ABIT motherboard compatibility

[Haig]

Hi Chuck,

Can't explain what the problem is. Some people have solved it with different type of solutions and the majority of the clients with the lockups are still locking up.

We even got a call today from a client who had fixed his lock up problems by disabling bitmap caching. So we can add this to the list of possible solutions I guess

I was hoping that we would be able to reproduce this problem properly with BC's board but do you think we were able to?

Nope!!

Right now, we just have nothing but guesses and a bunch of "try this try that..."

For all I know, it could be related to the light bulb in your room BTW, that has happened to 2 clients in the UK!!

Clients can still call tech supp and ask for an RMA if they wish.

Haig

[bcronin]

Sheesh. Sorry to hear you were not able to recreate the problem, but there is definitely >something< about that board, because the replacement board I got hasn't locked up once yet, and I've been pushing it *real* hard. With the old one I barely had to breathe and it'd lock up. Since so many seem to be having similar problems (and since the new board is working fine for me), I tend to doubt that there is anything inherently wrong with all of our systems and that there is still some sort of problem with the affected Max's (most likely a timing issue of some sort given the symptoms). I wish I knew more about hardware so I could help you find it (I'm great at finding timing-related problems in software ;-). In my opinion your best bet to get this solved would be to get someone with a failing Max/machine to lend you their entire machine so you could set it up in your test lab for analysis. The offer I made to Alex (to let you guys come here to my house with a team of testers) is still open so if you get really desperate you can bring my old board back here and have a go at it, but otherwise I can't afford to not have my machine here, since I work from home most of the time.

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

[Haig]

Hi BC,

I will give myself until Monday to reproduce this or else I will get someone's Max and mobo in exchange for a new set. Don't have much of a choice on this one. I agree that something is wrong with your 1st card but not in our systems for some reason. Right now, we left a few Abit systems running 3D and OGL demos in a loop over the weekend.

Maybe this is our problem or a combination of us and Abit or something else.

I was reading a forum on the BP6 and it looks like this motherboard also has a share of its problems with lock ups, mobo and some cases with ground problems, and power supplies. You can check it out at
http://www.bp6.com/discus/

Anyway, nothing is still for sure yet.

Haig

[ALBPM]

I just helped a friend build his new system today. It has a P3-500, ABIT BX6 rev2, G400 MAX, etc..
He has it in an Enlite case with a 235W power supply. No problems so far. His MAX was made in Ireland but has the two wire metal bladed fan.
He upgraded from a pentium 200 and was totally blown away by his new P3-500 system.

Paul

[dwright]

G400max testing continued:
1. I removed the 3rd stick of memory and there was no change to the lockups. The memory has behaved and worked without problems for over six month's. I agree that memory can cause these problems and I did have to try various brands until I got chips that worked on my bh6. I was worth a try. Thank you cjolley.

2. I removed the sblive (had it's own irq5). This made a change to the number of lockups. Where I would have had 5 or more I now get only 1. Before removing sblive I could predict the lockups. Now that is not possible.

My results so far seem to show sblive contributes to the lockups but it is not entirely at fault (sblive just magnifies the problem). With my old card a fury 32meg I did not get lockups. I don't want the old card back.

I could to unplug the g400max fan and try some tests when the computer is cool. What do you think? Has anyone tried this?

[bcronin]

Ok Haig, but I could recreate it for you in less than 5 minutes. I note someone else seems to suspect that their Sblive might be contributing. As it happens, I have an Sblive as well. I did not try removing it (and the replacement card works fine with the Sblive, so I really doubt the Sblive has anything to do with it per se). If I can help in any way please don't hesitate to call on me. Although my new board is working great, I am contemplating getting another one for my son's machine. It is also based on an Abit board (the BM6) with a Celeron 433
and so, I'm a little gunshy. I'd prefer to know that the problem has been identified and solved before ordering another one.

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

[asiriann]

Haig,
I was one of the people experiencing the lockup issue on a ABIT BE6(see previous posts). I checked the voltages with Sandra and MBM4. Here is what I've got;

Core 0 Voltage :2.00
Core 1 Voltage :N/A
+3.3 Voltage :3.28
+5 Voltage :4.97
-5 Voltage :-2.98 - This concerns me!
+12 Voltage :11.80
-12 Voltage :11.86

Hope this helps in some way. Now I need to figure out how concerned I need to be over the -5 volt issue.

