Unstable for years and fed up... help!!!!

[Crazified]

Hi everyone! I'll try and keep it short and sweet. any help would be GREATLY appreciated!

I got this system a couple of years ago and it's been unstable ever since. Here's the lowdown:

ASUS p4t533
512mb 1066 RDRAM chips
(2) 80gb drives raid 0
2.53mhz intel p4
550 watt pws

This machine will blue screen at the drop of a hat. Games are essentially impossible to play; bombing out after anywhere from a few mintues to up to 20 minutes. having the machine sit idle will typically make it "last longer", but it will eventually blue screen and reboot (or lock up).

I read other postings regarding "instablility" and went ahead and downloaded memtest86. Here's a screenshot (if it works). Ok... I can't figure out how (if I even can) attach a screenshot. So I'll type out what it says:

Pass 6%
Pentium 4 (0.13) 2539 MHz test 77%
L1 cache: 8 20810mb/s test #3 [moving inversions, 8 bit pattern]
L2 cache: 512k 17754mb/s testing 108k - 512m 512m
memory: 512m 1503mb/s pattern 02020202
chipset: intel i850 (ecc: disabled) /fsb : 133mhz

I was running this thing for over an hour with failures galore... here's some more stats:

walltime: 1:24:16
cached: 512m
rsvdmem: 148k
memmap: e820-std
cache: on
ecc: off
test: std
pass: 0
errors: 9221986
ecc errs: 0

you really should have seen the "red" flying by on this thing.

OK... SO WHAT DO I DO???? Is this really a massive failure of my RAM, or are my settings somehow run to make it fail? I've been running the machine (to type this note) for about 30 minutes.... but it's really a crap shoot. Sometimes explorere fails and shuts down, sometimes it hangs, sometimes it blue screens. It will ALWAYS eventually run into a problem.

Is it as easy as I've got crappy RAM or are some of my settings off??

Any help would be GREATLY appreciated!

Thanks!!!!!!!

 

[Zepper]

Turn machine off. Unplug from wall. Clear the CMOS - that will restore to default settings. Reboot go into BIOS settings and adjust just the basic things like the time and date, drives, boot sequence, etc; and run memtest again to see what happens. If you still have errors, go into the BIOS settings again and look for a way to adjust the voltage on your memory. Adjust it upwards and retest one small step at a time to see if it gets any better. If not, replace the memory OR the whole mobo and go with one that supports DDR menory. I hope it doesn't come to that. I would also try them in another machine if you can find one that uses RDRAM amongst your friends - just to be sure that it's the RAM and not something else.
. If you are lucky and your memory is name-brand, you may have a lifetime warranty on the modules. Worth checking into if you have assured yourself that they are bad.
.bh.

 

[Rike]

Originally posted by: Crazified

ASUS p4t533
512mb 1066 RDRAM chips
(2) 80gb drives raid 0
2.53mhz intel p4
550 watt pws

T

2.53Mhz? I think you mean Ghz? But anyway, that doesn't sound like a standard speed. Is the machine OCed? If it is, that would be the first place to start. Get it clocked normally.

How long has this system been running this install of the system? And what OS? It sounds like my old P3 machine running Win98. Did a fresh install of XP Home and what do you know, everything was soomth.

 

[Malak]

Also, 2.53 ghz on a 133 mhz FSB??? Nothing sounds right.

 

[LTC8K6]

2.53Ghz is a standard speed. It's a P4B. 19 multiplier.

 

[Rike]

Originally posted by: LTC8K6
2.53Ghz is a standard speed. It's a P4B. 19 multiplier.

Opps, good call. Can't do math.

Edit: BTW, welcome to the AT forums Crazified! Lots of smart people here.

 

[JBDan]

There are quite a lot of possibilities here I have this same processor in a Dell and you have the 850 (or 850E) chipset which supports RDRAM 133 quad pumped=533MHz FSB. Start with this. What is your bluescreen error message? Any error messages upon reboot?

 

[MDE]

Try to find the standard timings and voltage for your memory and set those in the BIOS. If Memtest is giving you errors, then your memory is most likely the culprit.

 

[DaveSimmons]

If memtest is giving RAM errors, don't bother trying to re-install Windows!

Until your RAM is error-free your Windows will never work right and your files will get corrupted left and right.

Follow Zepper's instructions before doing anything else.

