Chronic instability on MSI K7T Pro2A -- attempt to contact tech support failed!

[Modus]

My "suggestion" is only the obvious: that the problems a handful of people have experienced with this board amount to nothing compared to the vast, silent majority who have found it perfectly stable, including several competent resellers who report no problems.

I'm sorry, but it's a little arrogant to assume your own problems are so important that a company should change its entire policy just for you. Like every other poor schmuck (and believe me I've been one) who lands a deffective board, you MUST follow the ordinary channels, beginning with your reseller.

Get a replacement, get some support from your reseller, email MSI and wait for support, reinstall the new board, strip down your hardware, follow process of elimination to eliminate any mutual incompatibilities (some RAM will work fine in one board but puke in another.) and settle down. We'll help you out.

Modus

 

[XMan]

Leo, just out of curiousity - are you still undervolting your processor in the quest for the quiet computer?

 

[WetWilly]

LeoV,

I've gone back over the thread, and there could be several causes of those symptoms. BTW, I can reproduce the problem with having to cycle the power several times on my KT7 if I overclock slightly too far. From your system specs, you aren't overclocking, so humor me - I have a couple of other questions:

1) Does disabling ATA/100 using IBM's utility stop ALL problems or just IDE-related problems?
2) Do you have problems in Linux?
3) Which version of the 4-in-1s are you using with Win98SE, and which components of the 4-in-1 did you install?
4) What are the slot/IRQ assignments of the cards in your system?
5) What kind of DVD-ROM did you have in the system, and was the drive properly identified in the BIOS device enumeration screen? Same for the CD-RW.
6) Which GeForce drivers are you using?
7) Have you adjusted the AGP driving value for the GeForce2 MX card?
8) Have you increased the CPU core voltage or are you running it at default?
9) I saw you have Kingmax PC133 memory - how many sticks do you have, and have you tried running with either one stick of memory or with totally different memory?

Hawkeye_(BEL)

True, it's unfortunate that not everyone got on the bandwagon with fixing the 686A southbridge BIOS issues. That's why I said "most." I'd had a Tyan Trinity 400 (Pro 133A + 686A) and it also gave HDTach scores around 57K MB/sec with an ATA/66 drive. AOpen's turned their act around, unfortunately it seems like their Pro 133A boards were for practice. In Tom's review of the first boards with the 686B, the AOpen AK73Pro (KT133/686B) was at or near the top spot for nearly every benchmark, which also discounts his theory that there's a problem between the KT133 and 686B. If AOpen can pair the KT133 and 686B without a performance hit then the problem lies elsewhere.

 

[jpprod]

Whenever I watch DVD movies, the system will lockup sometimes with a BSOD, random reboot, or black screen with a green line on top.

This is the exact problem I was having, until I forced hard drive from UDMA/100 to UDMA/66 mode. I am fairly certain that the problem is caused by a 686B southbridge's instability issue when there's lots of traffic across PCI bus. I've been able to reproduce BSODs and complete lockups in following situations:
- Copying a large file from hard disk (or partition) to another with UDMA/100 ebabled.
- Watching a high-resolution movie on Windows Media Player with sound (either my SBLive or integrated) and UDMA/100 enabled.
- Copying large files over 100mbbs LAN. I got a BSOD even with UDMA/66, but adjusting the upper limit of PCI ethernet adapter's bandwidth and enabling setting "PCI bus efficiency" has stabilized LAN transfers.

So far my system's been up for more than twenty four hours, but that's because I haven't done anything very intensive on it. I am going to swap this motherboard for an Asus A7V within a couple of days.

And for the person who's going to blame my sucky components or configuration for these crashes in an above post,
- I've tried various memory configurations (1/2/3 DIMMs) and clock rates (100/133MHz)
- I've finally done a clean install of Windows, gone trough a load of different display, sound and VIA driver revisions and patched to latest BIOS. Out of the things I've done only disabling UDMA/100 (completely turning DMA off, or reverting to UDMA/66) has had a major impact on stability.
- my old MSI K7Tpro rev1.0 was perfectly stable in any torture situations I threw at it.
- these issues have been reproduced on several other systems.

