This say QDI....

http://www.hardwarecentral.com/showthread.php … INGS-for-QDI-MB

well this board is either extremely rare or part of a specialized machine, because that's the only links google will show...

I noticed that too. Hmmm..... I did not see much more than that by googling "motherboard vip596P3"

It's posible it's rev 2.0 of V5P596P2. I figured out witch jumpers set the FSB, so I'm now able to set the thing to as low as 20MHz or as high as 75MHz (not stable at 75, won't boot) 😀 I think I can get 286 speeds out of it at 20MHz with cache disabled - but I'm having some issues with L2 cache. It's a great board for simulating a 486 with a fat FPU or a 63MHz socket 3 POD 😀

The thing is I remember there were 2 versions of the socket 4 pentium - one running at 5v and one running at some other voltage 5.2 or 5.4v - as such I need to find witch jumper sets that...

If you decide to sell it contact me. Socket 4 boards with VLB are really hard to find and I am looking for one 😀

I imaging P2 at the end would be revision 2 and P3 revision 3.

As for page faults - if it is bad memory, did you try testing one memory stick at a time? Memtest and Memtest86 are good thorough tests for bad memory.

The thing is I remember there were 2 versions of the socket 4 pentium - one running at 5v and one running at some other voltage 5.2 or 5.4v - as such I need to find witch jumper sets that...

As far as I know, the 60MHz Pentium ran at 5V, and the 66MHz version at 5.15V. If there are other versions, I am currently not aware of them. I'm really curious if there were actually Socket4 boards (other than the official intel ones) that provided 5.15V for the 66MHz part.

debs3759 wrote:

If you decide to sell it contact me. Socket 4 boards with VLB are really hard to find and I am looking for one 😀

Not really sure about that. I'd be more interested in a trade for some other rare parts - namely a 100MHz capable PoD or preferably a 133Mhz Cyrix/IBM 586. Alltought right now I just want to get an original pentium system running and mess around with it for a bit.

debs3759 wrote:

As for page faults - if it is bad memory, did you try testing one memory stick at a time? Memtest and Memtest86 are good thorough tests for bad memory.

The board will only post with 2 simm modules (64 bit) so it's impossible to test them one at a time in this machine. I have tested them in a 586 and they were error free. I also tried other modules - same problem. I also downclocked the CPU to 50MHz - no dice.

Right. It seems emm386 refuses to load. I get an "insufficient memory" error.

If you have any questions regarding stability, performance tweaking, feel free to ask. I had one of these OPTi Cobra chipset boards in 1995 and now again one NOS board from AT&T.
These boards are fun and run stable when configured correctly. But the video performance lags behind the competition. Despite this, the memory performance is remarkable and above the Intel Mercury based boards (Batman + Revenge, Hendrix and so on.)

A nice piece of computing history.

On Topic: You did use FPM memory? Did you try with disabled L2 cache? I wonder why this board has a 64k TAG RAM instead of the 256k my boards were and are using. A manual of one of my boards advises to use 256k as TAG RAM.

FGB wrote:

If you have any questions regarding stability, performance tweaking, feel free to ask. I had one of these OPTi Cobra chipset boards in 1995 and now again one NOS board from AT&T.
These boards are fun and run stable when configured correctly. But the video performance lags behind the competition. Despite this, the memory performance is remarkable and above the Intel Mercury based boards (Batman + Revenge, Hendrix and so on.)

A nice piece of computing history.

I actually have a couple of issues with the board. First, I can't seem to be able to get EMM386 to run. Regardless of what I try, i get "Insufficient Memory, emm386 not loaded". The second is artefacting in 640x480@16 bit with any video card I try, PCI or VLB. It seems to be worse for VLB since these artefact at 640x400@256 col. I eliminated all possible software causes, all that's left is the hardware.

- Do you have the jumpers set to 1 WS for the VLB?
- The board should have options for the source of the clocks. External or internal via dividers. I recommend to set the clocks for the PCI / VESA Bus to external, it will use one of the 33MHz crystals to generate the clocks.
- There should be also a jumper for BACK TO BACK IO DELAY. I recommend it to set it to "ENABLED".
- Try to test without cache. These boards are sensitive and highly rely on cache stability to perform.
- Did you always close the turbo SW? My manual says: "if you use EMM386.EXE, the system will not allowd to enter non-turbo mode" - However, the statement may only be true for the specific board (MB-5DVP)
- I can provide another BIOS, you can flash and try.

FGB wrote:

- Do you have the jumpers set to 1 WS for the VLB?
- The board should have options for the source of the clocks. External or internal via dividers. I recommend to set the clocks for the PCI / VESA Bus to external, it will use one of the 33MHz crystals to generate the clocks.
- There should be also a jumper for BACK TO BACK IO DELAY. I recommend it to set it to "ENABLED".

I only have approximate jumper settings for the board... no proper manual or jumper settings anyware so I can't really set anything up 🙁

FGB wrote:

- Try to test without cache. These boards are sensitive and highly rely on cache stability to perform.

Yup. Tested with cache disabled and I get the same error and video artifacts. Also tested with 128 and 512kb of cache, 8kb and 16kb TAG, as well as completly replacing ALL cache chips with known working ones.

FGB wrote:

- Did you always close the turbo SW? My manual says: "if you use EMM386.EXE, the system will not allowd to enter non-turbo mode" - However, the statement may only be true for the specific board (MB-5DVP)

I did not. But I don't think it has a turbo switch.