 

[JBDan]

Originally posted by: DaveSimmons
If memtest is giving RAM errors, don't bother trying to re-install Windows!

Until your RAM is error-free your Windows will never work right and your files will get corrupted left and right.

Follow Zepper's instructions before doing anything else.

Correct there!

 

[caz67]

Ah , the good old RDRAM days..!! I used to have that board.

 

[gsellis]

Are you using XP? I did not see a reference except in a response (or I could be blind). If it is not XP, what are the Stops (Blue Screens)? If you have not been able to see them and ignored them so far, try changing this registry setting HKLM \ System \ CurrentControlSet \ Control \ CrashControl \ AutoReboot to REG_DWORD to 0. This will stop the system at the Blue Screen and you can get some of the details. The Stop may (some do not) be documented in Eventvwr in the System section.

But, if memtest is having errors, that sure sounds like you may be onto something

 

[PingSpike]

I see no reason to troubleshoot windows issues until he can get this machine to pass memtest. Windows may or may not be causing problems, but given the rampant memtest failures...the memory is the most likely culprit and should be the first priority.

In short, follow Zepper's advice. If and when you get it passing memtest, it will likely be a good idea to reinstall windows since its been running some time with unstable memory.

 

[gsellis]

Originally posted by: PingSpike
I see no reason to troubleshoot windows issues until he can get this machine to pass memtest. Windows may or may not be causing problems, but given the rampant memtest failures...the memory is the most likely culprit and should be the first priority.

In short, follow Zepper's advice. If and when you get it passing memtest, it will likely be a good idea to reinstall windows since its been running some time with unstable memory.

I think it is memory too (memtest says so too). Checking the stops can help prove it above and beyond that. The more random the stops (A, 50, 7D, 7E, this module, that module), the more it points to memory from the Windows debug side. If a specific control fails though, it may point to another mobo issue that is affecting memory (low precentage! - it usually is memory).

Yes, Zepper's advice.

Just wanted to pass on the alternate check. Most people do not know how to turn off Autoreboot from the Reg. This is import if you are using PE (MS or Bart's) to try to get to the information about a crash that happens at the start of Windows XP. You can load the system hive on the machine to change the reg setting to at least see the Blue Screen info.

 

[jvarszegi]

Unstable for years and fed up... help!!!!

fat and deranged is no way to go through life

 

[otispunkmeyer]

errors: 9221986

that has to be the highest i ever saw

 

[CraigRT]

instability is quite often either memory or power related.. since memtest is failing, either up the voltage on the RAM, or try swapping it out.. and likewise for the PSU if problems persist.... just my best guessing. if I had the system here on my desk I am pretty sure I could fix it for ya, but I dunno quite for sure like this.

good luck!

 

[Crazified]

Hi everyone! First of all... thank you so much for all of the great feedback! There is DEFINITELY a lot of knowledge here and I'm grateful for all of the prompt replies.

To answer some of the questions that have been posted back:

-I'm running xp pro
-I've reset the CMOS... and still having problems.
-I haven't been at the machine when it blue screens, so I can't supply any data from that yet. Below is what windoes gives me upon the restart.
-I don't know anyone who uses RDRAM (to be able to test on)
-I can't seen to find a spot in the BIOS to adjust for voltage settings on the RAM? Does my system allow for it??

Ok: here's what Windows give me on the error report. I understand 1% of this:

BCCode: 24 BCP1:001902FA BCP2:B1BA88B8 BCP3:B18A85B8 BCP4:F83A270F OSVer: 5_1_2600 SP:1_0 PRODUCT: 256_1

Here's some more interesting facts:

I'm still running with Service Pack 1 because I can't get SP 2 to completely install before the machine crashes. I've given up on it. I can't get a full virus scan to complete either for the same reason.

When I first got the system, I couldn't even get WINDOWS to install. It would always bomb out before I could get it to complete. I called ASUS and they said to make sure that the board was properly grounded and told me to place electrical tape between the the board itself and all of the screws and the post holes to where the board mounted to the tray. That did the trick to get me up and running, but if I recall, the errors (ie -blue screens) were the same or similar.

I've edited the reg to have it not reboot when it crashes. As soon as I get the data from the screen, I'll pass it on.

Thanks again everyone for all of the help so far!

Oh... and if it turns out that it IS the memory... it's Samsung brand, so I assume it has lifetime warranty? How would I go about getting it swapped out? I bought it from mwave.com. Would I get replacement memory sent to me beofre I had to ship it out?

