Best way to get a KT266A stable (4in1s or not?)

[kinney]

My last ditch effort before I order a nforce2 is a format and reinstall of xp. The system is a Gigabyte 7VTXE (KT266A), Athlon XP 1700+, 512MB Crucial PC2100 (2x256), VT GF3, Linksys nic. right now my xp install includes the latest 4in1s and latest nvidia drivers but i get reboots daily without warning and xp says its a system error detected by the cpu and they offer no fix.
I had an Audigy, a Hauppauge WinTV Go and a 3Com 56k gaming modem in the system prior to my trying to get the system stable. I took out the tv tuner and modem at first and it went from rebooting every 10minutes without notice to 1-2 hours. I have the interrupt handling mode set to PIC instead of APIC also. I got the Audigy on its own IRQ (5) at that point. Then the error reported by windows was that their was a device driver causing the problem. So I pulled out the audigy and now it reboots roughly every 12-24 hours and I get the error I listed at the beginning of my post that its a system bus or design error. Now I'm looking at the cpu or motherboard chipset.

I think a format is due before I order a nforce2 and will probably try no 4in1s, no pci cards and see how it goes. If its unstable try the 4in1s, still with no pci cards. If its still unstable at that point get the nforce!

I welcome all opinions and suggestions (please!)

Thanks in advance.

Athlon 1700+, KT266A, 512MB PC2100, Geforce 3, Audigy, Z560s

 

[Clarentavious]

The 4in1 driver pack is pretty much *needed* for VIA chipsets. I doubt the set packed with Windows XP is up to date, VIA updates their 4in1 all of the time. http://www.viaarena.com/?PageID=2

The thing that concerns me most about VIA chipsets is whether they are responsible for data corruption.

I've also had problems with my 266A by Abit, and Crashman mentioned Creative Labs based soundcards (from the SB Live and on, this would include the Audigy) can be a source of this.

I've had to mess around with mine to get it working, but I've never experienced random reboots when runnning 98 SE or 2000 (never run XP on this system though).

I'm personally going to switch to an SiS motherboard. I want to see how a single controller chip fairs out against a north and southbridge solution. IMO, while Nvidia has proven themselves in making video cards, the nForce chipsets haven't been out very long, and I'm not ready to trust Nvidia with a motherboard yet.


-----

Benchmarks don't lie

 

[kinney]

Data corruption on the IDE bus due to a soundcard? I doubt that.. I can replicate the data corruption bug in the kt266a with or without my audigy.

For me, all I have to do is start transferring large amounts of data over the IDE bus and the transfer fails.

I think the sis 745 is a great chipset but I don't want to trust sis... too cheap. I had a ecs k7s5a, the revision i had a flaw in the board where the cpu wasnt getting enough driver impedence or something. Not sis's fault.

I think the nforce2 has alot to offer with all the integrated features for the price, the board I'm going to probably order is the leadtek with the mcp-t (soundstorm).. put the audigy out to pasture until might have to build a P4 system if its not stable on the nforce.

But I think an audigy would work perfect on any system board but VIA. I've read the reason is via's chipsets dont offer a pci bus feature that intels have for quite some time, so creative assumed that most chipset manufacturers would have implemented it by now.

Athlon 1700+, KT266A, 512MB PC2100, Geforce 3, Audigy, Z560s

 

[RCPilot]

Have you checked out your PSU??? It sounds like to little power or it's failing. No motherboard will be stable without a proper & working PSU.

Skinny

Signature, I'm still learning & having fun doing it!!!!!!

 

[Clarentavious]

SiS boards are cheaper than most because they generally only use 1 controller chip. The controller chips are where much of the price comes from. The circuit board and other connectors don't cost as much money.

Until Nvidia released the nForce chipset, you had no other alternative at all aside from ALi (Acer) and AMD, if you were going to use an AMD CPU. I think AMD has quit making motherboard chipsets, and I'm not sure ATi ever made a chipset for AMD processors. Not like they would sell many motherboards even if they did.

The other choice would be to go with an Intel processor. Intel hasn't been flawless either. Let's not forget the MTH issue with the 820 boards, and how "well" the Williamette cores performed.


-----

Benchmarks don't lie

 

[phsstpok]

The old VIAHardware.com had a series of articles on the Soundblaster Live/VIA 686B Southbridge controversy. I think the site is gone but the articles can still be found at <A HREF="http://www.sudhian.com" target="_new">www.sudhian.com</A>.