FGB wrote:

- I can provide another BIOS, you can flash and try.

If it's for my board, then by all means, thank you.

So far I've been able to ID the following jumper settings:

JP6 1-2 8kb TAG RAM | 2-3 16kb TAG RAM

64kb Cache - JP21 open | JP20 open| JP19 open
128kb Cache - JP21 closed | JP20 open| JP19 open
256kb Cache - JP21 closed | JP20 closed | JP19 open
512kb Cache - JP21 closed | JP20 closed | JP19 closed

Cache uses banks 0 only JP22 2-3 | JP23 2-3
Cache uses bank 0 only with 256k cache - JP22 2-3 | JP23 1-2
Cache uses both banks JP22 1-2 | JP23 1-2

CPU Speed:
75MHZ - JP24 closed | JP25 closed| JP26 open
66MHZ - JP24 closed | JP25 open | JP26 closed
60MHZ - JP24 open | JP25 closed | JP26 open
40MHZ - JP24 open | JP25 closed| JP26 closed
20MHZ - JP24 open | JP25 open | JP26 open

debs3759 wrote:

If you decide to sell it contact me. Socket 4 boards with VLB are really hard to find and I am looking for one 😀

Have you seen this one? It's not socket 4, but socket 5 and has VLB slots.

http://www.ebay.com/itm/301728666510?_trksid= … K%3AMEBIDX%3AIT

It's pretty cool. I already have a nice socket 5 board (UMC chipset, turbo button, supports FSB 20, supports 2x multi) - PCI tough, witch I want for a socket 4/5 machine so I can use a Voodoo card.

I bet your board has a Turbo switch. I never saw any Socket 4 Board without one. But it's likely that your problem isn't related to the Turbo switch, though.
Regarding the BIOS: Most BIOSes of the OPTi Premium / Cobra chipset are interchangable. One can take a BIOS of a different board and try. Currently I have a Shuttle HOT-523 BIOS installed in my AT&T Motherboard which enables the memory / cache timings and makes the board A LOT faster. I mean.. really A LOT. But this won't help you at this point unless your BIOS is corrupted.

FGB wrote:

I bet your board has a Turbo switch. I never saw any Socket 4 Board without one. But it's likely that your problem isn't related to the Turbo switch, though.
Regarding the BIOS: Most BIOSes of the OPTi Premium / Cobra chipset are interchangable. One can take a BIOS of a different board and try. Currently I have a Shuttle HOT-523 BIOS installed in my AT&T Motherboard which enables the memory / cache timings and makes the board A LOT faster. I mean.. really A LOT. But this won't help you at this point unless your BIOS is corrupted.

I have memory and cache timings in BIOS. Also PCI level/edge and irq settings. The only thing I haven't tried is setting PCI IRQ to "Level" instead of "Edge".

BTW, the thing score 49 pts in speedsys - that's as much as a 133MHz 486 - is that normal?

FGB wrote:

I bet your board has a Turbo switch.

I've studied the PCB quite a bit and found no possible trace of a tb switch header. There's a high-res photo in the first post - look it over yourself. It would be cool if this was the issue.

49-50 points are normal for a Pentium @66MHz. But of course SpeedSys doesn't reflect real world CPU performance in its CPU scores (For example, a 386 without FPU scores much higher in Speedsys than a 386 at the same speed with a FPU installed.)

Will take a look at the picture later, maybe we find that pin header. //edit: Can't find one.

kanecvr wrote:

It's posible it's rev 2.0 of V5P596P2. I figured out witch jumpers set the FSB, so I'm now able to set the thing to as low as 20MHz or as high as 75MHz (not stable at 75, won't boot) 😀 I think I can get 286 speeds out of it at 20MHz with cache disabled - but I'm having some issues with L2 cache. It's a great board for simulating a 486 with a fat FPU or a 63MHz socket 3 POD 😀

The thing is I remember there were 2 versions of the socket 4 pentium - one running at 5v and one running at some other voltage 5.2 or 5.4v - as such I need to find witch jumper sets that...

VIP596P3 i think that mean its having 3 VLB slots on the board

VIP596P2 this saying it having 2 VLB slots on the board.

REV is just the reference design version number of the board.. This means a higher number is a changed motherboard design.

Robin4 wrote:

VIP596P3 i think that mean its having 3 VLB slots on the board […]
Show full quote

kanecvr wrote:

It's posible it's rev 2.0 of V5P596P2. I figured out witch jumpers set the FSB, so I'm now able to set the thing to as low as 20MHz or as high as 75MHz (not stable at 75, won't boot) 😀 I think I can get 286 speeds out of it at 20MHz with cache disabled - but I'm having some issues with L2 cache. It's a great board for simulating a 486 with a fat FPU or a 63MHz socket 3 POD 😀

The thing is I remember there were 2 versions of the socket 4 pentium - one running at 5v and one running at some other voltage 5.2 or 5.4v - as such I need to find witch jumper sets that...

VIP596P3 i think that mean its having 3 VLB slots on the board

VIP596P2 this saying it having 2 VLB slots on the board.

REV is just the reference design version number of the board.. This means a higher number is a changed motherboard design.

Some jumpers are missing on my board - completly - not even the pads. JP4, JP5, JP7 for selecting internal or external clock for example... FGB stated that I might get better stability by selecting external clock, but it seems I can't since the jumpers for it are missing... Also, jumper layout is slightly different. The only jumpers that are where they are supposed to be are half of the cache selection jumpers and the VLB WS jumpers.