Tony

[asiriann]

Nevermind! Just found out that the -5 cannot be read on the ABIT BE6.

Tony

[Scytale]

Hi guys. I'm sure in some cases the SBLive is a lock up factor. I'm also very sure that in other cases it's not: I have tested with all cards removed, and all integrated peripherals (except for a single EIDE channel) disabled, only to find that the system would lock up even sooner.
Ahem, not to mention that I don't even own an SBLive...

And just to let you guys know: Last week Haig sent me a plastic bladed (3-wired) fan to try, but I found it didn't improve things a bit.

I'm as puzzled as the next guy, and like BC I offered Haig my system components for exchange. Maybe it will not be necessary, I guess we'll know by Monday...

[Star F.]

Haig,

IMHO it doesn't make much sense to have known critical G400/mobo combinations run some demos for days. They may pass simply because no critical situation ever occurs in these demos.

All we have seen so far on this forum seems to point in the direction of a manufacturing or component defect in early-production G400s. Some part might be at the limits of electrical tolerances and if the mobo has such parts as well, then, under certain conditions, a lockup occurs.

I doubt that only ABit Boards are concerned. They're probably the worst quality, but Gigabyte isn't much better, at least not my 686BX. And almost every other brand name has appeared in the lockup reports here. ABit is just the most frequent.

With the Gigabyte 686BX, I have no lockups when starting D3D apps, they come at random points (it seems) when the app is already running. It makes no difference whether the machine is totally cold or has been on for hours.

The lockups seem to occur randomly, but maybe there's a hint. Some of my games, as well as the Techo Demo, have never locked up yet. Others do within 10 minutes a the most. You cannot really predict them, but most lockups do occur when the cpu has a relatively high burden in addition to its geometry tasks. When there are many actors, for example, or unexpected events. Or, typically, in racing games when a new lap is about to begin.

It may well be that turning off the bus mastering only reduces lockup-prone situations to almost zero.

ANYWAYS:
It's about time Matrox made it clear whether there is a problem in the hardware of some G400s only or a logic-related bug that might affect all G400s. In the first case they should start a call-back campaign, in the second they should try to fix the problem with a new BIOS or driver. And, if it can't be solved this way, if it really is the hardware design - then they will have to do a redesign and offer a new card to anybody who claims one. Whatever has to be done, the sooner they act, the smaller will be the number of customers they loose (they're losing already).

And you guys at Tech Support better use the real games that were named in numerous posts here. Which should also give you more job satisfaction.


-------------------
p2-450, Gigabyte 686BX (FSB @100), 384 MB pc-100, Adaptec 2940UW, Accton EN1207D-TX Fast Ethernet Adapter, Diamond S90 (A3D, Aureal 1), G400 SH 32MB

AGP 1x doesn't help; removing network adapter + sound card doesn't help; might be that devil of a SCSI adapter (which I can't do without, no IDE at hand), but then again, it's the most standard component I have.

[kooguy]

If you get lockups in games, I suggest you install the drivers *and* DirectX over again. I haven't gotten a lock up in games since. But there still lies another problem. When I surf the net, sometimes there is a complete lockup. Usually when I am using Netscape although it happens with IE too but just not quite as often. I have to Ctrl+Alt+Delete or sometimes hit the reset button after that. That is basically the only problem I have left.

[rinkol]

Haig

You mention the Abit BP6 as also having a tendency to have problems with the g400 max. This would not be surprising since it is very similar to the BE6; I am pretty sure that hte only differences, aside from a few more front side bus speeds, are directly associated with provision for the dual processor. One thing that would be common to the BE6 and BP6 and different from other mother boards is the inclusion of a controller (HPT366) for ata66 hard disks. Conceivably, this could be a factor, possibly as subtle as slightly slowing some timing signals by introducing a bit of extra loading or modifying some of the handshaking involved in bus arbitration. I have the impression that some lockup problems reported by others are correlated with SCSI adaptors. If this is correct, then these cases might be related.

