3dfx Archive | |
http://www.falconfly.de/cgi-bin/yabb2/YaBB.pl
This & That >> This & That >> BAD CMOS checksum error http://www.falconfly.de/cgi-bin/yabb2/YaBB.pl?num=1179570944 Message started by Obi-Wan_Kenobi on 19.05.07 at 12:35:44 |
Title: BAD CMOS checksum error Post by Obi-Wan_Kenobi on 19.05.07 at 12:35:44
Hi all I have this scary thing with my CrossFire machine.
the following parts are: AMD Athlon64 X2 4400+ Toledo 2x 1GB PC3200 Corsair XMS Platinum CL2 ASUS A8R32 MVP Deluxe AMD 580X chipset aka ATi CrossFire Xpress 3200 Sapphire X1900 XTX Sapphire X1950 XTX CrossFire Edition 3dfx Voodoo 5 5500 PCI Macintosh Creative Audigy2 Platinum the motherboard has been replaced just by trying to fix this problem, but that didn't seem to help, what could this problem be? a dead CPU maybe? since I have tested the old as new motherboard the same error keeps on occuring right after the VGA bootscreen of the X1900XTX as X1950 XTX CFE as 5500 PCI Mac, I booted the system with all 3 different cards as primary VGA and that didn't help either. :-/ Also I took usage of the bios recovery floppy but that didn't seem to work on the old mobo so I thought that it would be pointless to do it on the new one as well. |
Title: Re: BAD CMOS checksum error Post by MrFossey on 19.05.07 at 12:45:59 Quote:
Quote:
P.O.S.T. means Power On Self Test. That's the procedure where the system is checking the CPU, the RAM, harddrives, opticals and when errors like yours pop up. Quote:
You might give it a shot, though. But not before you're sure the battery on the motherboard is correctly. Check if it's seated correctly and if possible replace it. Also check wheter the CMOS Clear jumper is installed correctly. |
Title: Re: BAD CMOS checksum error Post by Obi-Wan_Kenobi on 19.05.07 at 13:21:20
well after all tests a friend on mine found out the when using 2 ram modules in dual DDR mode the error occured when only using 1 ram modules everything was cool.
So we found out that the memory controller which is located in the northbridge aka the CPU is broken. time to replace it I guess ::) |
Title: Re: BAD CMOS checksum error Post by outlaw on 19.05.07 at 13:27:23
Try other mem modules as well in varius combinations of dual channel. There is no mem ctrlr in any 939 chipset. A64s have imcs and some of them behave pretty strange with several mem chips (ex. hynix high latency mem chips).
|
Title: Re: BAD CMOS checksum error Post by MrFossey on 19.05.07 at 14:26:02
@outlaw: A part of the Northbridge is located in the AMD Athlon64 processor. Originally the memory controller located in the Northbridge. This is still the case but since a part of the Northbridge is in the CPU the memory controller is in the CPU ;).
Obi-wan Kenobi already mentioned that a bit: Quote:
It's indeed possible the CPU is broken but I suggest you test a bit more before getting a conclusion yet. Try every RAM combination Try different chips. Try to fiddle around with the latencies (timings) and if possible the RAM voltage. The CPU works at a low voltage, the RAM works at a high voltage. In some cases extreme differences (default or lower voltage for the CPU, but 3,2V+ for the DDR-SDRAM memory) can cause problems and might even damage the CPU! |
Title: Re: BAD CMOS checksum error Post by outlaw on 19.05.07 at 16:01:21 wrote on 19.05.07 at 14:26:02:
Can't understand whats the difference in mentioning the mem controller as a part of the northbridge and then some part of the northbridge in the cpu......the northbridge chip is the North Bridge....end of story. A chip located on the board. The imc is in the cpu and the northbridge has nothing to do with that, don't confuse things. The other fact you mention has to do with mem vtt that passes through the imc. The only problem with high vdimm is when you have high difference between vcore and vtt. That can be overcome by raising vcore to shrink the difference between the two voltages. As simple as that. |
Title: Re: BAD CMOS checksum error Post by FalconFly on 19.05.07 at 18:23:23
The possibility of the CPU being "broken" is extremely low.