Conclusion: there is something seriously wrong in MSI motherboard - VIA 686B southbridge - KT133 northbridge combination. I wish it can be fixed with driver and/or BIOS revisions. However, I am not going to wait for them - 95% stability is not acceptable for a server system.

Suggestion for ppl with K7Tpro2A, UDMA/100 hard drive (in that mode), DVD drive and 100mbbs ethernet adapter: can your system stay up from the following test:
1. begin copying a file of 500-1000 megabytes or more from one disk or partition to another.
2. start uploading or downloading a file of aforementioned size over LAN at 100mbbs.
3. launch a DVD or high-resolution MPEG4 movie.
If it can, we can begin to assume that the stability problems me, LeoV and other people - with quite a bit of previous computer experience BTW - have been having were isolated incidences caused by individual bad boards.

 

[WetWilly]

jpprod,

You had all that stuff - TBird, mobo, original GeForce, 2 hard drives, CD-R - stable with a 230W power supply? What brand was it?

 

[jpprod]

Wetwilly: Yes

[edit] CWT (Channel Well Technology). Not a known brand PS but, just an old ATX power supply that came with this OEM case. It "sucks" Meaning the fan blows air into the case, not out. Main reason why I've sticked with it is it's quietness - fan's almost completely silent.

[edit2] Just a though, but could this have something to do with the fact that the AC current here in Finland is 240 volts? That 230W might in reality (though not theoretically) be more here than it is over there

 

[Vinny N]

jpprod:

Hmmm, a PCI bus traffic problem?...

That's interesting...

Might suggest that my SCSI card was being choked for a moment on the PCI bus and so windows spit out a "cannot write to drive C:" BSOD...

That one time I got a Fatal Exception 06 BSOD, it was while watching tv on a WinTV card, it passes everything through the PCI bus and video is displayed as an overlay...wonder if the PCI bus was struggling again...

 

[jpprod]

Vinny_N: Certainly more proof to the theory A certain web site could investigate into this...

 

[WetWilly]

Edit

 

[AndrewR]

No one seems to have mentioned the Windows shutdown problem and problems with hard drive cache, it seems. Not sure if you saw the stories about it, but there is an indentified problem with processors faster than 900MHz (or thereabouts) when running Windows 98 in that when you shutdown, the drives do not have enough time to empty their cache before Windows cuts the power. I was having these errors all the time when I shut down or restarted, and it dawned on me when reading the story that the problem was my second drive being idle at the time which was killing the shutdown process (I think). It seems to have been fixed with the patch from MS.

I'm running an MSI K7TPro under Windows 2000 with a Duron 700 without any problems at all. Rock solid and a pleasure to use. My other board is an MSI 6309 with a PIII700@933. Also very good with only minor Windows problems (other than related above).

 

[Vinny N]

Red Dawn:

...almost sounds like the cream of the crop ribbon is passed around...
Abit got popular, starts pushing more boards out the door, then we hear more about people having problems...
Asus is then hailed as the best stuff around, but eventually we get the A7V, and some people have problems...
MSI is now touted as the most stable boards, they get more boards out of the door, and we hear about more people having problems.
Sounds reasonable...

 

[Leo V]

Information overload, information overload!!

"email MSI and wait for support"

Modus, this may come as a surprise to you, but MSI doesn't have a tech support email! They do have separate marketing and sales emails, just no support email :disgust: All they have is the long-distance answering machine that doesn't return calls. So indeed, I'm contacting John about a replacement/refund.

Xerox Man, the first thing I did was set the CPU to default voltage/speed. That's Step 1 in troubleshooting.