One thing that is not clear to me about Matrox' trouble shooting approach. If substituting a different G400 fixes the problem (reported by at least one poster), then it would seem to me that it would make sense to examine the characteristics of that (and other failing) g400s in detail and search for differences relative to other g400s that work. I would not be surprised if there were common factors such as a crystal frequency near the tolerance limits, or video logic chips showing speed behaviour near the fast or slow tolerance limits. Can the problems be usually fixed by replacing the g400 or not???

While I don't have a g400 max, I do have a g400/bp6 combo that is giving me headaches with lockups (I have reported this to Matrox technical support). With my setup, one thing that really strikes me is how I can make small changes, and see the lockup behaviour get worse (or better). With the tech demo, the problems range from locking up at the beginning of the tech demo (no corruption, locking up a few seconds into the tech demo, a really noticeable jerk at the same point, a lockup a few seconds later, or a lockup after 10 or 20 minutes. In some cases, depending on the setup, there is video corruption when the lockup occurs. For a given setup the behaviour is usually very repeatable (though to a lesser extent when the lockups occur 10 or 20 minutes into the tech demo). The only way I can get any reliability is to use the oldest driver I have (5.11), disable video caching etc. in the bios and apply the agp x1 registry hack. Without the latter, I get a lockup 10 or 20 minutes into the tech demo. This seems to suggest to me that my setup, even with the agp x1 registry hack, is on the edge, and that more aggressive drivers push it over. Indeed, one of the discouraging things here is that the newer drivers make things worse, rather than better.

Robert Inkol

[Markus Eisenstoeck]

Hi!

I need some advice. A week ago I bought me a new system at a dealer here in Austria equipped with a P3500, Abit BH6 board and a OEM G400DH 32MB. Yes I have a live and yes I have a Adaptec 2940U2W. I get the usual lockups described here (q3test lockups in opengl or random bitblits on screen forcing me to reboot).

I tried all the bios tricks, upgraded to lastest drivers, bios etc for board and every card in the system that is flashable. Finally Forcing to AGP 1x solved the problem (so it seems).

This solves the problem - but I BOUGHT AN AGP 2X card!

Should I bring back the card to the dealer and request a new one?

Is there any backing from Matrox - or will I just hear that the card is OK (maybe they test it and in their system it works... seems possible after reading all the posts).

Markus

[cjolley]

Markus Eisenstoeck,
First try my sugestions 4 posts back.

Run system information if you have win98 and look for conflicts too.

If that does not work, from what I have read here Matrox is very good about exchanging or even taking back boards at the online store. I suppose that shows their attitude toward their customers.
Your retailer can hardly expect you to keep the card if it does not work in your machine. It's in their best interest to do a swap rather than lose a sale. (though, you might have to point that out to them )

It's worth the effort!

chuck


------------------
ABit BH6 w/ Celery 333@500, 128mb gh@cas2, 10gb IBM@7200, SB Live Value@44kh, noname CDRom@40x, Mitsumi CDRW@2x2x8, Zoom@56k, Princeton EO75@1024x768x32x85hz, USB mouse,Matrox G400 MAX!!!!

[cjolley]

When I first got my MAX It and my SB Live did not play nice together. The solution was too simple.
1. Make sure you are using the latest SB Live drivers.
2. Pull the Live out of it's slot and throw it across the room. (well maybe not ).
3. Put it back in in the center PCI slot.
4. Set bios PNP OS? to yes
5. Set bios resource allocation to automatic.

Everything just hummed along after that.
Win98 redetected the Live and moved a bunch of resources around, then all was well.

BTW you will pick up an IRQ if you switch from ps/2 mouse to usb ( the usb will get one anyway so you might as well put it to use).

chuck


------------------
ABit BH6 w/ Celery 333@500, 128mb gh@cas2, 10gb IBM@7200, SB Live Value@44kh, noname CDRom@40x, Mitsumi CDRW@2x2x8, Zoom@56k, Princeton EO75@1024x768x32x85hz, USB mouse,Matrox G400 MAX!!!!