Bizarre (presumably) motherboard problem
For your consideration:
Epox 8RDA3+ (nFORCE2 Ultra 400) motherboard
Athlon XP 2500+ (unlocked) usually running @ 11.0x200 @ 1.70v
512MB Corsair XMS PC3200 C2 RAM @ default voltage
Windows XP Professional
Came home and found my system locked up. Pushed reset, PC speaker goes into alarm mode. C1 - C3 errors, which correspond with memory problems. Removed & reseated memory, no luck. Installed a 256MB Corsair VS PC3200 stick, no luck, installed another 256MB Corsair VS PC3200 stick, no luck. Put Corsair XMS back in, powered up, tapped insert, was able to get to BIOS.
Began changing CAS, TRAS, etc. timings, no luck, system always freezes upon reboot unless tapping insert. Up VDD voltage, no change, up VDIMM voltage, no change, up CPU voltage, no change.
Lower FSB to 185, no luck. Lower FSB to 180, boots just fine. Loads Windows no problem, and then the REALLY funky **** starts hitting the fan. Run a Sandra 2004 benchmark and see that my memory score has dropped from ~2800 to ~1400. WTF?, I say to myself.
At this point, I open up the My Computer properties and see that even though my processor is listed as an Athlon XP 3200 (at that point running 12.5x180), the CPU speed beneath is listed as 1.25GHz (12.5x100, obviously.)
Now, I don't have a clue what is going on. The BIOS verifies that I'm running 180 and/or 166 and/or anything between 100 and 180. Windows and Sandra, however, tell a far different tale. Attached a few pictures. The "Athlon XP 2700" photo shows the system running 11.0x180, and the "Athlon XP 2800" shows the system running 12.5x166, the third is just Sandra confirming the system running at 100MHz FSB.
I've been building systems for many years now, and I have never seen or heard of anything like this before. ESPECIALLY since the BIOS claims that all is cool and even Windows claims that the CPUs are 2700, 2800, 3200, etc speed but then also noting that they're running off of a 100MHz FSB.
I'd LOVE to get some input on this one. Has my mobo lost its mind? Has Windows? Have I?
Epox 8RDA3+ (nFORCE2 Ultra 400) motherboard
Athlon XP 2500+ (unlocked) usually running @ 11.0x200 @ 1.70v
512MB Corsair XMS PC3200 C2 RAM @ default voltage
Windows XP Professional
Came home and found my system locked up. Pushed reset, PC speaker goes into alarm mode. C1 - C3 errors, which correspond with memory problems. Removed & reseated memory, no luck. Installed a 256MB Corsair VS PC3200 stick, no luck, installed another 256MB Corsair VS PC3200 stick, no luck. Put Corsair XMS back in, powered up, tapped insert, was able to get to BIOS.
Began changing CAS, TRAS, etc. timings, no luck, system always freezes upon reboot unless tapping insert. Up VDD voltage, no change, up VDIMM voltage, no change, up CPU voltage, no change.
Lower FSB to 185, no luck. Lower FSB to 180, boots just fine. Loads Windows no problem, and then the REALLY funky **** starts hitting the fan. Run a Sandra 2004 benchmark and see that my memory score has dropped from ~2800 to ~1400. WTF?, I say to myself.
At this point, I open up the My Computer properties and see that even though my processor is listed as an Athlon XP 3200 (at that point running 12.5x180), the CPU speed beneath is listed as 1.25GHz (12.5x100, obviously.)
Now, I don't have a clue what is going on. The BIOS verifies that I'm running 180 and/or 166 and/or anything between 100 and 180. Windows and Sandra, however, tell a far different tale. Attached a few pictures. The "Athlon XP 2700" photo shows the system running 11.0x180, and the "Athlon XP 2800" shows the system running 12.5x166, the third is just Sandra confirming the system running at 100MHz FSB.
I've been building systems for many years now, and I have never seen or heard of anything like this before. ESPECIALLY since the BIOS claims that all is cool and even Windows claims that the CPUs are 2700, 2800, 3200, etc speed but then also noting that they're running off of a 100MHz FSB.
I'd LOVE to get some input on this one. Has my mobo lost its mind? Has Windows? Have I?
0
Comments
You could also try reflashing the BIOS (and go ahead and update the BIOS if there is one available).
Haven't run memtest, though I'm fairly certain it isn't the memory since I tried three different sticks. Also, when I first tried booting with the XMS, I never used the old Insert trick to get to the BIOS. I only resorted to tapping insert after three different PC3200 sticks failed to boot.
I presume that it could have always gotten to the BIOS using the insert trick.
I haven't reflashed the BIOS yet, but I am using the newest public release BIOS available. I could probably find myself a newer beta bios over at the AOA forums....
Do you think it's possible that the BIOS could have gotten itself corrupted during a non-flashing computer freeze? Never heard of such a thing happening, which isn't to say that it hasn't, of course.
Methinks me board be fooked. I'm well within warranty period, and I live an hour and a half from Newegg, so I'm sure I'll have a turn around time of a day or so... But I just don't want to have to pull all my **** out of my box, pack it up, send it off, wait a day, put it all back together again, bla, bla, bla. Dammit! I'm just pissed. Very stressed and, as a result, pissed.
