Jump to content
Corsair Community

Problems when using 3 DIMMS - BSOD


blaise

Recommended Posts

I am having a problem when I load all 3 slots in my motherboard with RAM.

 

I am using CORSAIR ValueSelect 512MB 168-Pin SDRAM Unbuffered PC 133 System Memory - OEM Model #: VS512MB133. My motherboard specs: 168-pin DIMM Banking, 3 (3 banks of 1), Chipset=VIA Apollo Pro133A, Error Detection Support= ECC and non-ECC, Graphics Support = AGP 4X, Max Unbuffered SDR SDRAM = 1536MB, Module Types Supported = Unbuffered only, SDR SDRAM Frequencies = PC100 and PC133, Supported DRAM Types

= SDR SDRAM only, USB Support = 1.x Compliant.

 

My system runs just fine when I have only 2 DIMMS loaded with 512MB each. As soon as I place a DIMM in the 3rd slot I get all kinds of stability issues, including the blue screen of death.

 

I have ran memtest86+ on each DIMM and on each DIMM placed in each of the 3 slots. Every test will pass except when I run all 3 DIMMS at one time. I will get numerous errors especially on test #5. So this rules out the memory and slots as being a problem. I know the RAM is compatible, but for some reason my system doesn't like it when all 3 slots are full!

 

I wonder if my power supply could be the culprit? It seems to be kind of a cheap and weak power supply: Sun Cheer 235 WATT, 3.3v@14a, 5v@23a, 12v@8a. Could it be that the PSU is not putting out enough power to run all 3 slots and the 1.5GB of RAM. Or could it be that my power management settings are set wrong in the BIOS. I have the HAL set to Standard Computer.

 

Also, I read somewhere that when using over 2 slots or over 1GB of RAM you should use conservative settings in you BIOS and that you should also use ECC RAM. Is this true? Could this be my problem?

 

Any help would be appreciated! I would hate to waste $80 by letting that extra 512MB RAM DIMM sit on my desk.

Link to comment
Share on other sites

Welcome to the forums, blaise!

 

Yes, when using that many DIMM's, you should probably relax the latencies in your BIOS. While I'm not sure what settings you've got to pick from, look for:

 

CAS Latency

RAS to CAS Delay (tRCD)

Row Precharge (tRP)

Cycle Time (tRAS)

 

I would add 1 to each of the first three and 2 to tRAS and see if that helps. You could also add 0.1v to the memory voltage.

 

If you can watch the voltages in your BIOS, see what the rails are doing and how much they are fluctuating with all three modules.

 

PLUR

CK

Link to comment
Share on other sites

Thanks for the reply.

 

I wish I could relax the latencies, but my BIOS will not give me that option.

 

The only options my BISO gives me are to use SPD mode or set the DRAM frequency, DRAM CAS Latency, DRAM Integrity mode and CPU In order queue manually.

 

I will disable the SPD mode and try setting the DRAM frequency down to 100Mhz instead of 133Mhz. I will also set the DRAM Latency to 3. What should the CPU in order queue be set to?

 

There is no option to change the memory voltage. I can monitor it and it runs about 3.3 - 3.4 volts. Is this OK?

 

I have also noticed that with 3 modules installed, the CPU temp. will run higher (97F instead of 86F).

 

Any other suggestions?

 

Thanks,

Mike

Link to comment
Share on other sites

Actually when I am running at PC133 or SPD is when I have the problems, so the only other more relaxed setting is to try is PC100.

 

I have the CAS latency set to 3.

 

The "CPU In order queue" is set to 4-level.

 

There is also an option for "CPU Host/PCI Clock (MHz)" this is set at Auto.

"CPU Ratio Selection" is set to 4.0X

 

Should I change these, or is their any thing else I can try?

Link to comment
Share on other sites

No, those are all fine... but the speed of the RAM might be the key issue, the chipset might just not be able to handle 3 modules at 133 rather than 100... if you increase the memory voltage up just one more 0.1 and that doesn't work, then it may just be a chipset limitation.

 

PLUR

CK

Link to comment
Share on other sites

Would it be better to run 1G of RAM at 133Mhz or 1.5G of RAM at 100Mhz? Which option would provide the best performance?

 

As I stated in my first message, my system runs stable with (2) 512M DIMMS of RAM. I really haven't had a chance to try 1.5G ( 3 512M DIMMS) at 100Mhz yet. However, I know that 1.5G will not work with SPD enabled or the 133Mhz option selected.

 

I am at work right now and I have my system running memtest86+ at home with (3) 512M DIMMS running at 100Mhz, 3 CAS. We will see if I get any errors.

 

There is no option to change the memory voltage. I can monitor it only and it runs about 3.3 - 3.4 volts. Is the only way to change this voltage through the BIOS?

