What are the chipsets of the cards you are trying? most cards of the era should work.

Please post a full pic of your board (showing the way you jumpered it along with which CPU).
If PCi is the problem, then maybe you setup your PCI incorrectly? Maybe you're inadvertently overclocking or underclocking your PCI bus in such a way it makes your PCI cards problematic.

dirkmirk wrote:

What are the chipsets of the cards you are trying? most cards of the era should work.

I use pretty standard solutions: S3 trio 64V+ and S3 Trio Virge. I have in my arsenal PCI version of geforce2, and i accept that this card may not work - it does in my P100 config.

Tetrium wrote:

Please post a full pic of your board (showing the way you jumpered it along with which CPU).
If PCi is the problem, then maybe you setup your PCI incorrectly? Maybe you're inadvertently overclocking or underclocking your PCI bus in such a way it makes your PCI cards problematic.

Pics attached. Looks correct, not much to do with jumpers on this board.

Yesterday cache chips were replaced - same result.

Maybe the PCI is configured incorrectly - than again - how to clear this setup with black screen - the answer is: 'clear cmos jumper' - just how to do this without pins on the motherboard - should I solder them on?

How can i hard reset CMOS on this board?

In BIOS try "Load Power On defaults" and "Load Setup defaults".

To troubleshoot select slowest BIOS settings and disable all the things not essential. Use ony 1 memory stick and VGA card. Also try ISA VGA card if you have one at hand. Try different PCI slots. Disable FDD/HDD... and so on. Start with disabling everything.

Double check all the jumpers, especially the one for PCI bus divide.

Here's a better and bigger picture of Your mobo
http://img37.imageshack.us/img37/3355/dscf0057ej.jpg

Here's the configuration information
http://www.arvutimuuseum.ee/th99/m/A-B/33492.htm

JP15 and JP16 should both be on. I think You currently have it set for a 40mhz fsb which might be the problem.
JP17 does not exist on Your mobo, so ignore that.

kixs wrote:

In BIOS try "Load Power On defaults" and "Load Setup defaults". […]
Show full quote

How can i hard reset CMOS on this board?

In BIOS try "Load Power On defaults" and "Load Setup defaults".

To troubleshoot select slowest BIOS settings and disable all the things not essential. Use ony 1 memory stick and VGA card. Also try ISA VGA card if you have one at hand. Try different PCI slots. Disable FDD/HDD... and so on. Start with disabling everything.

Double check all the jumpers, especially the one for PCI bus divide.

Something wrong must have happened because I cant get picture everytime I try now, so setting options in BIOS is not possible. PCI diagnostic card is giving me "00". That doesn't look good 😒

After unplugging all [cpu, ram, vga] the error is still "00".

Is there an option to reset BIOS on this mobo with jumpers?

Do you have a standard 486DX? Like with 33 MHz?

Use that to start your trouble shooting.

This board has a LOT of jumpers that you need to get right. So triple check everything 😀

Imperious wrote:

Here's a better and bigger picture of Your mobo http://img37.imageshack.us/img37/3355/dscf0057ej.jpg […]
Show full quote

Here's a better and bigger picture of Your mobo
http://img37.imageshack.us/img37/3355/dscf0057ej.jpg

Here's the configuration information
http://www.arvutimuuseum.ee/th99/m/A-B/33492.htm

JP15 and JP16 should both be on. I think You currently have it set for a 40mhz fsb which might be the problem.
JP17 does not exist on Your mobo, so ignore that.

I practicaly copied that jumper settings from Your link (minus the cache ones), nothing changed. Added jumper, so now there are jumpers on both JP15 and JP16, nothing changed.

Fortunately i took pictures from bios before problems started:

opiate wrote:

dirkmirk wrote:

What are the chipsets of the cards you are trying? most cards of the era should work.

I use pretty standard solutions: S3 trio 64V+ and S3 Trio Virge. I have in my arsenal PCI version of geforce2, and i accept that this card may not work - it does in my P100 config.

The bottom pic (the one with the cache chips visible), according to the link with the manual, JP5, HP6 and JP7 need to be 1-2, 1-2 and 1-2. Yours seems to be set to 1-2, open and 1-2.

