Appeared to have fixed my Abit/G400 Max lockup problem

[ALBPM]

Here's a pic of where the contacts for the two wire fan are located.


This next pic shows what I believe could be the problem for some of us. Have you noticed how hard it is to get the end of the bracket to fit down between the edge of the mobo and the case. This is why I bend the end of the bracket in addition to the top of the bracket.


Sorry Chuck I'm not going to attempt to draw that one

Paul

[This message has been edited by ALBPM (edited 10-09-1999).]

[bcronin]

Hmmm, interesting. Are the contacts for the 3 wire fan in a different place? I don't recall having had any special difficulty getting the G400 Max seated in my AGP slot. From your pic it looks like you have a socket 370 based board. There's plenty of room in my Inwin Q500 tower case and the BX6R2 mobo isn't especially cramped in. Alas (what am I saying? ;-) I can't test this theory out since I already returned my "bad" 2 wire fan G400 Max. The 3 wire fan version is working just fine for me with no special tweaking of the card, bracket, etc. Got a theory to explain why? Hey Haig, can I have the bad one back? I'll try it in my son's machine and let you know if it works or not ;-)

------------------
Cheers,
BC

[ALBPM]

Hey bc,

What your seeing as a socket 370 is about 1/8 of an Alpha P3125 on a P3-550. I'm using an ABIT BE6 motherboard.

Paul

------------------
P3-616(112x5.5), ABIT BE6, 256MB PC100, MILL G400 MAX-32MB on BNC, Quantum3D Obsidian2 X-24 on D-sub, IBM Deskstar 18GB ATA/66, 21"Trinitron, SCSI CDR/RW, SCSI ZIP,SB Live!Value,USR Voice Faxmodem Pro(USB Ext),Epson 1520 printer,Umax1200S scanner,WN98 and a (Very Messy Desk) ;-P

[Haig]

Hi guys,

After seeing Paul's diagrams last night, I asked a few clients to do pull out our card and put it back in. But this time, do not screw in our bracket into the case. Every client that emailed me back has confirmed they have no more lock ups. 2 clients have also confirmed the same but they are running more tests.

I have sent an email to my contact in Abit europe to see how deep their AGP socket is because someone is out of spec.

I have confirmed that our brackets are all in spec.

This leaves the following:

- The depth of the AGP slot in the Abit mobo's.
- The height of the risers where the mobo's sit on.
- The holes in the cases themselves are off by a tad.

I think that trying to get any kind of answer from the case manaufacturer's is a lost cause but you guys can try if you like.

Thanks to Paul and all others who were emailing tips after tips after tips

BC, pin 3 and 5 are the same gnd and vcc pins for all G400 cards.

If you want your old board back I'll trade ya

Haig

[This message has been edited by Haig (edited 10-09-1999).]

[xortam]

Haig, If you give me you're E-Mail address, I'll send you my RMA# and E-Mail log from my debugging session with Matrox Technical Support. The log will describe some of the experiments I tried. Since I still have both "different" Maxes, I'll pull out the new one and compare the form factor on the two. I'll look for some of the things various users have been reporting.

Oh yes, and what is the client ID# you refer to? I have the order #.


[This message has been edited by xortam (edited 10-09-1999).]

[Rob M.]

You can make that 3 clients are running more tests. Strange thing is, I can wiggle it around, push it down (hard), even rock it back and forth in the slot, while it's on even, and I havn't had a crash in about 4 hours. I'm not quite ready to eat my words from my previous post yet though, I've had the card work fine for 4 days without this crash before.

One last problem for me to solve then, is there anyway for me to order a new fan? Mine's LOUD (much louder than all other fans in my system put together), and it's driving me nutz!

BTW. Love those q3 results with 5.30. Hope to see these optimizations worked into all games soon(customers are NEVER satisfied )

[Zypher]

LOL...its boiling down to a case problem.

Lots of cheap *cough* noname *cough* finger-slicer cases have very tight fits, the good ones have some leeway in the screw positions (esp for those bigass dual boards) and/or use plastic mounts. (theres a reason all the bigwigs use some/all plastic mobo mounts, among other things)

Never skimp on cases or monitors...most of the time they outlast 2 or 3 computers

[Whisper]

Well I did mention to everybody in this forum that this was possibly the problem in this thread
http://forums.murc.ws/ubb/Forum5/HTML/004858.html

Not that many people seemed to pay any attention to it, what can you do though??

