VOGONS


First post, by feipoa

User metadata
Rank l33t++
Rank
l33t++

I have been having the most difficult of times getting an ATI FireGL X1 256M APGro card working. I have had 4 cards in total. The test boards are a Tyan dual Tualatin and a Supermicro dual Tualatin.

1st card - screen shows artifacts straight away at power on, but can still view BIOS, DOS and Windows.

2nd card - doesn't show any artifacts at power on, but after about a minute in Windows (with drivers installed), artifacts start to appear. Over the course of a few minutes, the quantity of artifacts on the screen increases.

3rd card - the screen stays blank at power on, however I can hear the computer booting to a floppy, the HDD, etc. If I use a DVI to VGA adapter, the card will display the first second of the screen when it comes to life, e.g. the point just before memory count starts. However, the screen stays frozen on that screen and the computer continues to boot with this frozen screen.

4th card - exactly the same outcome as the 3rd, except I have to give the fan a little kick with my finger to start spinning.

Cards 1/2 are from the same source. Cards 3/4 are from the same source, but not the same as cards 1/2. It is interesting how cards from the same supplier exhibit similar behaviours. Are there slight variances in OEM versions of these cards? I was under the impression that all of these were Dells, that is, only DELL had the AGPPRO FireGL X1 in 256M.

Does anyone have any idea what is going on with these? Has anyone been able to get one of these cards working? I suspect there isn't a compatibility problem because card #2 worked just fine, albeit with some artifacts after time. It was able to run 3DMark2001SE just fine.

Cards 1&2 had been returned. Cards 3/4 I still have in possession. Why do I want one of these cards? From a historical perspective, I think it would be neat to have at least one working AGP PRO card in my ultimate dual Tualatin setup.

FireGL_X1_256M_AGPPro_top.jpg
Filename
FireGL_X1_256M_AGPPro_top.jpg
File size
1.18 MiB
Views
1658 views
File license
Fair use/fair dealing exception
FireGL_X1_256M_AGPPro_bottom.jpg
Filename
FireGL_X1_256M_AGPPro_bottom.jpg
File size
1.67 MiB
Views
1658 views
File license
Fair use/fair dealing exception

Plan your life wisely, you'll be dead before you know it.

Reply 1 of 10, by The Serpent Rider

User metadata
Rank l33t++
Rank
l33t++

Have you tried more beefy motherboard? Like something based on i875P or even workstation Xeon.

I must be some kind of standard: the anonymous gangbanger of the 21st century.

Reply 2 of 10, by feipoa

User metadata
Rank l33t++
Rank
l33t++

I'm pretty sure that my dual Tualatin boards (based on HEsl chipset), Tyan Thunder S2567 and SuperMicro P3TDE6, are the only boards I have which sport an AGP PRO slot. I have an Intel S875WP1-E motherboard, but it only contains a regular AGP slot.

One other interesting factor about the second pair of FirePro cards is that they both came with a 4" cable connected to that small white PCB-bound connector you see in the photo. The other end was left dangling. I'm not sure what the other end is supposed to connect do. Do you know? The first two cards didn't have this cable, and the first two cards appeared to sort of work, albeit with artifacts.

Plan your life wisely, you'll be dead before you know it.

Reply 3 of 10, by The Serpent Rider

User metadata
Rank l33t++
Rank
l33t++

I'm not sure what the other end is supposed to connect do. Do you know?

Most likely external TV in/out support. DIN connector on the card is reserved for 3D stereo glasses.

Without another motherboard it's hard to tell. But it's possible that all cards are in zombie state. Unfortunately all R300/R350 based cards were very prone to suddenly die or malfunction. Some are victims of the Samsung BGA memory (Radeon 9700 and FX5800 series notoriosly) and some just died due to problems with early GPU BGA flip-chip design.

I suspect all cards also came with horrible yellow rubberlike thermal padding under the cooler? This may explain why 2 cards just freeze from the start - they overheat.

Last edited by The Serpent Rider on 2018-12-15, 18:10. Edited 1 time in total.

I must be some kind of standard: the anonymous gangbanger of the 21st century.

Reply 4 of 10, by feipoa

User metadata
Rank l33t++
Rank
l33t++

Serpent Rider, I appreciate your valuable insight. My gut tells me its not a motherboard issue (I realise that is not very scientific). I suspect both conditions you mention may be at play. The first pair of cards have fairly common memory related artifact issues. The second pair of cards may be the same, or also the heatsink issue you describe. I will remove the heatsinks and inspect the thermal padding.

Plan your life wisely, you'll be dead before you know it.

Reply 6 of 10, by vlask

User metadata
Rank Member
Rank
Member

They just dying, same here. Had one working in collection. After 1 year i took it out to retest it - dead, screen artefacts.
So i got 2 more cards, of them one dead too, second working (so far). They are just bad made. I tried to push by my fingers memory modules when card was working and found out, that it changes style of screen artefacts. So i guess that memory modules are dying.....
At last one nice video....
https://youtu.be/tQEhWCc45Ko

Not only mine graphics cards collection at http://www.vgamuseum.info

Reply 7 of 10, by feipoa

User metadata
Rank l33t++
Rank
l33t++

This is useful information, thanks guys. I was unable to remove the heatsink without risking damage. I intend to return the two cards, so I didn't want to push it.

Sounds like these cards in particular have passed their shelf life. This is unfortunate. I really hoped to get a fairly high-end AGP PRO card in this system that can also do games. I'm aware that there are Wildcats, but those don't game. I guess I'll have to be content with my Quadro FX3000 in my dual Tualatin w/dual channel U320 SCSI RAID and dual channel memory. I fear that even if I found a working FireGL X1 256M AGP PRO that it would stop functioning in the coming years.

Plan your life wisely, you'll be dead before you know it.

Reply 8 of 10, by vlask

User metadata
Rank Member
Rank
Member
feipoa wrote:

I was unable to remove the heatsink without risking damage. I intend to return the two cards, so I didn't want to push it.

I tried to remove it on first dead card in try to reflow main chip. At the end i removed it, but with small resistors around gpu that was glued on orange something supposed to be thermal paste....it was rock solid glued to chip.....
Dont try, youll end with damaged gpu like i did. Just try push each memory chip from both sides at once to pcb by fingers and watch if artefacts on screen changes, if yes, youve got your bad memory.....

Not only mine graphics cards collection at http://www.vgamuseum.info

Reply 9 of 10, by feipoa

User metadata
Rank l33t++
Rank
l33t++

The pair of cards which had artifacts have been returned. When I had them in possession, I did try pressing the memory chips, but nothing changed with the artifacts.

The second pair of cards I currently have in possession do not contain artifacts. The screens stay blank.

Plan your life wisely, you'll be dead before you know it.

Reply 10 of 10, by feipoa

User metadata
Rank l33t++
Rank
l33t++

Received refund for the two cards. Seller didn't want them sent back. Is there any salvageable parts on these? Maybe the DVI connector? They are pretty bulky to store in my broken PCB box, so might be headed to e-waste.

Plan your life wisely, you'll be dead before you know it.