VOGONS

Dead M919 Board - What to do next?

Topic actions

First post, by Pickelhaube808

Posted on 
User metadata
Rank Newbie
Rank
Newbie

A few weeks ago, I bought a PCChips M919 3.4B/F board for $50 that the seller stated was "tested and working". However, after receiving, putting in a known good processor and RAM (Intel 486 DX4 100 + 2 x 32MB EDO), and setting the jumpers to the correct settings (assuming the manual is correct), I get no video, no nothing. At least it seemed to be an honest mistake on the seller's part as he immediately refunded me upon me bringing it up to him.

After pressing the power button, all that happens is that the CPU warms up. Plugging in a speaker results in no beeps, and plugging in a POST card results in no codes. Cards in the ISA and PCI slots do however turn on, so at least they are getting power. Oddly though, the board acts roughly the same whether or not the BIOS chip is inserted.

What I have tried so far:

  1. Removing and replacing CMOS battery
  2. Setting the BIOS jumper to 5v/12v/NC
  3. Using vinegar, then Deoxit, then 99% IPA on corroded connectors
  4. Soaking the entire board in 99% IPA to remove any grime
  5. Different combinations of EDO memory in various slots (It SHOULD support EDO, right?)
  6. Different CPUs (an Intel DX2-50, along with changing the jumpers to reflect it)
  7. Flashed newest BIOS revision to an SST29EE010 chip
  8. Checking continuity on some of the traces, all of tested ones had continuity
  9. Tried ISA and PCI video cards
  10. Replaced all electrolytic capacitors on the board with high quality Panasonic ones of the same values
  11. Replaced diodes with new analogs, even though the old ones tested good
  12. Replaced all transistors and other voltage regulation components on the board with their modern analogs
  13. Replaced main 14.31818MHz crystal oscillator

What I have not tried so far (but could):

  • Replacing any of the tiny SMD components on the board
  • Replacing any of the resistors, including resistor arrays/networks
  • Reflow or replace larger package chips such as the chipsets or other controllers
  • Reflow every single solder joint on the front and back
  • Build an M919 cache module (It shouldn't need one to boot, right?)
  • Poke around the board with a multimeter or oscope when it's turned on

Considering what I have already tried, I know this is either a hard nut to crack or I am just overlooking something major. Any advice on how to proceed would be greatly appreciated though to revive this awesome board.

Here are some images of what I've got:

The attachment 20230422_203917.jpg is no longer available
The attachment 20230422_204411.jpg is no longer available
The attachment 20230422_204140.jpg is no longer available
The attachment 20230422_204227.jpg is no longer available
The attachment 20230422_210020.jpg is no longer available

The lighting wasn't very good for what I had to work with when I took these, so please let me know if you want to see any particular part of the board so I can get you a closeup image.

Reply 1 of 5, by Nexxen

Posted on 
User metadata
Rank l33t
Rank
l33t

Have you checked traces around the edges? Closest to borders may have got cut/scratched.
Checked V of the cpu?

Maybe look for Necroware videos to find ideas.

No L2 Cache? Tried setting to "none"?

PC#1 Pentium 233 MMX - 98SE
PC#2 PIII-1Ghz - 98SE/W2K
- "One hates the specialty unobtainium parts, the other laughs in greed listing them under a ridiculous price" - kotel studios
- Bare metal ist krieg.

Reply 2 of 5, by Pickelhaube808

Posted on 
User metadata
Rank Newbie
Rank
Newbie
Nexxen wrote on 2023-04-23, 02:17:
Have you checked traces around the edges? Closest to borders may have got cut/scratched. Checked V of the cpu? […]
Show full quote

Have you checked traces around the edges? Closest to borders may have got cut/scratched.
Checked V of the cpu?

Maybe look for Necroware videos to find ideas.

No L2 Cache? Tried setting to "none"?

Good ideas to start out. I guess I will have to find the pinout for the DX4 and probe the bottom of the board while it's running. I also thought about the fact there's no COAST module in there, but the CPU should(?) be able to run without it.

Reply 3 of 5, by SScorpio

Posted on 
User metadata
Rank Oldbie
Rank
Oldbie
Pickelhaube808 wrote on 2023-04-23, 02:32:

Good ideas to start out. I guess I will have to find the pinout for the DX4 and probe the bottom of the board while it's running. I also thought about the fact there's no COAST module in there, but the CPU should(?) be able to run without it.

It can run without the custom cache module. Is not COAST and trying a real COAST module can fry things.

A post card would let you see it there's any life.

Maybe try replacing the BIOS with a freshly flashed chip.

Reply 4 of 5, by Thermalwrong

Posted on 
User metadata
Rank Oldbie
Rank
Oldbie

1. Press on either of the two main chips while powering up - in my case this took a dead board from "--" on the POST card to giving meaningful codes. These boards are often 1.2mm thick instead of 1.6mm thick so it can help. In which case, you should reflow the solder joints to resolve the issue. I revived a board by pressing / resoldering the ASIC by the memory slots.
2. Test with a 5V CPU - the essentially stops the BJT transistors, the chips with the heatsinks from being a problem. You can replace them with similar replacements if that's the fault.
3. Make sure the memory is in the correct slots - my M919 with the MASK ROM BIOS refuses to boot if RAM isn't in the lowest 72-pin SIMM slot. Later BIOS versions are better about this though
4. In terms of RAM, try FPM only for now. Try with EDO only when the board is known to be working and try the latest BIOS to see if EDO support improves, I think that improved EDO support on mine.

Reply 5 of 5, by Pickelhaube808

Posted on 
User metadata
Rank Newbie
Rank
Newbie
Thermalwrong wrote on 2023-04-23, 04:13:
1. Press on either of the two main chips while powering up - in my case this took a dead board from "--" on the POST card to giv […]
Show full quote

1. Press on either of the two main chips while powering up - in my case this took a dead board from "--" on the POST card to giving meaningful codes. These boards are often 1.2mm thick instead of 1.6mm thick so it can help. In which case, you should reflow the solder joints to resolve the issue. I revived a board by pressing / resoldering the ASIC by the memory slots.
2. Test with a 5V CPU - the essentially stops the BJT transistors, the chips with the heatsinks from being a problem. You can replace them with similar replacements if that's the fault.
3. Make sure the memory is in the correct slots - my M919 with the MASK ROM BIOS refuses to boot if RAM isn't in the lowest 72-pin SIMM slot. Later BIOS versions are better about this though
4. In terms of RAM, try FPM only for now. Try with EDO only when the board is known to be working and try the latest BIOS to see if EDO support improves, I think that improved EDO support on mine.

Definitely going to try #1 out. Unfortunately, #2 is something I had already tried as I did so with my DX2-50. I also replaced the BJTs with analogous ones from Mouser.

As for #3, I do happen to have two sticks, one in #1 and #2. If convenient, would you be able to also see what BIOS chip you have? I am curious as to if everyone has the "M919" badged EEPROMs or if people have gotten something like a 29EE010 working in the slot.

And as for #4, I will buy two or four sticks of FPM now just in case.

Thank you :^)

Go to top of page Go to top of page
Back to General Old Hardware
Close window
Close window