Chkcpu wrote on 2025-04-04, 18:51:

Hi byte_76, […]
Show full quote

byte_76 wrote on 2025-04-04, 17:26:

This is the first time that I've used a T48 programmer so I hope that I extracted the file correctly. […]
Show full quote

Chkcpu wrote on 2025-03-15, 08:59:

Hi byte_76, […]
Show full quote

Hi byte_76,

If you have an (E)EPROM Programmer, can you post a dump of the BIOS chip contents here?
Then I should be able to tell you if the BIOS is corrupted or valid and correct for your board with SMC I/O.
And it will be nice to have for TheRetroWeb as well. 😀

Cheers, Jan

This is the first time that I've used a T48 programmer so I hope that I extracted the file correctly.

I have attached the BIOS image. (zipped file)
Can you check it and confirm whether I saved it correctly and what information you can obtain from it?

Hi byte_76,

Okay, you got the T48 programmer and made the BIOS dump from your B887 board!

I’ve checked this 128KB Award BIOS and it is indeed a Rev.D with BIOS-ID:
02/16/96-SiS-496-497/A/B-2A4IBM89C-00

I used Award’s CBROM and MODBIN tools on this compressed BIOS and everything looks perfect. Apart from a few small differences in the BIOS Setup, the main BIOS module (original.tmp) in this Rev.D BIOS is identical to that of the VER.B “LAZER-TRON” BIOS. This also means that your Rev.D BIOS is correct for your board with SMC Super I/O chip.

I also tested your Rev.D BIOS in the LS-486e emulated machine for SiS 496 chipset in 86Box and it booted fine. This is the POST screen:

The attachment B887_RevD.png is no longer available

So now you know the BIOS is good, and that the fault must be elsewhere.
I hope you will find the problem.

Cheers, Jan

This is both good and bad news.

Thank you for testing and sharing the feedback.

I need help with troubleshooting this board please.

I don’t have much electronics experience as I’ve only learnt the basics as part of this hobby.

I have a multimeter.
I do not have an oscilloscope.
I have a post analyzer card.

I don’t yet have a 5V cpu but pretty sure this board should work with 3.3V - 4V CPU’s and I’m testing with my Cyrix DX2 66 that is marked 3.45V/3.6V/4.0V.

Could anyone mark the points on the board that I need to test with my multimeter and advise roughly what readings to expect please?

It is difficult to pinpoint specific test points, especially on a rare board as yours. But some general advice can be given.

Look at this helpful “How to repair a retro PC motherboard” video by Necroware:
https://www.youtube.com/watch?v=A0OXbKaYMbQ

He is showing a nice step by step “how to” to start troubleshooting.
Although he is using an oscilloscope for specific checks, you can also use a much cheaper logic probe in most cases.

Cheers, Jan

I’ve watched the video and it’s definitely useful but he basically found a broken trace and ended the troubleshooting there.

I have checked (in very thorough detail) for broken traces using my digital microscope and I’m sure that is not the issue here.

Are there any analyzers or other devices that can help me figure out what’s wrong with this board?

Otherwise, let’s take it one step at a time?

I’ve checked:

- No broekn traces or physical damage
- CPU voltages present on the socket
- Tested without memory and expansion cards but no beeps or post codes
- Reset flashes briefly on analyzer card at power on
- No other codes on analyzer card
- Reprogrammed the BIOS rom - No change

What can I check next? (Without an oscilloscope)

Is it possible that the capacitors may be faulty, resulting no post codes without them being shorted?

BIOS rom type

1) Is the board set for 512kb cache ? is it really 512kb cache ? U12 jumpers.
2) try the 2 modules os ram in the other bank (other two slots)
3) test with 4 sticks of ram

nuno14272 wrote on 2025-04-11, 10:45:

1) Is the board set for 512kb cache ? is it really 512kb cache ? U12 jumpers.
2) try the 2 modules os ram in the other bank (other two slots)
3) test with 4 sticks of ram

1) It’s configured for 256K and the board only has 256K soldered in.

2) I’ve tried all slots with 1 and 2 modules installed and without any modules installed. No change in behavior.

3) I’ll try but really not sure how this could help.

Maybe there is a faulty cache module but since they’re soldered in, it’s not so easy to diagnose.
There also doesn’t seem to be an option in the jumpers to disable the cache.

If I could find a manual or jumper settings document it would help but no luck with that so far.

Since this board does not use a Dallas RTC module, which chip on this board is the RTC?

The SiS 85C497 IC has the RTC integrated, according to the datasheet, page 3:

https://www.dosdays.co.uk/media/sis/SiS%20496-497.pdf

Possible foriegn matter in card slots?

Double check no shorts caused by screws or standoffs? (I know it feels insulting to ask this, but really.)

There are no screws or standoffs installed.

I did check the slots, using my magnifying glass and microscope but I didn't see any dirt. The board was clean but certainly not entirely spotless, so to be sure, today I have sprayed the board with IPA and left it to soak for a few minutes and then washed it with warm water and dish soap and a soft paint brush. (With CMOS battery and BIOS rom removed as well as the CPU socket lid which was very carefully removed. Nothing else is removable)
I used a blower to get the water out from under the components and dry off the board to the point that I could no longer see any moisture anywhere on the board. It is sparkling clean now.

