BP6 Stability Issues

[hyperspace]

The BP6 mainboard can have stability problems under heavy loads or when overclocking. Not all BP6's have stability problems. Many on the forum feel that most of the stalility issues are due to the mainboard's voltage regulation circuits for the CPU's. Since the majority of the modifications involve components associated with these circuits, the evidence weighs fairly heavy toward those observations. Keep in mind though, there many other things that can affect the stability of the your mainboard. Personally, I think the overall design is sound but I don't think Abit realised what the modern PC enthusiastic would try to do with a dual processor mainboard. Many serious BP6 modders have shown how to increase stability by modifying these circuits. Check the references below for more details about these modifications.

Yoichiro's BP6 Modification Guide

Anton Tomov's BP6 Modification Guide

Derek's Most Xcellent Original Dual PIII Notes

Stability issues may cause strange BSOD, your mouse and keyboard will lock up while surfing the WEB, the system hangs when using graphics intensive programs like PhotoShop or Quake III; just to mention some of the common lock ups. You will have to use the Power or Reset button in order to clear the lock up. BP6 mainboards that are revision V1.1 seem to be effected the most. The EC10 capacitor will be a 100uF. More information about this, in a moment. Some BP6's that are revision V1.0, will have the wrong EC10 capacitor. Check for the revision information on the external side of the ISA slot that is next to the edge of the mainboard.

There are a couple of good ways to determine if you are experiencing lock ups due to the BP6's own stability. One involves creating an color area-fill in PhotoShop which produces a very large memory load for the CPU's to deal with. In Photoshop, make a new image 100 x 150 CM at 150 DPI (total 149.7 MB). Now, use the paint bucket to fill the canvas with a random color (Any colour). During the fill process, the BP6 will lock up. You can also monitor the BP6's VTT voltage using a program like MotherBoard Monitor, MBM. The proper VTT voltage should be around 1.5 volts. It can vary from 1.47 to 1.53 volts. If the range of variation is more like 1.3x to 1.7x volts or better, the BP6 will have stability issues.

One of the primary causes for the BP6's instability, is due to the EC10 capacitor which located on the edge of the mainboard, between the two CPU sockets. If the value of this capacitor is 100uF (it will be a black capacitor), it will need to be replaced. Replace this capacitor with a high-quality, low-ESR radial aluminum electrolytic capacitor with a capacitance of 1500-2200uF and rated WVDC of 6.3-10 volts. Capacitors with higher voltage ratings, may be too big for that area on the mainboard. There are a few other components close by. Abit's official Fix is replacing the EC10 with a 1500uF, 6.3 volt electrolytic capacitor.

Some BP6 owners have corrected their stability issues by replacing the Q6 voltage regulator. It seems as though the EC10 modification is usually sufficient for resolving stability problems. If you continue to experience stability issues after this modification, please check our the link for the Q6 modification:

Q6 Fix

Important Note: Soldering on a mainboard is NOT for the faint of heart. It is one of the most important aspects of micro-electronic assembly. If you don't know how to Solder on a mainboard, find someone who can. You will be much happier.

Recently, there has been much WEB press about the "Copied Electrolyte Formula" used by several capacitor manufacturers. It affects many different mainboards including the BP6 (and VP6). Check posts on the forum for more information. Defective capacitors will appear discolored (dark greyish), bulging, or oozing. The electrolyte ooze will appear reddished-brown in color. Many BP6ers have replaced all of the voltage regulation capacitors around the CPU's and the System temperature sensor, RT3, reporting excellent stability after the modification. Some BP6 owners have reported scorched or burnt induction coils in the voltage regulation circuitry area. This is most likely due to the faulty voltage regulation capacitors. Information from other WEB sites confirms this particular phenomena.

Based on the work done by so many different individuals and the recent information regarding low quality capacitors, it seems clear that the best modification would be replacing all of the CPU and SYSTEM voltage regulation capacitors with 4700uF, 6.3 volt, radial aluminum electrolytic capacitors. Additional voltage modifications to other voltages circuits may be necessary for those wishing to overclock beyond 110 Mhz FSB. Check out Yoichiro's WEB site listed above for more information about modifying these other circuits.

Important Note: Power is very important for stability on any dual system. Two processors, more RAM, more hard drives and CDROM drives, most if not all the PCI slots filled, high-end AGP graphics card, more cooling fans, etc. All this makes for one power hungry system. Make sure power consumption is not an issue.

[hyperspace]

Overclock to 75 Mhz FSB and try to copy the data.

[Derek]

I've got a newer BP6 than the ones targeted by Abit for the fix, but it also locks up under 2k when running the photoshop test.

It also had a habit of rebooting randomly before I replaced the highpoint controller with a PCI Promise UATA-100 card.

Now it only reboots when it's under heavy load, usually when I'm right in the middle of converting media files.

I've checked the voltages in MBM, and the vcore is showing a range from 1.47 to 1.53.

Would it be worth it to replace the caps, and vrm?

Yes it would be! Replacing the caps should fix the remaining lockup problems.

[kuun]

ok,

i don't have lockup problems when im OC'd to 75 mhz FSB (500's @ 563) but when i try to play with the 80mhz FSB (600mhz) it'll lockup randomly when not doing anything, and sometimes i'll lockup hours after boot and sometimes it won't load the first program in starup...

im thinking of replacing the EC10 and seeing if that will fix it

i have rev 1.0 on both my boards

Kuun

[harker]

If you are using the onboard controller 366 I would 1st get a promise card and bypass that to see if that works, Easiest solution I feel and since I abandoned that I don't get lock ups any more

[kuun]

EW EW EW

DIE HPT366!!! DIE!!!