How are things going atm?

And could you still post a pic of your entire motherboard?

PhilsComputerLab wrote:

Do you have a standard 486DX? Like with 33 MHz?

Use that to start your trouble shooting.

This board has a LOT of jumpers that you need to get right. So triple check everything 😀

And this is some really good advice here and to me kinda a staple when troubleshooting 486 boards.

HOST CLK:PCI CLK 1 : 2/3

That would currently set the PCI speed to 22mhz!! If The FSB is 33mhz that needs to be 1:1

You can reset the cmos by shorting out the 2 solder points with a paper clip or anything like that, just do it while
the power is off.
Your mobo has a Dallas RTC/Battery all-in-one, which is why You don't have a cr2032 battery soldered there.

Imperious wrote:

HOST CLK:PCI CLK 1 : 2/3 […]
Show full quote

HOST CLK:PCI CLK 1 : 2/3

That would currently set the PCI speed to 22mhz!! If The FSB is 33mhz that needs to be 1:1

You can reset the cmos by shorting out the 2 solder points with a paper clip or anything like that, just do it while
the power is off.
Your mobo has a Dallas RTC/Battery all-in-one, which is why You don't have a cr2032 battery soldered there.

It's not set to 33MHz, according to the pics of JP15 and JP16. Could be set to 40MHz and this would explain the divider visible in the BIOS. I couldn't find JP17, so no idea how this is set.

Tetrium wrote:

It's not set to 33MHz, according to the pics of JP15 and JP16. Could be set to 40MHz and this would explain the divider visible in the BIOS. I couldn't find JP17, so no idea how this is set.

If You read slightly above in the thread, He has sorted that out. The page I linked to with config information says JP17 is probably not installed, so can be ignored.

Imperious wrote:

Tetrium wrote:

It's not set to 33MHz, according to the pics of JP15 and JP16. Could be set to 40MHz and this would explain the divider visible in the BIOS. I couldn't find JP17, so no idea how this is set.

If You read slightly above in the thread, He has sorted that out. The page I linked to with config information says JP17 is probably not installed, so can be ignored.

This was after he had taken the pics of both his jumpers and BIOS screens though and as he didn't change his cache jumpers, these are probably still set incorrectly.

edit: Nvm, just noticed JP6 is a block of 2x2 pins and (if the right pair of pins are 1 and 2) is actually set correctly (I thought the right pair was JP7, but I misread). My apologies 😊

And I actually edited the wrong post, should've edited this one and not my previous one, time for bed 😁

edit2: But anyway, dunno what the issue with his motherboard could be, except perhaps he could try a 5v DX-33 part for testing or maybe try another motherboard.

Imperious wrote:

HOST CLK:PCI CLK 1 : 2/3 […]
Show full quote

HOST CLK:PCI CLK 1 : 2/3

That would currently set the PCI speed to 22mhz!! If The FSB is 33mhz that needs to be 1:1

You can reset the cmos by shorting out the 2 solder points with a paper clip or anything like that, just do it while
the power is off.
Your mobo has a Dallas RTC/Battery all-in-one, which is why You don't have a cr2032 battery soldered there.

I tried everything, every posible jumper setting, even tried ISA graphics card that is working in another retro-build - no change, still no picture.

So the only thing left is to reset cmos - the question is: what should i short:

- JP42 like the manual is suggesting? Note that in manual it is said that pins 2 and 3 are closed, and they are clearly not. Pins arent even soldered to the board.
or
- two soldering dots in that place where battery should be?

Ehhh, It pains me to see good motherboard like this dead 😐

opiate wrote:

I tried everything, every posible jumper setting, even tried ISA graphics card that is working in another retro-build - no chang […]
Show full quote

Imperious wrote:

HOST CLK:PCI CLK 1 : 2/3 […]
Show full quote

HOST CLK:PCI CLK 1 : 2/3

That would currently set the PCI speed to 22mhz!! If The FSB is 33mhz that needs to be 1:1

You can reset the cmos by shorting out the 2 solder points with a paper clip or anything like that, just do it while
the power is off.
Your mobo has a Dallas RTC/Battery all-in-one, which is why You don't have a cr2032 battery soldered there.