Thanks again!

 

[Crazified]

Ok... here are some stats from the blue screen. (It will vary... this is the first I can report on). What does this tell me?

PAGE_FAULT_IN_NONPAGED_AREA
Stop: 0x00000050 (0xc18ae31c, 0x00000000, 0x804ef379, 0x00000000)

I'll report more as they come up! Thanks!

 

[gsellis]

A 50 is almost always hardware IIRC. The memory sounds bad, mismatched, a bad PSU, or a bad mobo. Since you are getting the errors in Memtest, start there. I am guessing it has to be paired since you have 1066. Did you get 4 128's or 2 256's? If you have 4 sticks, pull 2 and see if you can find a pair that are stable.

 

[SickBeast]

Originally posted by: gsellis
A 50 is almost always hardware IIRC. The memory sounds bad, mismatched, a bad PSU, or a bad mobo. Since you are getting the errors in Memtest, start there. I am guessing it has to be paired since you have 1066. Did you get 4 128's or 2 256's? If you have 4 sticks, pull 2 and see if you can find a pair that are stable.

4 sticks of memory in a system can often lead to instability in itself.

The first thing I would do would be to raise the memory voltage, as someone suggested earlier. The other option is to change the memory timings. You might want to consult Samsung's website/documentation to find out what the reccomended timings for the memory are. If RDRAM uses CAS latency, I would lower that setting marginally.

I'm not convinced that it's a bad stick of memory, but repeating the memtest on each stick individually isn't a bad idea at all. :beer:

 

[stevty2889]

He can't do memtest on each stick individualy, since they have to be installed in pairs. What caught my attention is he said a 2 year old system with a 550w PSU. My guess is you are using a generic PSU...could you please let us know what brand your PSU is, and the amps it has on the rails, 3.3v rail, 12v rail, and 5v rail...And I know from experience, that a generic PSU, can cause memtest errors, blue screens, including the page fault errors you are getting. Thats the problems I was having with my 400 powmax(generic piece of crap). Replaced that POS psu, and my lock ups, crashes, BSOD, and memtest failures all went away. Of course that was after it took out a $180 motherboard, which I originaly thought was bad, then the same problems came back on the new motherboard. Fortunatly I came to my senses and realized it was the PSU, before it took out another motherboard..is it specificly the p4t533 that uses the 32bit rdram modules, the the p4t533-c that still uses 16 bit modules. If it has the 32bit modules, then they don't actulay have to be installed in pairs, so you could test your individual sticks.(not sure if you still need the continuity rimms or not, since thats the only board I know of that even supports 32 bit rimms).

 

[Crazified]

Ok... to answer some of the questions that you guys are asking (sorry I didn't post them before):

- I have (2) Samsung 4200 256mb chips. The board itself only has 2 slots to use.

- My PSU is an Antec TruePower 550. I don't exactly know what you mean by "amps it has on the rails", but this is what the box says:

"Individual +5v, +3.3v and +12v voltage feedbcak circuitry"
"Very low minium load: 0.8A at +12v"

Does that help?

- I just tried testing the sticks individually (with the continuity rimm in place) and it worked! What I found was interesting:

With the first stick installed and a single pass through all 10 tests in MemTest, I got a total of 326 errors (328 on test #4, 12 on test #6, and 26 on test #7). I swapped the one stick of the other, ran the test again and got 10 errors; all on the #4 test. BUT!!! I decided to let it continue to run and let it do multiple passes. I can back about 10 mintues later and the error count was up to about 1,100 and the machine had locked up.

Once I installed both sticks again, I ran MemTest again and got the same results as before (ie - errors counting into the Millions). How is that possible?

Thanks again guys!

 

[stevty2889]

Sounds like your ram is either bad, or needs the voltage adjusted. There should be some setings in your BIOS to adjust the ram voltage, or possibly a jumper. Look in your manual if you still have it. It may be under frequency/voltage control setting in bios, or maybe advanced chipset settings. If you have an Antec truepower, I doubt that is the problem, since Antec makes quality PSU's. To find out the amps, look at the sticker. under the voltage numbers, it will say for example 20a meaning 20 amps.

 

[Crazified]

Apparently there is no way to adjust the voltage for my RAM. I checked the maual cover to cover. No jumpers and, in the BIOS, all I've got is "regular" and "turbo" mode.