out the TC4069, and a logic 0 would flip the bit turning output high and negating the input. I'm seriously wondering if this motherboard EVER WORKED given what I just discovered. It's almost as if they installed the incorrect IC from the factory....

Good investigating ! I cannot see the TC4069 in the picture. There are a few Tant caps around U37/U38 also c40, c62 and around the CPU that could be near shorting though doubt that would effect your scope reading much.
Still think something could be up with bad traces around 82C311...is late am not thinking clear....

Horun wrote on 2020-12-21, 05:52:

Good investigating ! I cannot see the TC4069 in the picture. There are a few Tant caps around U37/U38 also c40, c62 and around the CPU that could be near shorting though doubt that would effect your scope reading much.
Still think something could be up with bad traces around 82C311...is late am not thinking clear....

It and a 74LS06 (open collector buffer) are located underneath the KB controller. I don’t think the traces from the 82C311 are the issue. Everything looks good. I suppose there could be some corrosion underneath it. But I do not suspect that. I think this is an issue with either the logic ICs, or the 6818 RTC itself. As the 82C311 isn’t battery backed, and it’s data lines appear intact. So I’m thinking it has something to do with the RTC, as it has 50 bytes of user ram, and is battery backed.

Well, I figured the issue out with this board this morning.

Knowing that the issue was a problem with being able to read or write from the CMOS memory, I knew the problem was with the HM6818P circuitry. After some further thought, I couldn't see how the 6818 would have any issues with it at all. Everything seem perfectly fine with it. And it seemed to be generating the various different signals. And while I have a few MC146818P replacement chips coming in the mail, I thought there may be something else that is the culprit. Examining the circuit, the 6818's data lines are driven by a 74LS245 (octal bus transceiver). After probing around on the data lines between the transceiver and the 6818 I determined that there were no active signals. This led me to believe that it must have something to do with the output enable line for the transceiver. A quick probe on this pin showed that there was some initial activity. However, it was only pulling the signal down to about 4.5v. Which is well above the threshold for TTL logic to change states. This was my problem.

So I pulled the 6818 and Keyboard controller, and used a magnifying glass to inspect the area. Seeing that the trace to this pin ran underneath some components in the area I knew I would not be able to trace it around. However, I have a multimeter with continuity tester, so I decided to beep it out. After probing around for about 10 seconds I got a beep on a pin to the keyboard controller. But that didn't seem right. Inspecting the area again I noticed this is not anywhere near where the trace ran off to. So I keep probing around , and eventually find where the trace goes. And then notice a second trace from that pin down near where the Keyboard controller goes. Looking back at the pin that showed it was connected I noticed that the leg of the IC socket was bent over touching the board near that trace. I quickly used an small screwdriver to pry the pin up, and then tested for continuity again. This time it no longer beeped out.

At this point I decided to check the board to see if that made any difference. Sure enough, the board booted right up, displaying the amount of installed memory, and a warning that the CMOS memory was corrupt. After messing around with the bios I decided that this board is in fact working. Now to find an late 80s case to put it in.

Moral of the story: If you got a board that shows life, but wont boot, check all the small stuff before the big stuff. Make sure IC legs are not bent, take a close look around the area that your bios is telling you there is a problem. And if you still can't figure it out, donate it to me. I'll probably get it working and add it to my collection. 😉

Good work ! Glad you figured it out.

jdgabard wrote on 2021-01-01, 22:33:

And if you still can't figure it out, donate it to me. I'll probably get it working and add it to my collection. 😉

Damn, I will. You know quite a bit about early computer design.

Nice job !

Meanwhile I have completed the UH19 page with the bios and photo you've posted here 😀
http://www.win3x.org/uh19/motherboard/show/3539

Deksor wrote on 2021-01-02, 15:00:

Nice job !

Meanwhile I have completed the UH19 page with the bios and photo you've posted here 😀
http://www.win3x.org/uh19/motherboard/show/3539

Nice. You want me to post the PLD code as well? Those things do occasionally go bad. I can break out my programmer and read it in a bit if so.