Random hangs on KT133 and KT333 with raid

On Thu, 11 Sep 2003 15:56:53 GMT, John wrote:

Ive had tons of problems with these boards.

I think it would be more fair to menton specific board makes,
revisions, and system configs, since obviously something is unique and
causing this, but NOT lumped all together like a rant, rather tackle
one problem at a time.

I think Im pretty sure now whats causing the problem.
On the KT133a boards I think its poor cooling on the Northbridge chip.
You know those little bitty heatsinks and fans they have. They are
cheap and are the first to go on boards that are a year or so old.

Yes, I never use 'em, unless someone insists on one, then I PRE-lube
it, before it ever fails the first time... wait until it stops or gets
loud to lube it and you'll already have a 1/2 worn out fan that
vibrates more due to the bearing wear.

I got an RMA on my old ASUS KT133a and was extremely annoyed when the
replacement - of course a refurb which was also old - had random
hangs. Not as bad as the RMAs board but still there. I gave it to
someone so its still running with the occaisonal SYSTEM file
corruption crash and hang. Im pretty sure now that its caused by the
intermittent failure of the little bitty fan on the Northbridge chip.

.... or memory or power supply or drivers... say, are you running
Windows? Then that too.

I caught it a while back completely stopping and oiled it. It worked
fine after that crash. Then it crashed again recently and I checked it
last week and the fan had failed again. I replaced it with a teeny
weeny fan you can get at COMPUSA - they have a pack of 3-4 little
heatsinks and one little fan thats a perfect replacement for it for
like $10 Im expecting NO MORE PROBLEMS like that since its a new fan
until a year or so.

Expect that new teeny-weeny fan to fail too. The smaller (and
thinner) it is, the less precisely it'll be balanced, the smaller and
shorter a bearing it'll have. For best performance use minimum
40-50mm x 15mm thick fans. 15mm is just about minimum thickness
necessary for a fan to have dual ball-bearings. Of course with this
larger fan you don't need but enough RPMs to get it to spin-up, so
it'll be quiet too.



Now for my hairpulling ABIT KT333 board. My other one was a complete
meltdown after a year. Not sure what was wrong with it - it worked
great for a year. But the replacement 1.1 revision - a newer one has -
you guessed it random hangs ! This REALLY annoyed me. Well I think i
isolated now since I ran into tons of problems last night ripping some
stuff. It generally hangs when doing intensive data transfers. Its
hard to isolate since its random but this time it consistently
failed,hung. Last night the problems were huge - and happened over and
over again at around the same spot.
It wouldnt finish ripping some video stuff in fact, After I switched
the Hard Disk to the built in controllers all the problems went away
so - so far it seems obvious that its the High Point controllers.

Did you update the BIOS? Newer Highpoint driver?

I guess everyone with an older board should double check and replace
that little bitty fan and make sure the heatsink isnt clogged with
dust if you are getting hangups and or corruption/crashes. I like my
new ASUS nforce2 board cause they replaced active cooling with passive
- a huge heatsink only.

Fortunately the industry is moving in this direction.

As for the highpoint - not sure what to make of it. Terninally buggy
drivers ? I dont think so. My original KT333 worked OK. Does the
1,1 have highpoint compatiblity problems vs the older 1.0 I had ?
That would be weird for a board to get worse.

Too many variables to say... might be different components, memory,
power supply, bios settings, Windows config, etc, etc, or you just got
a bad board. I do know people use these controllers without the
problems you've had, but it could be anything, even cooling.


You could just disable
the extra IDE connections or get a Maxtor add on card if you need one
- but that seems ridiculous since you paid for both. Does the
highpoint chip need cooling ? It doesnt feel that hot though the south
bridge chip does get pretty hot - it has no cooling at all. Not sure
if its relevant since there are no problems when Im not using the
Highpoint.

It shouldn't need cooling, but I've not noticed those Via southbridges
getting near as hot as the nForce southbridges do. Perhaps you have a
hot-spot in the case, dead air?

Maybe Im imagining this but it seems like the South bridge gets hotter
when Im using the Highpoint addon controller. Whats the connection
between the SB and Highpoint? I dont know.

The highpoint is a PCI-bussed device. The SB IS the pci bus
controller, evething PCI goes through it, but also the Onboard IDE is
contained within it, but isn't on the PCI bus. The onboard controller
is therefore a more efficient, lower-latency controller that can
exploit more bandwidth particularly when other PCI devices are being
heavily accessed (given same ATA speed) except in situations where
RAID's benefits outweigh that.


Dave

John wrote:

But at least now Im pretty sure its the
highpoint. First time I was able to duplicate it. Move it back on the
highpoint - problems , move it off of it - no problems.


People said the same thing about SB live sound cards being "the problem"
when used on Via chipsets when these cards have no problems on any other
chipset. Some Via chipsets do have problems with PCI latency and maybe this
is causing the problem with this PCI controller? Before I decided the
controller is the problem, I'd try it on a non-Via board.