"You can lead a horse to water, but you can not make it drink"

SteveC and Alfie also made very good points to, that may also help.

[This message has been edited by Whisper (edited 10-09-1999).]

[MultimediaMan]

This is reminiscent of some problems experienced with Gigabyte Motherboards and Matrox Adapters about a year ago.

Another thing to try: assuming your case is a little out of whack, loosen all of your mounting screws on the motherboard, and insert the AGP card, making sure it is fully seated. Retorque all of the motherboard mounting screws evenly and everything should be wonderful.

Don't you just love the joys of custom computer systems? Don't you wish everyone did?



------------------
Junior Member

[Scytale]

Hi guys. It looks like we have screwed up where we shouldn't have. After Haig mailed me with his suggestion, I seated the card without the screw and clocked my P2 to 4x100 (to avoid other problem factors). And behold... no lock ups for hours, not even with Expendable, my usual crash bench.

I ran some more tests and I found that besides the seating of the card being a factor, another factor still was (AGP) bus speed. As long as I kept my AGP at 66ish (x2) I didn't get lock ups, an FSB of 112 however was enough to get lock ups again (but only with Expendable, so it seems).

Of course I have tested all this FSB stuff before, but not screwing up made the difference.

At this time I'm running at 4.5x100, AGP2X+BM, and everything seems to be hunky dorey, including Expendable. Using my old MAX... since the DHL courier screwed up on the delivery of my replacement card.

Anyway, I think a big congrats in the direction of Haig and Paul is due! Two and a half months of puzzling seems to have paid off!!!

So... Thanks guys! I'm sure my next card will be from Creative again...

Erm...

[Scytale]

LOL, join the club. I took the card out and it appears I can't get it seated correctly again. Once more the damn thing is kicking like a mule.

It was fine until I touched it. I have played a game of reseating and booting for over twenty times now, but I just can't get it right any more... so for now I'm back at AGP1X. Doh.

[Rob M.]

heh, I havn't touched a thing since I got mine working, for fear that it'll stop again.
Anyone who got the replacement card to fix their problems wanna try re-inserting their card a few times to see if they can get the lockup problem again?

After all, if this was the problem, then people should've been reporting it with every G400, not just with the metal-bladed ones.

[This message has been edited by Rob M. (edited 10-09-1999).]

[xortam]

Haig, I compared my two Maxes this morning and they appeared to have identical form factors and bracket allignment... except: The Canadian board seems to fit a bit more snugly than the Irish board. Also, Irish board has some abrasion (running horizontally/widthwise) on the first guiding slot tab (no contacts, opposite edge from bracket end); the older Canadian board is unmarred. Again the Irish board fixed my lock ups. I'll send you the details (client id# ?) when I get you're E-Mail address. BTW, The ASUS P2B-S motherboard is housed in an Intel OEM XEON case.

[urtho]

I am also have the same problems with my vanilla G400-32MB. It runs fine with AGP1X. I have tried everything suggested here and the lockups are less frequent but they still happen. I have a BX6R2 and a SBLive. I don't have any problems when I put my old AWE64 back in. The G400 is on IRQ11 and the SBLive is using 5 and 10. The SBLive is in PCI slot 4. I thought I would post here because it would seem that it is mostly the MAX with the problems but that isn't true.

[Markus Eisenstoeck]

I did my fair amount of tests with my G400 32DH and different positions in the AGP slot of my BH6 during this evening.

Depending on how I bend the metal end of the G400, I can reproduce everything from blinking bits on the screen, no working opengl at all, only working with busmaster off, working only with AGP 1x and finally running with AGP 2x ...

I've run a timedemo loop in quake 3 for the last 2 hours on my p3 500@550 and have not experienced one single lock up.

The only thing that still hangs after about 5 minutes is ssystem (http://www1.las.es/~amil/ssystem), an opengl solar system simulator. But I guess this is more of a driver problem, because it does freeze the display, but I can reset with ctrl-alt-del, something not possible when q3test locked.