A much higher possibility exists that either a RAM module became shaky or defective or simply dirty. Signal errors are the most common reason for RAM related problems on modern machines. Can be caused by dust, thermal effects or any other effect that increases a reduced signal-to-noise ratio on the Memory signal lines. If nothing helps (e.g. clean, use of other RAM slots, deactivating hot timings e.g. 1T Cmd), going to a lower RAM clock often helps. What needs to be done is to gain enough stability to run memtest86 or MicroSoft RAM tester, and ideal case Prime95. Many 'odd' errors can also cause weird system behaviour, beginning with damages to Data cables (HDD, Disk, CD/DVD), ageing Power Supply and static charging on the Motherboard (Dirt, bent Case mounting or alike). It just requires alot of testing and a healthy amount of spare components to isolate the cause. The best start is to remove all unneeded components (e.g. take only one basic PCI Card) and then begin testing. |
Title: Re: BAD CMOS checksum error Post by MrFossey on 19.05.07 at 21:23:17
[offtopic]
wrote on 19.05.07 at 16:01:21:
So basically the Memory Controller is in the Northbridge part of the CPU. Don't believe me? :P See this picture Source: AMD Functional Data Sheet, 939-Pin Package [/offtopic] |
Title: Re: BAD CMOS checksum error Post by Obi-Wan_Kenobi on 19.05.07 at 23:08:47
well people the strange thing is that my setup has been running stable without problems for like 12 months and the PC store where I bought the gear has become very unreliable so sending stuff back would give an outcome of no garrantee and I have had to many troubles with things like that so therefore I kinda left that out of mind as well.
The ram I used was reccomended for this motherboard, I used class A TwinX 2x 1GB PC3200 CL ram modules from Corsair XMS Platinum. Just good DDR400 modules the system had been working in Dual DDR 400 mode for 12 months and last monday morning when I wanted to do some CrossFire research the problem occured, since I don't do silly things like OCing and volt mods and stuff I suppose it's just a macanical failure. Also I have removed all the parts from to the Voodoo5 CPU and one ram module. With a single stick of 1GB PC3200 in either of the 4 slots the system boots without a problem but If I use both in either the 2 blue or black slots I get the error, this kinda explained it\, the error even occured when I put 2x 512MB PC2700 from Kingston in it, so this explained that my Athlon64 X2 4400+ couldn't drive the Dual DDR setup , the same happened on the new as the old Mobo, so it would be very strange if the new one would be broken all of a sudden right :) the CPU gives the problem as far as we could find out, the same happened when we used the X1900 XTX as the X1950 XTX CFE as the Voodoo5 the VGA cards are in good shape so are the memory modules it's the CPU which can't drive the Dual DDR setup, only single DDR works on both boards with all 3 VGA cards. @ Mr Fossy You are correct the Memory Controller is in the Northbridge part of the CPU, but this was something I allready knew 2 years ago every A64 design has the northbidge in them the smae goes for Opterons as Sempron64's |
Title: Re: BAD CMOS checksum error Post by paulpsomiadis on 20.05.07 at 05:12:24
Dang! That's a bit of a B!TCH... :o
...hope you can get a good deal on a replacement CPU! :-/ (and this time maybe a bigger/more efficient CPU cooler - although I'm just guessing with that one!) ::) |
Title: Re: BAD CMOS checksum error Post by Obi-Wan_Kenobi on 21.05.07 at 21:17:49
yeah I did find something of better quallity a nice Opteron 180 :) no more Athlons here just Opterons my experience with Opterons has been better , just better parts I guess
|
3dfx Archive » Powered by YaBB 2.4! YaBB © 2000-2009. All Rights Reserved. |