I will leave it over night to be absolutely sure it's dry before I do anything further with it as far as power is concerned.

Thinking about this more....I have had 2 similar boards (ls486e, zida...whatever) that woke up after extensive solder reflow of predominantly the pci slots. One of these boards would sometimes work with a 5v cpu. There was no obvious bad solder joints, I can only assume that one of the traces on either side (or possibly one of the inner layers?) wasn't connecting.

Unfortunately washing the board has not changed the symptoms at all.

I have tested memory slots full populated and also no change.

I have not tried reflowing the solder as that is quite a big job and my soldering iron is not very good. It will work fine to solder a few components but touching all the solder points for each slot would probably leave them in a messy state and even less likely to work. I would leave that as a last resort.

Have you checked the pins of the ram slots? I had two boards which seemed dead and without post codes which was caused by a single bent pin in a memory slot touching a neighboring pin. Unbending the pin brought both boards back to life. This may not be the issue but is worth checking in my opinion.

I just looked at his high res photos of the board. The ram slots are very clear, and in GREAT focus.

I do not see either foreign matter or bent pins.

The two centermost slots have an odd circular pattern on the bottom end, which might be a crack.

It would be very unusual for that to be a molding mark, because the shape covers TWO different slots.

I would suggest getting an ISA ram card, like a lotech 1mb card, and popping THAT in.

That was just a hair on the slots.

The board has since been washed so it’s sparkling clean now.

The symptoms are still the same though.

Is it possible that the board does not support 3.3V CPU’s (or anything below 5V) despite having multiple voltage options?
It is a socket 2 board, not sure if that matters.
Is there a difference in pinout that might not be supported by this board?

I don’t have a 5V CPU to test yet.

256KB is not only the upper jumper ? 2 jumpers is 512kb.

Did I miss a post that contains progress, or is this board still stuck at "no POST code on the analyzer card all"? If there still is no single POST code, we don't need to consider the RAM type or the cache jumpers. To get to the first POST code, you basically need:

• A working CPU. I assume the CPU is working.
• A working CPU clock signal. Even with just a multimeter, you can detect "clock stuck high" and "clock stuck low". The clock pin at the CPU should read between 1.5 and 3.5V at 5V signalling voltage (which this board will use even with 3.3V CPUs).
• A working reset circuit. If your POST analyzer card displays the RESET signal, it should be lit for 0.1 to 0.5 seconds, and then get dark. Also, the RESET signal at the CPU should be high for a short time and then go low. The 80486 has an active high reset signal.
• Nothing that disturbs the front-side bus. Broken L2 cache chips and VL cards are able to disturb the frontside bus, so I suggest to remove all plug-in cards and all cache chips.
• Working connection of the front-side bus to the chipset. Broken solder joints at the chipset might cause issues. Both the SiS496 and the SiS497 need to be connected properly.
• Working connection of the SiS497 to the X-Bus, which contains the BIOS chip.
• Nothing that disturbs the X-Bus. One might remove the keyboard controller to prevent a broken keyboard controller from messing up the system, but I have no idea what side-effects a missing KBC might have, and how far into the POST the system can get without an KBC. I suggest to leave it installed.
• Proper connection from the X-Bus and the SiS497 to the ISA bus.

You don't need the ISA bus to work for beeping. So even if the IO-Write signal on the ISA bus were broken, as long as it isn't already broken on the X-Bus, you should be able to get beep codes if RAM is missing.

Assuming, RESET is working, try to verify that the system tries to read from the BIOS after reset. Measure the voltages at /CE and /OE of the BIOS chip with a meter. I expect steadily around 4 to 5 V while the system is idle or in RESET (reset button connected and pushed). The voltage as shown by a meter should drop for one reading at least when you let go the reset button. As the meter isn't actively measuring the whole time, repeat the process several times. If there is no reaction to reset on either of the pins at the BIOS chip, the issue is somewhere between the processor and the BIOS. While you checked the traces with a micrososcope, my next step still would be to check continuity with a meter from all the 30 address lines and the bus control lines R/W, M/IO and C/D and ADS at the 80486 socket to the corresponding pins at the chipset. Also check for shorts between neighbouring signal pins at the chipset - but don't be tricked by neighbouring pins that are intentionally connected to the same level, like GND.

You haven’t missed anything. The board is still NO POST.

Thank you for the feedback. As a novice in electronics, I’ll try to the best of my understanding.

On the reset, If I press the reset button while powered, the reset light flashes on the analyzer card as well. Not sure if that means anything.

The processor and the north bridge get warm while powered.

The cache and KBC are soldered.

When I try display, I am testing with a PCI display card and nothing in the ISA slots but of course I’ve tested without any ram or cards.

This board has socket 2.
Maybe despite the voltage regulator and jumper options, it doesn't actually support processors below 5V.
Perhaps BIOS support is needed and is not present for the 3.3V CPU's.

I don't have a 5V CPU to test at the moment but I have ordered one and will be testing it as soon as I can.