WetWilly:

1) After disabling ATA100, jpprod hasn't had a crash yet; neither have I. Obviously the whole point of the Pro2-A is ATA100 support, so that's little consolation.
2) Linux (2.4.0) runs fine so far (zero crashes), but I hadn't run the sort of intensive programs I've used under Windoze. Long compilations didn't seem to stress the disk nearly as much.
3) Tried the latest official 4-in-1 (4.25a) and the latest beta ones (4.28). So has jpprod.
4) Long IRQ list, but no conflicts were reported. Removing the 3com nic seemed to help, but I hadn't the time to thoroughly test all combinations of PCI slots. I did try some, and it didn't help any.
5) DVD-ROM: Toshiba M1212 6x/32x DVD-ROM, detected as a CD-ROM drive in BIOS and Windoze. But capable of playing DVD's with DMA disabled.
CDRW: Plextor 12/10/32A, detected properly. Haven't tried extensive burning yet, only a couple CDR's so far.
6) Latest official Detonators, 6.31. The videocard is ruled out, anyway--I've been using a Riva128ZX for the past week, same problems.
7) No, but again the videocard is ruled out.

At this point, I'll add that I've already tried the "Fail-Safe" BIOS settings, which miserably failed to stop the crashes.

8) Default, default, default, of course!!
9) Memory is also ruled out. Until just recently, I used a single, high-quality 64MB MemoryMan stick, same problems. The 2x 256MB PC133 Kingmaxes can't be the culprit.

So as you can see, I've pretty darn much exhausted the possibilities, wasting a good many hours in the process.

 

[WetWilly]

It *is* an interesting theory about the PCI bus. Another question - how many options does MSI give you in the BIOS for tweaking/optimizing PCI transfers? This reminded me of one of Abit's BIOS options - "PCI #2 Access #1 Retry" - which deals with putting/removing a limit on how long the AGP bus can query the PCI bus. Considering the rather odd symptom of the blank screen/green line under heavy load, perhaps the AGP bus is timing out and crashing? Now I think about it, this was the same problem with the Vortex 2 sound card and older drivers. I know some Abit users solved graphics problems by adjusting this setting (I don't recall off the top of my head how it was set).

 

[Wolfman35]

RED (and anyone else who knows me): My experience has also been that there was a bad batch of K7TPro2A's. In a recent (01-05) shipment of 20 Pro2A Boards exactly half (10) had some combination of the problems listed here and 3 had the infamous "Multiplier Bug". Seems to me that MSI has experienced the same fate as Abit did when they got swamped with orders in Late 98. Rush to production at the expense of QC. My Solution was a switch to the Epox EP8KTA2. That batch of 20 Epox boards (01-16) had ZERO defects including proper functionality of ATA100 (IBM 75GXP 45G Drives) Several of the Epox systems got Zip drives (Both EIDE and USB) which I could never make work (properly) on the MSI Pro2A Boards. My experience has been that this is an Issue with MSI rather than VIA and only exists in SOME MSI Boards. Not that the VIA chipset is without flaws but the I-815 has flaws as well. MSI I-815 (6337) boards have a higher defect rate than the Asus CUSL2 as well but (so far) it's been an acceptable percentage (~10%) of 100+ 6337's. MSI's QC has definitely gone downhill as thier market share and production demand has skyrocketed.

Modus (et al): You simply cannot judge the quality of a component based on limited experience with a few random parts OR a Review based on ONE Part. There are definitely some Issues with the MSI Pro2A Boards that do not seem to show up (as often) in other boards based on the 686B chipset.

This is all Sooooo reminiscent of the Abit situation 2 years ago!!!! If MSI Starts charging $25 per RMA I'm going to start wondering!!!!

 

[John]

I sold him the board, and I offered to have a replacement shipped to him. I have sold ~50+ boards and 4 people have had various issues so far.

 

[McMadman]

I have the same board, have had the same problems as I said in an earlier message. The only difference is my problem was mostly fixed by installing the 2150 busmaster drivers manually, it's not perfectly stable (win98se), but is a lot better now than it ever was. I haven't pulled the ata100 sticker off to see the week of the chip.

I tried to disable ata100 using the maxtor util (dropped it down to 66) it didn't change any of my crashing problems.

