PC Hoarder Patrol wrote on 2024-11-05, 08:19:

Late, but not the latest...see […]
Show full quote

FLareonlover139 wrote on 2024-11-05, 00:20:

also the board version is AA 750780-608, where in the timeline is that for revisions

Late, but not the latest...see

The attachment OR840 ProcSupp.pdf is no longer available

The -701+ revisions introduced support for 512MB RIMMs

so it is not one of the early buggy ones, also I am fine with it not being super fast as I have an overclocked 2.52ghz athlon XP with a phasechange cooler, 2GB of ram and a 7800GS so a slower dual P3 works for me

[/quote]

well after inheriting there board, I thought it might be something simple but my first burn in test found some very slight instability after running for 10 hours (memtest86+ v5.01 crashed after 10 hours with a 4X256MB pc800 40ns ram kit a single 733mhz P3 and a terminator) so I ordered the caps from the shopping list and will be recapping it once the main caps come are off backorder
[/quote]

Well - 10h Burn-in test i would already consider as fully working.
Considering the Age - not sure if 10h is actually such an good idea - could be PSU, GPU - ofc the agin condensators - even CPU
+ its a Rambus Board -they all were a bit ...at the limit already back then - hot, fast = can be instable either way.

With the Revision - yes, puts you bit midrange, bit lower - might not be either way a perfect stable one anyway.
Could test some other Rambus vendors/modules - might not even be able to run with 512er's or P3 1000.
And even if, might be instable - my Orig. Or840 does run currently best with Elpidas - its weird - several Samsungs it wont even start - Infindeon and Kingston work, but ..something feels off - with the Elpida it runs nonstop.

Well good luck - I still didnt forgot my experiment with turning an IBM Or840 into an orig one - still didnt sacrifeced the one IBM one i have..(yet) ^^

freekymachine wrote on 2024-12-02, 13:22:

well after inheriting there board, I thought it might be something simple but my first burn in test found some very slight instability after running for 10 hours (memtest86+ v5.01 crashed after 10 hours with a 4X256MB pc800 40ns ram kit a single 733mhz P3 and a terminator) so I ordered the caps from the shopping list and will be recapping it once the main caps come are off backorder
[/quote]

Well - 10h Burn-in test i would already consider as fully working.
Considering the Age - not sure if 10h is actually such an good idea - could be PSU, GPU - ofc the agin condensators - even CPU
+ its a Rambus Board -they all were a bit ...at the limit already back then - hot, fast = can be instable either way.

With the Revision - yes, puts you bit midrange, bit lower - might not be either way a perfect stable one anyway.
Could test some other Rambus vendors/modules - might not even be able to run with 512er's or P3 1000.
And even if, might be instable - my Orig. Or840 does run currently best with Elpidas - its weird - several Samsungs it wont even start - Infindeon and Kingston work, but ..something feels off - with the Elpida it runs nonstop.

Well good luck - I still didnt forgot my experiment with turning an IBM Or840 into an orig one - still didnt sacrifeced the one IBM one i have..(yet) ^^
[/quote]

PSU was known good and the GPU was a rage 128 also known good, cpus are a pair of 733s and it is stable but only if SMP is not used, so it is an 840 that cant run 2 CPUs

if I try and run any version of memtest in SMP mode it crashes right to the bios boot vector

Hi all. Does anyone have a bios dump for or840 from ibm? I will be glad if you share.

Cool.Tweak wrote on 2025-02-02, 19:05:

Hi all. Does anyone have a bios dump for or840 from ibm? I will be glad if you share.

Try here... https://theretroweb.com/motherboards/s/intel- … -outrigger#bios

PC Hoarder Patrol wrote on 2025-02-02, 21:51:

Cool.Tweak wrote on 2025-02-02, 19:05:

Hi all. Does anyone have a bios dump for or840 from ibm? I will be glad if you share.

Try here... https://theretroweb.com/motherboards/s/intel- … -outrigger#bios

He's not there. You need a dump, not a file update. For hardware firmware. The BIOS from the regular or840, which I extracted from the second board, is not compatible with the board from IBM.

Cool.Tweak wrote on 2025-02-03, 00:04:

PC Hoarder Patrol wrote on 2025-02-02, 21:51:

Cool.Tweak wrote on 2025-02-02, 19:05:

Hi all. Does anyone have a bios dump for or840 from ibm? I will be glad if you share.

Try here... https://theretroweb.com/motherboards/s/intel- … -outrigger#bios

He's not there. You need a dump, not a file update. For hardware firmware. The BIOS from the regular or840, which I extracted from the second board, is not compatible with the board from IBM.

Not sure I follow, so maybe check that link again. At the bottom of that page, IBM BIOS PLJT44A / PLKT44A is available to download which is the latest version available for the IBM OEM version of the OR840 and is the one I'm currently using on my board (check the pics on the left and you see my actual board - is this the same as yours?)

You can see the IBM support page for this BIOS release here... https://www.ibm.com/support/pages/flash-bios- … -type-6868-6878

Help Needed!