Yes I =WILL= blame flakey problems like this on Via as I've been down the
road of dozens of reinstalls, swapping cards/driver updates etc and finally
replace the board with an SIS or nvidia and it runs smooth as silk with no
bios flashing and trying older versions of chipset drivers etc etc etc.
Dave seems to like Via boards, I can't fathom why anyone would want to
fight with the buggy 4in1 drivers etc these have.
--

Stacey

On Fri, 12 Sep 2003 03:02:11 -0400, Stacey wrote:

John wrote:

But at least now Im pretty sure its the
highpoint. First time I was able to duplicate it. Move it back on the
highpoint - problems , move it off of it - no problems.


People said the same thing about SB live sound cards being "the problem"
when used on Via chipsets when these cards have no problems on any other
chipset. Some Via chipsets do have problems with PCI latency and maybe this
is causing the problem with this PCI controller? Before I decided the
controller is the problem, I'd try it on a non-Via board.

Yes I =WILL= blame flakey problems like this on Via as I've been down the
road of dozens of reinstalls, swapping cards/driver updates etc and finally
replace the board with an SIS or nvidia and it runs smooth as silk with no
bios flashing and trying older versions of chipset drivers etc etc etc.
Dave seems to like Via boards, I can't fathom why anyone would want to
fight with the buggy 4in1 drivers etc these have.

I like nForce boards too though, it's just that I haven't seen all the
problems you have with Via boards and so haven't been avoiding them,
so presumably have used more, modern Via boards. I'd more likely
avoid a Sis board than a Via, but partly because they seem to have a
lower overall construction quality, plus a certain group of
manufacturer(s) uses a lot of Sis chips but provides relatively poor
support. Then again I'd sooner use an Asus Sis-based board than a
Biostar or ECS/PCCHips/et al, nForce or Via, so to each his own.


Dave

John wrote:



Its just bugging the hell out of me. Why did my original work OK with
the same drivers and then this one doesnt ? Even with the updated
drivers. Now I know why the KT133a ASUS problem was - the
intermiitent fan failure on the northbridge. But whats the problem
with the KT333 ? The temps and fans work OK. And I sure cant blame
the components - not only have I replaced everything and done
everything to rule them out - they work fine on my Nforce2 board.


Isn't it obvious? It's the Via board. The first one may have worked fine,
this one obviously doesn't. That's another problem I've had with Via based
systems. I'll build 2 or 3 with a certain hardware/software mix and then
the next one I build acts flakey..
--

Stacey

kony wrote:

Then again I'd sooner use an Asus Sis-based board than a
Biostar or ECS/PCCHips/et al, nForce or Via, so to each his own.



I guess I have ALWAYS used top tier boards so I don't blame the issues on
the board makers. Sure some of the Via systems I've built ran great. It's
just after a while you start to notice "Damn if it doesn't seem like every
=headache= system has a Via chipset.." I'm sure I could say the same about
every PC chips system etc if I used junk like that!

--

Stacey

On Mon, 15 Sep 2003 05:25:01 GMT, John wrote:


The writer is dependendable as usual , its a Liteon which Ive had
numerous older models. And Ive verified burns and they come out
perfect on all kinds of media. BUT - one thing that bugged me. My old
32x Liteon burner could burn any SVCD on all media at the max 32x.
My 52x Liteon can only burn max 32x SVCD too. No improvement.
Well the thing is - all the problems seem to come at the end of the
burns of SVCDs that are very long. I think it has a problem with poor
overburning. It can do it but only up to 32x.

When burning at 52X, the burner doesn't actually burnt at 52X until
the outer tracks, so a shorter file might never be burnt at speeds
over 32X even if that's the speed set to burn. I think a Google
search will find a regsitry mod for Nero that shows realtime burning
speed.. just don't recall any more details about the reg key though.


I updated the firmware to the latest since my firmware was quite old,
several verions older. The latest one interestingly says -improve
overburning. As far as I can tell despite several versions in
firmware updates - there seems to be ZERO improvement. Still has the
same problem.

If the capacity of the disc isn't being exceeded, it's not
overburning. Even if it were, it would be burning at the same speed
on the overburn as the last non-over-burnt portion of the disc, so it
would seem that overburning could definitely have this result, it's
just the amount of time it was burning at this "excessive" speed per
media that's the problem. Most discs can't burn well at over 32X on
any burner, result in correctable errors.


I bought a cheapo BENQ 48x on sale for my second PC.
It burns the same SVCDs perfectly on the same media at 48x the max
speed.

The BenQ isn't a bad drive but I think it has a more fraglie
mechansim, and also rubber belts for tray, so may wear out faster and
be more subject to heat-stress.

The last test Im doing is put the Liteon on my second PC to rule out
any PC factors. But Im wondering either my Liteon is a bit off in
overburning though the firmware that says "improve overburning" seems
to suggest it might not be, or maybe the 52246s versions relatively
poor at overburning and they improved it later in the new 52x series.


Im going to open up my backup Liteon and see if it has the same
problem too. I bought that a few months after my other Liteon. The
Benq I bought last month so maybe the newer burners all are better I
dont know.

I'd just burn at 32X anyway... even if the BenQ makes readable discs,
it likely has a lot of correctable errors (given same media) which can
only be bad for readability after the disc has aged.


Dave