Reinstalling windows made it stable for a little while, but after a couple days of use it became unstable once again.

I do not know if i could get it exchanged for a newer board (I ordered from monarchcomputer.com) but after numerous attempts to get my aiw radeon cables, the idea of waiting the weeks-months for them to get a different motherboard dosen't sound like a good idea.

Has anyone that uses a buggy board had any luck getting it working under win2k?

 

[Leo V]

Red Dawn, FYI indeed I've tried two PSU's. This one is a 300W AMD-approved supply.

 

[LXi]

Oh no, MSI is going to turn into Abit II. I hate to say it but jumping conclusions is probably not going to help. I doubt MSI will change their roots to success --QC, in expense to make production rushes.

 

[John]

<< Oh no, MSI is going to turn into Abit II >>



Puhleeeeeeze

 

[Leo V]

&quot;No one seems to have mentioned the Windows shutdown problem and problems with hard drive cache, it seems.&quot;

I forgot to mention, but I had already applied Microsoft's fix for this problem. So it isn't relevant here.

 

[Wolfman35]

John: Serious question for you. I know you sell alot of the MSI Boards but how many of that 50 have you integrated yourself? Realize that many of the &quot;Issues&quot; don't show up under all configurations. Many people don't use ATA100 drives, USB devices or Multiplier Adjustments so those problems wouldn't show up for them. Also ... were ALL of those 50 boards Pro2A's? 4 out of 50 is 8% and that is well within the normal defect rate for Motherboards in general and about the rate I saw with the K7TPro2 (non A) but these &quot;Issues&quot; seem to be isolated to the Pro2A and especially later production so you have to limit your perspective to the time frame when these problems started sufacing (best I can tell since late December).

While I don't believe that MSI's problems (and YES they DO have problems) are as servere as the dying BIOS chips ect. that Abit had with the BH6/BX6R2, any reasonable systems integrator has to see the similarities of the situation. Also, keep in mind that building and configuring systems in quantity can provide more insight than simply selling components. Many system integrators got burned on the Abit problems a few years back (Ask Russ for one). Perhaps, with a little FACTUAL information, we can prevent that from happening again. As Red Dawn pointed out in an earlier post, when the Abit sitution occured in late 98 many Abit fanatics emphaticaly denied that a problem existed but as we now know ... it DID. I hope that situation does not repeat itself with MSI but to blindly deny the possibility (in light of ALL the reported problems) is somewhat closed minded.

Edit: Lxi ... MSI has already made the production increases. It remains to be seen at what (if any) expense those production increases were gained.

 

[holdencommodore]

I have been lucky, no major problems with my m/b. If i do have problems, I just write to support@msi.com.tw. I usually get a response within a few days.
To clarify this issue with the 686B chipset, I think that Anand should do his own comparison of the Pro2 and the Pro2-A, to see if Tom's report is correct. If there is an issue, hopefully it can be resolved with a BIOS flash...

 

[tjdavis1138]

I'm more anxious than ever now to get my Pro2A back and see if it has any of these problems. Man I hope not. Couldn't sell it if it did and what a pain to try and get my money back.

 

[Leo V]

holdencommodore, thanks for the email address. I'll write them, asking whether they're aware of this issue. (Of course, they'll lie through their teeth...)

 

[nortexoid]

i didn't read all the posts, so somebody may have said it before me, but what the f*ck does Intel vs. AMD have to do w/ your apparently small-scope problem?...

if VIA's chipset is to blame, so be it...the chipset and chip are entirely different.

people and their &quot;sweeping&quot; generalizations...&quot;my isolated problem w/ via chipsets means via chipsets are the worst and never work, nor do AMD chips that run on via platforms, and my IBM drive must suck too cuz i ran it on this chipset and chip, and etc.&quot;

ASUS probably has even worse tech. support than MSI, so don't be too upset. (i, for one, am NEVER buyin another Asus product unless it's completely compelling to do so..but that's just based on tech support)