Do Athlon XP's have problems? UPDATE: it's VIA's fault!

[Stark]

I used to have an Athlon 850 on an MSI K7T Turbo w/Raid and it was great. Solid as a rock. Then I upgraded to a retail XP 1800+ (the fastest msi officially supported). After bumping the FSB to 133, the computer would hang when I'd turn on the power. The only way to get the video to initialize and the post screen to show up was to hit the reset button on the case. I tried the latest bios, and it still does the same thing.

I normally leave the computer on all the time, so I just decided to ignore the problem and hit reset whenever I have to re-start from a power off state (the problem never happens on a windows reboot). My computer was fast and it wasn't that big a pain... I was still sold on the Athlon XP.

Fast forward to a couple of weeks ago when I helped a co-worker put together a system (he built 95% of it). Epox EP-8KHA+ DDR motherboard with a retail 1900+... and he's getting the exact same problem!! The only thing our systems have in common is the cpu. I'm now convinced that there is some defect associated with the Athlon XP.

Has anyone else seen this, and is there any fix? (I'm not about to buy a P4 anytime soon, but if there is truly a bug, the world must know!!)

UPDATE:
I finally decided to spend some money and upgrade my motherboard to an MSI nforce420 board with 256mb kingston DDR. Guess what... no more reboot problem. Now I'm questioning Via's chipsets... I thought they had greatly improved since the k6-2 days, but apparently they still have issues that creep up down the road.

BTW, the nforce board is sweet. I have found my new "general purpose workstation" setup to recommend to anyone who asks. nforce+Athlon XP+DDR memory = all you need in a basic system.

 

[jar5tyle]

That's odd... I'm interested to hear if anyone else has had that problem. 1700+ here running solid as a rock

bump

 

[Rob9874]

Yeah, my 2000 is running perfect.

Oh wait, I have a 2000MHz Pentium 4. That's why!
juts teasing

 

[Insane3D]

Don't know about the MSI, but the Epox board has a jumper for 100 or 133mhz FSB operation. This jumper comes in the 100mhz position by default. If you simply set the FSB to 133mhz in the bios, you are still using a 1/3 PCI divider and your PCI will be way out of spec @ 44mhz (133/3-44.3333). Make sure if the board has a FSB jumper, it is in the correct position..

 

[Siddhartha]

I have been running a XP1900+ for two months without any problems. Check to see you installed the heatsink properly.

 

[Stark]

<< Don't know about the MSI, but the Epox board has a jumper for 100 or 133mhz FSB operation. This jumper comes in the 100mhz position by default. If you simply set the FSB to 133mhz in the bios, you are still using a 1/3 PCI divider and your PCI will be way out of spec @ 44mhz (133/3-44.3333). Make sure if the board has a FSB jumper, it is in the correct position.. >>


The jumper on both boards is at 133MHz. I don't know about the epox, but on my msi board, the boot problem goes away at 100MHz.


<< Check to see you installed the heatsink properly. >>

That was the 5% of my co-workers setup I had to help him with. He put the heatsink on backwards and was lucky not to fry the cpu... about 1/2 the core was touching the thermal pad and 1/2 was on the bare heatsink aluminum. He couldn't figure out why it was locking up all the time. Thanks to the adventures of JonnyGuru, I had the answer for him.

If not an AMD issue, what about a VIA problem? Do the 133A and 266A boards have some issue with running at 133?

 

[xylem]

Judging by the fact that there are tens of thousands of these cpus operating properly, it's doubtful that there is a flaw in the cpu. More likely, it is a conflict between the cpu and the motherboard settings or between the motherboard and installed components.

 

[Anzu]

That's really, really odd.

As the sig shows, I have a 1600+ running about 25% faster than that, and it's never given me a bit of trouble.

 

[Ausm]

<< Yeah, my 2000 is running perfect. Oh wait, I have a 2000MHz Pentium 4. That's why! juts teasing >>



If Projectile vomiting was an Olympic sport I think I would have just won the gold medal

Ausm

 

[Ausm]

Sounds like a heat problem to me....what type of cooling are you using

Ausm

 

[GasX]

<< If not an AMD issue, what about a VIA problem? Do the 133A and 266A boards have some issue with running at 133? >>

I have an Asus A7v and you can't take the FSB much past 105. It is a known issue. It is something to do with the VIA chipset.