I tried everything, every posible jumper setting, even tried ISA graphics card that is working in another retro-build - no change, still no picture.

So the only thing left is to reset cmos - the question is: what should i short:

- JP42 like the manual is suggesting? Note that in manual it is said that pins 2 and 3 are closed, and they are clearly not. Pins arent even soldered to the board.
or
- two soldering dots in that place where battery should be?

Ehhh, It pains me to see good motherboard like this dead 😐

ASUS A7V133 I worked on also had this 'feature'. I just used a flatbed screwdriver to short these solder dots.
Dunno for sure if it'll work on your board, but that's what I did and it seemed to work?

These A7V133s didn't have the Dallas chip though, so dunno how this will affect this.

It's time to write down the time of death. 🙁

Clearing CMOS [shorted jp42 and the battery pins] did nothing. The battery pins are actually connected with JP42 BTW.

At the end nothing helped but big thnx for all your help guys!

Time to hunt down another socket 3 board with PS2 - now that is a challenge 😀

Opiate: You are attempting to short the wrong two solder pads for clearing the CMOS. The implementation of JP42 was for one possible design of this board which incorporated the CR2032 battery and a different clock module. All the MB-8433UUD boards I have seen, which include the first version all the way to the last version (v3.1) have a Dallas/ODIN Real-Time Clock module, which has an integrated battery. You should be setting the JP12 jumper to 3-4 to clear the CMOS. Then open the jumper to return to normal operation. Jumper headers 1/2 should be solder pads only. The caveat to this is that some RTC modules do not have the ability to clear the CMOS settings. Such modules are those which do NOT contain an A suffix, e.g. DS12887+ or DS12887, or DS1287. I am pretty sure that most of these boards were outfitted with an A-type RTC, usually an ODIN OEC12C887A, which allows for clearing the CMOS data preferences.

I have seen this board get hung-up on some PCI cards, whereby only a clearing of the CMOS was able to resolve. For those boards without the A-type RTC, you might need to desolder your RTC and replace it, or wait a few more years for the battery to no longer hold charge.

You mentioned earlier about a GeForce2. This board will work with a GeForce2 MX400 if you are using an Am5x86 or Pentium Overdrive. I could not find any functioning drivers for such card if you are using a Cyrix/IBM 5x86. If you decide you want to try a GF2, use NVIDIA Detonator drivers v12.41 w/Am5x86. If using a POD, newer drives up to 77.72 will work.

You mentioned that JP17 is missing the header. I've seen this happen on the older version 2.0 boards. Newer version 2.0 boards have the jumper header in place. I'm not sure why they didn't change the PCB revision. The newer version 2.0 boards are a lot similar to the early 3.0 boards. Having JP17 soldered closed means that you cannot use FSBs of 20, 50, 60, and 83 MHz, but 25, 33, 40, and 66 MHz are available to you.

Concerning your intermittent PCI video issue (I realise it is no longer intermittent, it doesn't work at all), I wonder if your crystal is failing. It would be helpful to probe the PCI clock with an oscilloscope. I suppose your PLL could also be failing, which on this board is socketed and easy to replace. I have had issues on my version 1 board, which did not allow me to run the PCI bus at 1:1 (CMOS setting), even with a 33 MHz BUS. From that experience, I just avoid version 1 boards. Version 1 boards don't have any version numbers printed on them.

Also, this board works best with FPM memory. And if you are trying to use a CF card with the onboard IDE, I find that you need to set IDE Primary Master PIO to PIO-4, not AUTO.

I have looked over your BIOS settings. I will point out some of the differences from what I usually set.