Anyway, there were two separate issues, data coruption and audio problems (crackly sound). Viahardware.com did some tests. They determined they could reproduce the sound problems with various chipsets including VIA, Intel, and AMD. They tried multiple sound cards and only the Soundblaster Live cards had the problem. (The article was written before Audigy came onto the market)

The more important issue was the 686B bug. When the PCI bus was saturated on motherboards which have the VIA 686B Southbridge (this includes VIA KT133A and KT266 boards as well as some AMD 760 boards) then data loss can occur. Not everyone ran into this problem but many did. An individual named George Briese wrote a small fix, a Latency patch, which eliminated the problem but at the cost of some performance (I think). The patch also seemed to fix some PCI timing issues as I recall. Viahardware determined that on known problem boards, those that actually exhibited data loss, the patch seemed to correct the problem. Supposedly, manufacturers incorporated a similar fix in BIOS updates. Likewise, later VIA 4-in-1 drivers (4.28 and on I think) allegedly had a fix as well.

The latency patch did not fix the Soundblaster problem. This issue was complicated by the fact that not all Soundblaster Live owners could not reproduce the problem.

The last I heard about this is that if one has the sound problem there is a workaround (but not a fix). If the Windows main sound level is set low and compensated by setting one's speaker level high then the problem gets masked. I think keeping the sound card, itself, at low volume levels prevents it from introducing the noise or reduces the relative noise level. In any case it works.

Well that's basically the last I have heard on both issues. As far as I know the data loss problem has become a thing of the past as VIA no longer uses the 686B Soutbridge. Soundblaster Live cards are still available but I haven't heard many recent complaints on sound quality of these cards so perhaps that issue was fixed.


<b>99% is great, unless you are talking about system stability</b>

 

[Clarentavious]

I sometimes get weird fuzzy noises out of my speakers. I don't think it is due to my Soundblaster, but probably just occasional signal interference. I have an SB Live 5.1 X-Gamer. The first SB Live cards were the Value and another one that had an additional feature that wasn't included until the X-Gamer series was released. They weren't 5.1 either (the best speaker setup at that time was 4 point surround stereo or something).

I know Creative Labs has had alot of problems, but as for their sound cards, I think they are probably the best sound cards available period (outside of professional Hollywood recording devices and machines designed to deal with forensic acoustics).

There really aren't that many other chipset makers for sound cards, and "sound blaster 16/compatible or better" has been recommended on the bottom of game boxs for years.

Let's not forget Creative was the one to introduce EAX for Environmental Audio. The company itself has had its fair share of problems (such as in vendoring video cards and making drivers), but I wouldn't say their faults have lied in the area of sound.

The data corruption I have experienced has been rare, but it isn't cool if the files are important. It has often been when I have written CDs. I have a Cyberdrive 32x12x48 with Exaclink. I usually write my CDs at 28x I only use high quality media like iMation.

Do you think setting potential performance boosting options that "could" cause instability might be to blame? My 266A is by Abit. What I mean are things such as BIOS options like: CPU fast command decode, 1T data channel access for RAM, FSB overclocking, AGP fast writes, etc....

I know my FSB limit because if I go above 136 my PC freezes after a couple minutes of operation of any type (whether I leave it idle or do intense gaming). But if I leave it at 136, I never experience crashes or random problems, I would conclude it is stable at 136, but could this affect possible data corruption regardless?

Also, I happen to use 98 SE on this system. Do you think my OS is to blame? Do you think corrupted files would be less likely if I was using Windows 2000?

Also, I have heard there has been speculation among the industry that VIA may have stolen SiS' memory interface, and this is what accounted for the 10% performance increase from the 266 to the 266A revision.

VIA's processors, dating back to the first Cyrix, to the C3 now, haven't been all that great either.


-----

Benchmarks don't lie

 

[phsstpok]

I sometimes get weird fuzzy noises out of my speakers. I don't think it is due to my Soundblaster, but probably just occasional signal interference. I have an SB Live 5.1 X-Gamer. The first SB Live cards were the Value and another one that had an additional feature that wasn't included until the X-Gamer series was released. They weren't 5.1 either (the best speaker setup at that time was 4 point surround stereo or something).