 

[Insane3D]

"I have an Asus A7v and you can't take the FSB much past 105. It is a known issue. It is something to do with the VIA chipset.

LOL! The Asus A7V uses the KT133 chipset which does not support a 133mhz FSB. That is not a "known issue", it's what the chipset was designed to do...


Jeez people....let's get the facs straight here....

 

[mboy]

Running an Xp1600@1543 on an 8KHA+ and an XP1700 @1503 on a K7S5A and both are just fine. Only problem is when I try to OC the 8kha+ too much (currently maxed at 146fsb stable).

I would say due to the fact that there are MANY MANY MANY people using XP's without incident a design flaw is an impossibility!

 

[breathe]

I have an 1800+ on a Soyo K7ADA running beatuifully.

 

[Stark]

<< Sounds like a heat problem to me....what type of cooling are you using >>



We're both using the heatsinks that came in the retail cpu box. He has two extra case fans, and I run mine with the side of the case off and no extra fans. Plus a heat problem would creep up after the computer was on for a few seconds... it wouldn't prevent the system from initially starting up.

Ok, maybe it's not a complete cpu design flaw, but there is something weird going on here. This is the first time I've seen this problem (and I've built plenty of computers in the past with PIIIs, P4s, Celerons, Durons, and Athlons) and both times it was XP processors on via chipsets with 133fsb. I suppose memory is another possibility, but he has Micron DDR and I have memman infineon PC133, so that shouldn't be the culprit.

Until I find an answer to this problem, I'll have a hard time recommending the XP to people who ask.

 

[Stark]

Here's some more info from my co-worker:

When his computer hangs, his onboard diagnostic shows FF.
When he hits reset and it starts to boot, it flashes through a bunch of different codes, then shows FF.

 

[CTho9305]

whast your power supply?

 

[gdawson6]

hmmm...i do hope you changed your power supply after upgrading or we found your problem!

 

[Mem]

<< Until I find an answer to this problem, I'll have a hard time recommending the XP to people who ask. >>




No problems here with my XP 1700+ & Epox 8KHA+ all rock stable ,yes check the PSU since some are fine for lower AMD processors but once you start going 1900+ and above it can seperate the men from the boys so to speak.
There are no AMD XP cpu problems I`m aware of period.

 

[Stark]

I have a 300w that I got in europe last summer. I guess that's a possibility.

My co-worker has a 300w that came in this Newegg says it's an enermax and AMD approved.

I'm not looking forward to buying a new $50+ PS just to test a theory... I have dozens of 200-250w units at my disposal, but nothing better than that.

 

[mboy]

NO offense, but you look awfully silly bashing the XP chip when 100,000s of thousands are in use without incident. 200-250w supplies are not recommended for it. 300 should be ok tho, but not a cheapie. If their is an inherent flaw, why are my 2 working, the 1 I built for my dad, and 5,000 others here at Anandtech, etc working fine? I just can't see how you are blaming AMD when neither of you are even using recommended PSU's!

 

[Stark]

Hi mboy,

I respectfully encourage you to read my previous post more closely. Although I may be silly, each one of your points is addressed there.

I just checked, and I'm pretty sure my PS is actually an HEC 300w that is on AMD's approved list.

 

[sandorski]

It's not the cpu, it's most likely a power supply issue.

From the Epox EP-8kha+ manual: error code FFh indicates these possible problems:

1) BIOS chip replace inverse
2) Update wrong BIOS
3) Mainboard problem
4) ADD-ON card insert inappropriate

#4 would be the best place to start, re and re all your cards, ram, and even cables.

Double check #2, make special note of any mobo revision notes.

#3 could possibly be an issue, although it maybe caused by something shorting the board. Check all your standoffs for contact with mobo traces, or any drives which may touch the mobo.

 

[Stark]

update in first post.