PCI IDE IRQ Map To : PCI-AUTO
IDE Primary Master PIO - AUTO, if using CF card, set to PIO-4.
DRAM Read/Write = 0WS if using 40 MHz FSB or less. (Note my RAM is 60 ns or better)
L2 cache wait states = 2-1-1-1 if using 40 MHz or less and 256 KB of double-banked cache. (Note my cache is 15 ns or better)
Early Cache Write Mode = Enabled
Burst Copy Option = Enabled

What BIOS revision are you using? You might want to consider updating it. Updating the BIOS will also reset the ESCD, which may help with some PCI/ISA configuration problems.

feipoa wrote:

Opiate: You are attempting to short the wrong two solder pads for clearing the CMOS. The implementation of JP42 was for one poss […]
Show full quote

Opiate: You are attempting to short the wrong two solder pads for clearing the CMOS. The implementation of JP42 was for one possible design of this board which incorporated the CR2032 battery and a different clock module. All the MB-8433UUD boards I have seen, which include the first version all the way to the last version (v3.1) have a Dallas/ODIN Real-Time Clock module, which has an integrated battery. You should be setting the JP12 jumper to 3-4 to clear the CMOS. Then open the jumper to return to normal operation. Jumper headers 1/2 should be solder pads only. The caveat to this is that some RTC modules do not have the ability to clear the CMOS settings. Such modules are those which do NOT contain an A suffix, e.g. DS12887+ or DS12887, or DS1287. I am pretty sure that most of these boards were outfitted with an A-type RTC, usually an ODIN OEC12C887A, which allows for clearing the CMOS data preferences.

I have seen this board get hung-up on some PCI cards, whereby only a clearing of the CMOS was able to resolve. For those boards without the A-type RTC, you might need to desolder your RTC and replace it, or wait a few more years for the battery to no longer hold charge.

You mentioned earlier about a GeForce2. This board will work with a GeForce2 MX400 if you are using an Am5x86 or Pentium Overdrive. I could not find any functioning drivers for such card if you are using a Cyrix/IBM 5x86. If you decide you want to try a GF2, use NVIDIA Detonator drivers v12.41 w/Am5x86. If using a POD, newer drives up to 77.72 will work.

You mentioned that JP17 is missing the header. I've seen this happen on the older version 2.0 boards. Newer version 2.0 boards have the jumper header in place. I'm not sure why they didn't change the PCB revision. The newer version 2.0 boards are a lot similar to the early 3.0 boards. Having JP17 soldered closed means that you cannot use FSBs of 20, 50, 60, and 83 MHz, but 25, 33, 40, and 66 MHz are available to you.

Concerning your intermittent PCI video issue (I realise it is no longer intermittent, it doesn't work at all), I wonder if your crystal is failing. It would be helpful to probe the PCI clock with an oscilloscope. I suppose your PLL could also be failing, which on this board is socketed and easy to replace. I have had issues on my version 1 board, which did not allow me to run the PCI bus at 1:1 (CMOS setting), even with a 33 MHz BUS. From that experience, I just avoid version 1 boards. Version 1 boards don't have any version numbers printed on them.

Also, this board works best with FPM memory. And if you are trying to use a CF card with the onboard IDE, I find that you need to set IDE Primary Master PIO to PIO-4, not AUTO.

I have looked over your BIOS settings. I will point out some of the differences from what I usually set.

PCI IDE IRQ Map To : PCI-AUTO
IDE Primary Master PIO - AUTO, if using CF card, set to PIO-4.
DRAM Read/Write = 0WS if using 40 MHz FSB or less. (Note my RAM is 60 ns or better)
L2 cache wait states = 2-1-1-1 if using 40 MHz or less and 256 KB of double-banked cache. (Note my cache is 15 ns or better)
Early Cache Write Mode = Enabled
Burst Copy Option = Enabled

What BIOS revision are you using? You might want to consider updating it. Updating the BIOS will also reset the ESCD, which may help with some PCI/ISA configuration problems.

Thnx for your input, lots of good information, especially about bios settings - i do plan to use CF adapter.

Regarding the RTC, my board have the A suffix on the Dallas/Odin chip (You can see this in my pictures few posts up], yet shorting jp12 [3&4 pins] does nothing.

You mentioned that desoldering that chip could help - well, i'm not afraid of soldering iron, so that what i'll do.

BIOS rev is on the attached picture, if there is newer bios that maybe i could update it outside MB, saw people do that using some device from china.