I've run into the sound problem first hand. My friend still uses an ASUS A7V with the original Sound Blaster Live (full retail not Value). His static problem was very severe to my ears. We followed sound issue for about a year before I stumbled onto the advice to keep Windows main volume low and just turn up the speaker volume. The tip does work.

I found the article to which I referred early. Here's the link.

<A HREF="http://www.sudhian.com/docs.cfm/id/29/pid/53.sud" target="_new">http://www.sudhian.com/docs.cfm/id/29/pid/53.sud</A>

They compared several sound cards of the day (Aug'01) including SB Live Value, TB Santa Cruz, Philips, Zoltrix, Abit AU10, VIA AC'97. All having different sound chips.

It seems my memory was a bit faulty. According to the article, the corruption problem was traced to the <b>combination</b> of the SB Live and the Via 686B Southbridge. (Early and later chipset were not tested). Also, only KT133A, KT266 (not KT266A), and Intel i815EP motherboards were tested.

Noise problems were produced <b>only on the SB Live</b> and it did occur on the i815EP motherboard, not just the two VIA motherboards.

Data corruption was detected only with the combination of SB Live plus 686B southbridge. When the other sound cards were used or the other motherboard was used there was no data corruption.

As I recall reading in forums, including this one. The issue is easy to test. To do this you simply creat a large file, at least 100 MB. Copy the file, preferably to a second hard disk on a separate IDE channel. This cause the PCI bus to be saturated during the transfer. If the resultant file is corrupted then your system obviously has a problem.

To confirm if what you are seeing is actually the SB Live/686B problem just disable the sound card in device manager or better yet temporarily remove the sound card from the system. Repeat the above test. If this time you do not get a corrupted file then you just confirmed the problem.

Even if you don't confirm the problem, I recommend downloading the George Briese's PCI Latency Patch and see if it fixes your data corruption.

Anyway, read the article and judge for yourself if it is applicable to you. As far as I recall, the KT266A does not use the 686B southbridge but perhaps the problem exists in later chipsets. I don't know. The fix is worth a try. After all, you can't trust your system if it has a data corruption problem.


<b>99% is great, unless you are talking about system stability</b>

 

[phsstpok]

I forgot to comment on some of your issues/questions. Sorry if I get the order wrong on some of my comments.

I also have a Cyberdrive CDRW. Mine's the 36x12x48x model. I have found the best results with Taiyo Yuden manufactured discs. The discs have few or zero errors when tested. I can read them on every drive I have and the read speeds reach the maximum of each drive. TY manufactured discs generally come in the cake boxes with a screw cap above the cover. Also, if you are in the USA. All Fuji discs are TY.

Abit KT266A? Is that the KR7A? This is an excellent overclocking board. I seem to recall one person reaching a 180 Mhz FSB, for example.

136 Mhz? Hmmm, that's a pretty modest bus overclock. I wouldn't think you would have problems. How are your temperatures?

Funny you mention the SB 16 recommendation on games. I just realized that I have never owned a Creative Labs card. My cards were those "SB compatible" cards, going back to 1987 with what I think was called the Pro Audio Thunderbird, followed by a Pro Audio Spectrum, and a Turtle Beach San Tropez. Sadly, I will have to retire the last when I next upgrade my motherboard. The San Tropez is an ISA card.

I've been using VIA chipsets for years. I had all kinds of problem with an MVP3+ board until I installed a 300 watt ATX power supply, replacing a 230 watt AT unit. The last of my issues on that MVP3+ board were eliminated with the release of the 4.29 4-in-1 drivers. That board has been very stable ever since. I also had good luck with an Abit KT7 (KT133). Now, I'm using a KT133A mobo, Epox 8KTA3PRO. This one hasn't worked out quite as well but it's a combination of things not just the mobo. I'm getting tired of constantly installing/reinstalling 4-in-1 drivers. My next mobo won't have a VIA chipset.

I use Windows 98SE exclusively so I can't speak about Windows 2K. I, personally, don't have many complaints about Windows 98SE and when I do have problems I like being able to fix them without having to reinstall Windows.

Also, I have heard there has been speculation among the industry that VIA may have stolen SiS' memory interface, and this is what accounted for the 10% performance increase from the 266 to the 266A revision.

There has been a lot controversy surrounding VIA. This would not be the first time that they were accused of "borrowing" technology.


<b>99% is great, unless you are talking about system stability</b>