 

I have also noticed that with 3 modules installed, the CPU temp. will run higher (97F instead of 85F with 2 modules installed). What should the CPU temperature be running at?

 

For what its worth, I have a VIA 694X chipset. Pentium III FSB @ 133Mhz, AGP 4X and PCI plus advanced ECC Memory Controller, Supports PC100/133 SDRAM, VCM, & ESDRAM technology.

 

Thanks for all the help..

Link to comment
Share on other sites

I'm not really sure, but I highly doubt that your system can even run a program that would require more than 1GB of memory, thus I'm inclined to say that 1GB at PC133 is better than 1.5GB at PC100.

 

Yes, the BIOS would be the only place to change the voltage on your system.

 

PLUR

CK

Link to comment
Share on other sites

Well I ran memtest86+ with all (3) 512M DIMMS installed. And it passed 4 runs! This is after I played with my BIOS settings a little.

 

Here are the changes I made to my BIOS:

 

Disabled system BIOS cacheable and all shadowing. Enabled Cache Bus ECC. Disabled power management. Disabled SDRAM timing by SPD. Set DRAM frequency to 100MHz and CAS latency to 3. Disabled ClkGen Spread Spectrum. Set CPU Host / PCI Clock to auto. Changed CPU ratio selection from 5.0X to 4.0X. Changed CPU Vcore selection from Auto to 1.70V. I am really not to confident with messing with BIOS settings, but this is what I came up with from all the research I have done. There are still many functions in the BIOS that baffle me, and I can’t seem to find any information that can help me.

 

I am still looking into a new PSU, and I am trying to find out how to raise the memory voltage from MSI tech support (I guess there might be a jumper I can change).

 

I am going to try messing with the BIOS some more. Any suggestions?

Link to comment
Share on other sites

My DRAMM ECC setting is disabled.

 

I enabled the Cache Bus ECC.

I thought this would give me better stability because it catches and corrects single bit errors.

 

My system eventually crashed again with the "3 DIMMs @ 100Mhz" settings I listed above. It seems like the CPU temp was running high - in the 100F range. Is this to high? What should it be running at? Could this be from the weak power supply? Sun Cheer 235 WATT, 3.3v @ 14a, 5v @ 23a, 12v @ 8a.

 

 

DRAM ECC must be disabled. But if the system is stable I would run at these settings.
Link to comment
Share on other sites

Turn off all ECC functions, put the CPU back to 5x and run Memtest86+ to see if you're getting memory errors...

 

100 Farenheit isn't all that bad for a processor... the max die temperature is usually much higher than that... however, if you let Memtest run for several full passes and start accumulating errors after the first couple, it could be a heat issue.

 

PLUR

CK

Link to comment
Share on other sites

My system seems to run more stable (less shutdowns) with the L2 Cache Bus ECC enabled. I only ran one memtest86+ with this enabled and it passed numerous runs with (3) 512M DIMMS at 133Mhz.

 

memtest86+ results:

(3) 512M DIMMS at 133Mhz = fails numerous times during test 5, 6 and 7.

(3) 512M DIMMS at 100Mhz = Passes every run, ran 5 runs in a row and passed all of them.

(3) 512M DIMMS at 133Mhz = when CPU Host / PCI Clock is set to 122/30.50 with ClkGen Spread Spectrum is enabled, test will pass every run.

(3) 512M DIMMS at 133Mhz = when CPU Host / PCI Clock is set to 129/32.30with ClkGen Spread Spectrum is enabled, test will fail during tests 5, 6 and 7.

 

(2) 512M DIMMS at 133Mhz = passes every run.

 

128mb of ram consumes around around 8 to 10watts. I have 1.5G of RAM so that consumes about 120watts. At 3.3V this is 36.36amps. My power supply is only rated for 14amps at 3.3V. So this seems that I am only running at half the amps that I should be. Do you think this could be causing my problems??

Link to comment
Share on other sites

That would make it about 11 amps of power required to run the memory. So do you guys think my problems could still be caused by my power supply, since my PS is only putting out 14 amps at 3.3 volts. I am sure their must be other componets that use 3.3v on my mother board.

 

No its about 12 watts per module max no matter what the density of the IC is.
Link to comment
Share on other sites

All the specs that I have found and even Micro Star technical support states that the mother board is supposed be able to handle this type and amount of memory.

 

Has any one heard of this problem or even had this problem?

Link to comment
Share on other sites

  • Corsair Employees

All I can tell you would be if the modules all run one up at the tested spec but not when all three are installed that would suggest a issue with the MB.

And at the time this MB was developed the largest modules on the market were 256 Meg and some MB's based on this chipset would need to run the memory at PC-100 speed when all 3 slots were filled with double ranked modules.

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...