I need you to use a multimeter and measure continuity between the 12V ATX pin and AGP pin A1.
Measure between the yellow square in the AGP port, and either of the other yellow squares in the picture below.

Background
My OR840 has developed a nasty problem with its AGP slot: I'm unable to use most AGP cards.

At first I thought it was related to AGP signaling and proper voltage generation—AGP 1.0 cards run at 2x and use 3.3V, while AGP 2.0 cards can run at 4x while using 1.5V.
However, taking measurements right at the MOSFET (signaled in blue in the picture), I saw proper voltages when using a Voodoo3 (3.3V card), when using a GeForce2 (1.5V card), and when using a Radeon8500 (1.5V card)
So the voltage switching and generation seems to be working properly. I measured VREF for each card as well, and its within the expected ranges for each of them.

The plot thickens:

• The Voodoo3 gives me video and the system POSTs properly.
• The GeForce2 doesn't give video and I don't hear any POST beeps, so the system doesn't POST. Also, the fan on the card doesn't spin.
• The Radeon8500 doesn't give video but I hear POST beeps, the fan on the card spins just fine.

The fans gave me my next clue: the Radeon powers it from 5V, while the GeForce powers it from 12V.
When I checked continuity between the ATX 12V pin and the A1 AGP pin, there was none! In every other motherboard I have, there is a direct connection.

With the system running, I measured the voltage in the A1 AGP pin with the different cards. Here are the results I got

• Voodoo3 --> 0.01V
• GeForce2 --> 1.26V
• Radeon8500 --> 3.75V

So right now I have the following question before I can take action:
In the OR840, should there be a direct path from ATX 12V pin to A1 AGP pin, or are there elements in between?
Tracing proved impossible, so I need someone to help me!

• If there is a direct path, then I'll solder a patch cable, with a fuse just in case.
  --> In this case, the voltages measured with each card are just the result of the circuit in the card doing funny things without the 12V supply.
• If there isn't a direct path, then I'll need further help finding the faulty element.
  --> In this case, there is a faulty element in the path between ATX 12V and AGP A1, and the measured voltages are due to different load from each of the cards

Thanks in advance, everyone!

Took a chance on this based on the following

• All my other boards have a direct connection between ATX 12V and AGP 12V
• The only components that would make sense to put between ATX 12V and AGP 12V would be fuses
• Different voltage readings from the 3 cards above suggest voltage is being backfed from other circuits into AGP 12V
• Checking the schematics of the i840 dev kit, there is a direct connection between ATX 12V and AGP 12V
• I'm kinda impatient to wait for responses

I ended up soldering a small 24 AWG solid-core cable from the 12V pins in the AGP Pro part of the slot into the A1 pin in the normal AGP part of the slot. I added a 400 mA resettable fuse just to be "safe".

Cards are now working properly!
I guess testing a broken GeForce3 a couple of years ago vaporized the trace, I only just realized it was dead because I had a Voodoo3 in there since then.

I dont see any damages in the outer layers of the board, or any burnt elements. I wish there were a fuse I could just change instead of the patch at the bottom of the board, but for now Im just glad I diagnosed it correctly

Im also wondering if 400 mA are enough to power some of the hungrier cards that dont have external power connectors. Ill run some benchmarks with the Radeon 8500 and maybe a GeForce4 TI 4600.

I didnt have access to my OR840 boards otherwise I would have checked earlier. sorry.

Do you think the 4-pin connector is just bad? I dont recall having an issue with mine in the past.

luckybob wrote on 2025-04-23, 23:45:

I didnt have access to my OR840 boards otherwise I would have checked earlier. sorry.

Do you think the 4-pin connector is just bad? I dont recall having an issue with mine in the past.

hey luckybob, dont worry about it, thanks for always checking this thread!

You mean the molex tagged "AGP AUX PWR" on the board? Nah, every pin/path that should've been connected to 12V was connected to it, except AGP A1.
PCI pins were connected (A2 in the PCI connectors), the CPU VRMs were connected, the AGP Pro 12V pins were connected, and the AGP AUX PWR 12V molex pin was connected.

Only AGP A1 was severed, I think this fits the case of a bad GPU frying the line by requesting too much power through it.
Nothing but a flesh wound.

Speaking of the CPU VRMs, I took some pics of it to post in another thread, I'll add them here for reference as well.
Mine are Delta based.

Mine looks the same as I recall. There is a fuse in the top left, I'd give it a once over, if it was me.

It acts like a blown fuse, I done see an obvious one on the Pic. Hmm.

slivercr wrote on 2025-04-23, 21:25:

Took a chance on this based on the following […]
Show full quote

Took a chance on this based on the following

• All my other boards have a direct connection between ATX 12V and AGP 12V
• The only components that would make sense to put between ATX 12V and AGP 12V would be fuses
• Different voltage readings from the 3 cards above suggest voltage is being backfed from other circuits into AGP 12V
• Checking the schematics of the i840 dev kit, there is a direct connection between ATX 12V and AGP 12V
• I'm kinda impatient to wait for responses

I ended up soldering a small 24 AWG solid-core cable from the 12V pins in the AGP Pro part of the slot into the A1 pin in the normal AGP part of the slot. I added a 400 mA resettable fuse just to be "safe".