The timing for all of this is, naturally, nothing short of impeccable. If I ever get my hands on that bastard Murphy, so help me....
On a lark, I opened the case one last time, turned off the power supply one last time, pulled the power cable off the board one last time, reset the CMOS one last time, put it all together again, rebooted, reconfigured BIOS, and lo-and-behold: problem fixed.
Back to running at 11.0x200
Set everything to stock in the BIOS. Boot up and uninstall the processor from Device Mangler, err... Manager. reboot and let it redetect the CPU.
Worth a shot!
EDIT: You're too quick for me, Jake! Glad you got it worked out.
/me hears a rousing "Or what"
Good, for sure; just mystified as to why the problem remained the first two times I hard reset the CMOS but went away on third. If I believed in numerology or any form of mysticism I'd make something of it; but I don't, so screw it.
Maybe you just left the jumper on a few seconds longer this time or something.
It might be time to worry about stuff like that.
Either that, or it might just be the pahse of the moon.
Ok, when you change the jumper, this happens:
CMOS is cleared with electricity. That electricity comes from the same battery that runs the clock. To clear it is has to zero the table values, which are only set by the BIOS(when the jumper is in normal mode) or by the clearing when jumper is in clear position.
Now, the clock batteries used have been engineered to send a certain flow of watts for about 90-95% of their life, but they also to do that can only send same watts out of a fixed charge "pool" of power for this time that is their active lifetime. And the flow is tuned to what is used when active.
To clear the CMOS, an inverse polarity current is sent. And that charge has to be negative throughout the time it takes for all positive charges (whihc computer bios sees as binary 1's) in CMOS to be cancelled out. The modern computers with bigger BIOS code need a bigger cmos table to store all teh option selections possible for BIOS to need to track. So, for modern computers you cactually need to have the jumper in clear longer for the fixed flow to wipe all the values to zero than for older ones which tracked fewer options and selections by using smaller BIOSs.
It takes LONGER for a bigger table to get zeroed than a smaller one when that table is stored in circuits that are designed to hold the values unless negatively charged. I have seen computers with 2MB BIOS code take 15 sec to clear the larger CMOS tables they need to work right, and if it takes 30 seconds I know I need a new CMOS\Clock battery as the only way the time can almost double is if the battery pool of watts is drained down to alower state than the computer "wants" for normal operation.
WHY??? Battery is designed for fixed flow. It has a limited charge to start with. It drains at afixed rate for most of its usable life.
Lets say you have a computer motherboard, and you want it to start right up when you hook it up. To have that happen you need to have battery jumpered active. Motherboards used to come with CMOS jumper NOT set to normal to keep battery from using juice while motherboard was being shipped (by water, BOAT for most cases) and stored on a shelf until sold and while on shelf before being shipped. A fixed lifetime battery should have its life in use while IN USE, not while sitting active and not in use. Soem assemblers of computer parts, including motherboards, do not bother to park the jumper at assembly line. So, if the jumper is active as soon as assembly is cmplete, the motherboard circuits somewhere drain it for all that time.
Batteries have fixed lifetimes on shelf, and can be stuck in partway through that lifetime or AFTER that lifetime has been exceeded by a bit. If they are left jumpered active for along time, they are not fully charged when you get them to start with. I have had this combo of things lead to a motherboard with a half or more discharged clock\CMOS power cell when they get to me as builder of system.
So, what happens when you leave jumper on for too short a time??? CMOS table is partly wiped, partly has values that are unpredictable, and is partly unwiped with old values still stored in part of CMOS. BIOS freaks when this happens, sometimes.
The other way to clear a CMOS is not to use the jumper, make sure no power is coming into or remains in board, but board is grounded. CMOS needs a small amount of juice to "keep the right values set." So, if you pull battery and motherboard capacitors are drained and you let it sit, electrical charge for a 1 gets turned into a 0 as far as BIOS is concerned-- eventually, time with motherboard hooked to studs, power off, capacitors on motherboard drained, half an hour to an hour of sitting will drain most CMOS tables. IF BIOS is seeing all zeros, it sets the CMOS to some base settings when computer is powered up, then looks and probes and tries to set things from basic options it has in code for "default."
But if BIOS sees codes coming in that it is not coded to deal with, it freaks-- no post results, motherboard looks dead. If battery gets drained too badly, CMOS table gets "scrambled," it has random settings. If battery is jumpered too long, you are doing nothing with CMOS table but you are running the battery down. Will not hurt CMOS to leave jumper on too long, just drain battery.
Take a stopwatch or watch second hand or counting on a watch-- give jumper in clear a 15 second time count, move jumper back to normal, reconnect the motherboard power you disconnected beofre clearing CMOS, and see if board posts. If no post, disconnect motherboard power jumper, wait until power LED on motherboard goes out so you know capacitors are drained, leave the jumper set to clear for twice as long. IF you now get a post, battery is now drained by a huge bunch(between having been left jumpered to normal (active) for too long and being drained more than you normally need by being left in clear for too long. Then, if no post, you replace battery, clear CMOS for 15 seconds, rejumper to normal, and then try for a post. If still not post, put old discharged CMOS cell back in, and RMA board. But a new CMOS cell has fixed a no post half the time I have gotten an apparently dead board in with the CMOS power jumper arriving in normal position when taken out of bag.
John D.