[Derek]

I've got a BP6 (not sure which model) with 2 Celeron 466's in it. No overclocking or anything. I've got 2 HDDs on the HighPoint controller. Running Windows 2000/SP3. All has been fine for years on this box, but now it's spontaneously rebooting, and nothing I seem to do tells me what it might be. I've moved the drives from the 33MHz controller connectors to the 66MHz controller connectors, hoping it was a controller issue, but to no avail. (not sure why it was ever on the 33's, but this is my wife's machine that I haven't touched in years.)

Also, no matter what I select for the Boot Seq. options, I can't seem to boot from the CD-Rom drive (actually a Toshiba DVD-Rom/CD-RW drive). CDs are good, I can boot to them in other machines, but not with the BP6.

Any ideas on where to start troubleshooting this?

Regards,
David Hull

Check for dead capacitors. Spontaneous rebooting is a sure sign of dying capacitors.

[hyperspace]

What a difference replacing the EC10 cap makes.

The Photoshop fill test passes, but with a quirk.

Right where it usually locks up, it'll pause for about 10-15 seconds, and then continue filling.

The system hasn't rebooted as a result of my media converting tasks, but I'm curious if replacing the VRM would put a halt to the Photoshop pause?

The other capacitors on the board seem to be fine, and I'm not noticing a huge drop in voltage the way it used to do when under load.

Your pause may be due to the processes going on between PhotoShop and the OS. How much RAM is on your system? What OS are you using?

[Derek]

Try it on 2K Professional, I could only get Photoshop to work properly on Server a few times.

[hyperspace]

Try it on 2K Professional, I could only get Photoshop to work properly on Server a few times.

Isn't there a setting that allows you to control the foreground/background of Applicaions on the server? Being able to give Applications Foreground processing. I think its in the same diagolue box that allows you to optimise the server as a file server. (i can't remember... )

[Derek]

Try it on 2K Professional, I could only get Photoshop to work properly on Server a few times.

Isn't there a setting that allows you to control the foreground/background of Applicaions on the server? Being able to give Applications Foreground processing. I think its in the same diagolue box that allows you to optimise the server as a file server. (i can't remember... )

You’re absolutely right. But "Server" and "Advanced Server" will always use more resources than "Professional". The Server versions have plenty more services running in the background and the operating system uses resources differently. And then there is the issue of compatibility: Where Professional is compatible with most software, Server may not be compatible with the same software.

[acelondon]

Any of u ppl ever heard of a 660?

http://www.frozencpu.com/cgi-bin/frozen ... ?mv_pc=454

you'll never have power problems with this baby

Im running:
2 x AMD Athlon MP 1800+ @ ~5.2Ghz(mad OC)
custom Dual Cascade cooling
1 x maxtor IDE 20gig
4x maxtor IDE RAID 40gig(tweaked to 70Gig each)
1x dvd drive
1x CD-RW
6x vantec tornado

Ya, I'm Insane. Got a problem with it???

[hyperspace]

More Power good...



Thanx, FrozenCPU.Com!

(hyperspace drools)

[Beata]

I was getting some swings on Vtt between 1.3 and 1.6 V, with EC10=2200uF+100uF. Went to Radio Shack today to try out the Q6 fix, found out almost immediately how fragile the original part was.. Sunk the replacement regulator into place with heat sink. Brought the machine back up and didn't notice any improvement in Vtt. Decided to log out of X and do some testing with cpuburn.

System is using 2x Celeron 433 clocked at 433. 128K L2 cache.
First, I fire up a loop of sensor monitoring to check on Vtt. With the machine otherwise idle, Vtt ranges between 1.45 and 1.55 V. I start some burnP6 processes and continue monitoring that Vtt is between 1.4 and 1.55 V. Okay.. so I start a burnBX and that terminates after a few seconds with return code 254 (Memory error). I keep running that a few more times, and I notice that now Vtt is swinging more widely. I see the same thing when I run burnMMX instead (which terminates even more quickly than burnBX).

So I run burnMMX with different memory allocations. With 64K allocated it should run pretty much from the L2 cache, and here I see only the nominal Vtt swinging and burnMMX does not terminate with error. But when I run burnMMX with 128K or higher (which should make it page to memory banks) it begins again to terminate quickly with error.

The widest swing I have seen is from 1.15 V to 1.65 V.

Might it be that the value of EC10 is too high?

Oh, and yes the references at the beginning of this thread are stale.

[hyperspace]

Oh, and yes the references at the beginning of this thread are stale.

Links are fixed!

[Dave Rave]

with the ec10=2200uf+100uf
that sounds like you have them in parallel.
which means you have lowered your 100uf down to 95uf making it worse.
if you had two 2200uf only, it would be a good 1100uf.

[hyperspace]

with the ec10=2200uf+100uf
that sounds like you have them in parallel.
which means you have lowered your 100uf down to 95uf making it worse.
if you had two 2200uf only, it would be a good 1100uf.

Capacitors in parallel... Add

CT = C1 + C2 + C3 + ···

would be 2300 µF

two capacitors in series...

CT = (C1 × C2)/(C1 + C2)

would be 95.65 µF