The attachment 20250423_151857.jpg is no longer available

Cards are now working properly!
I guess testing a broken GeForce3 a couple of years ago vaporized the trace, I only just realized it was dead because I had a Voodoo3 in there since then.

I dont see any damages in the outer layers of the board, or any burnt elements. I wish there were a fuse I could just change instead of the patch at the bottom of the board, but for now Im just glad I diagnosed it correctly

Im also wondering if 400 mA are enough to power some of the hungrier cards that dont have external power connectors. Ill run some benchmarks with the Radeon 8500 and maybe a GeForce4 TI 4600.

Great job. Hopefully, what internal trace that may be vaporized haven't damaged any adjacent traces? But running flawless for years implies that it hasn't.

slivercr wrote on 2025-04-24, 01:15:

Speaking of the CPU VRMs, I took some pics of it to post in another thread, I'll add them here for reference as well.
Mine are Delta based.

Is this VRM supplying both CPUs or it's one supplied directly from the motherboard and the VRM only used when running dual CPUs?

SWZSSR wrote on 2020-06-05, 02:05:

Hey guys […]
Show full quote

Hey guys

I recently acquired this OR840 Engineering Sample System in Adelaide, Australia. My friend saved it from the crusher!

I hope to find what original expansion cards & hdd it had. PSU will need recapping as its shot.

Photos

Hi All... after a long long time I finally got the OR840ES System going.

Specs
2x600B Katami 133FSB (Was installed when found)
2x128MB PC600 Samsung RIMMS
TNT2 M64 Originally? Placed a G400MAX Inside (Date Codes Match)
LITEON PSU Fully Recapped with Service\Removal of Hotsnot Glue

Installed W2K. Did an ACPI Installed. Holy smokes that was slow!

Bios is very very early - P01.9910011410 !!!!

I can confirm only 128MB or less PC600 RIMMS work! Memory beep with anything larger or faster...

I can post more below if people are interested. I dont want to hijack silvercrs thread though!

H3nrik V! wrote on 2025-04-24, 04:15:

slivercr wrote on 2025-04-24, 01:15:

Speaking of the CPU VRMs, I took some pics of it to post in another thread, I'll add them here for reference as well.
Mine are Delta based.

Is this VRM supplying both CPUs or it's one supplied directly from the motherboard and the VRM only used when running dual CPUs?

Sorry for slow reply. From what I can tell, this supplies both CPUs and there's no way to start the board without it.

SWZSSR wrote on 2025-04-29, 14:21:

Hi All... after a long long time I finally got the OR840ES System going. […]
Show full quote

SWZSSR wrote on 2020-06-05, 02:05:

Hey guys […]
Show full quote

Hey guys

I recently acquired this OR840 Engineering Sample System in Adelaide, Australia. My friend saved it from the crusher!

I hope to find what original expansion cards & hdd it had. PSU will need recapping as its shot.

Photos

Hi All... after a long long time I finally got the OR840ES System going.

The attachment internals.jpeg is no longer available

The attachment cpuinfo.jpeg is no longer available

The attachment raminfo.jpeg is no longer available

Specs
2x600B Katami 133FSB (Was installed when found)
2x128MB PC600 Samsung RIMMS
TNT2 M64 Originally? Placed a G400MAX Inside (Date Codes Match)
LITEON PSU Fully Recapped with Service\Removal of Hotsnot Glue

Installed W2K. Did an ACPI Installed. Holy smokes that was slow!

Bios is very very early - P01.9910011410 !!!!

The attachment biosver.jpeg is no longer available

I can confirm only 128MB or less PC600 RIMMS work! Memory beep with anything larger or faster...

I can post more below if people are interested. I dont want to hijack silvercrs thread though!

Yooooo, good to see you! Hijack? Show us more!

Can you please try to back-up the original BIOS? I dont think we have many versions archived.

AA 750701-603 is Up and running, quick show off...

Next Posts is Everest

Everest 5.50 Tests

Windows 2000
- Intel OR840 AA 750701-603
- No Service Pack
- ACPI Install (Auto)
- 1999 ES Bios - Intel-OR840600.86E.0122.P01.9910011410
- Inbuilt Chipset & IDE Drivers of W2K
- 2x600B Katami 133FSB
- 2x128MB PC600 Samsung
- G400MAX
- Terratec DMX

Performance is BAD. Whole system runs poor, very unresponsive..I believe this is due to very poor ACPI compatibility on the engineering sample bios.

Next Test - Windows XP Service Pack 2

Windows XP Installed as MPS - Multi Processor System instead of ACPI Multi Processor
- Intel OR840 AA 750701-603
- Stock Drivers, Nothing Installed or Updated
- No Service Pack
- MPS - Multi Processor System Install (Auto)
- 1999 ES Bios - Intel-OR840600.86E.0122.P01.9910011410
- 2x600B Katami 133FSB
- 2x128MB PC600 Samsung
- G400MAX
- Terratec DMX

Memory Performance is MUCH better!!!

Next Tests is W2K\XP again with Latest BIOS only