PC Reboots When Maximum RAM is Installed

kc27

Distinguished
Jun 24, 2010
20
0
18,510
I have a PC with a Gigabyte GA-8SQ800 motherboard running Windows XP Home, Service Pack 3 The motherboard has four 184-pin DDR DIMM sockets. It supports memory sticks in the amount of 128MB, 256MB, 512MB AND 1GB of unbuffered DRAM. The maximum amount of memory it supports is 4GB.

The computer has always ran fine with two 512MB sticks of RAM installed (Kingston #KVR400X64C3A). I attempted to upgrade the amount of RAM to the maximum of 4 GB, and installed four 1GB sticks of 1GB RAM in the PC (Kingston #KVR333X64C25). After I did that, the computer started having issues. The PC could run fine for hours or as long as a day or two, then suddenly it would reboot. When Win XP came back up, I would see a "this system recovered from a serious error" message. The reboots were sometimes preceded by me starting a browser or email client. Sometimes it would just reboot on its own.

To troubleshoot the RAM itself, I ran MemTest on the RAM, and it passed. I removed all but one stick of 1GB RAM and tried running Win XP with that amount of memory. Still got the error. Tried the 1 GB stick in different sockets. Still got the error. Tried a different 1GB stick in different sockets, Still got the error. It doesn't matter which of the four DIMM sockets is used, which stick of RAM is used, or even if I put a 1GB stick in some or all of the sockets. I am using 1GB sticks of Kingston value RAM, number KVR333X64C25. I even borrowed a different stick of 1GB Kingston #KVR400X64C3A RAM, which is the same model number as the 512mb sticks I was using, and experienced the same problem. I also had the computer do this when I was running a utility from a boot disk, so Windows wasn't even in the equation at that time. Heat is not a factor, either. I've added additional cooling..

The PC is still useful to me, and I suppose if it does not like 1GB sticks of RAM I can just avoid them. But if someone has any ideas they can pass along on what might cause the motherboard not to be able to support 1 GB RAM sticks, I would appreciate the help.

Thanks in advance for your time.
 

bilbat

Splendid
Still amazes me when someone 'pops up' with one whose manual I don't have!
0265l.jpg

This one was interesting - found several GB sites for it, mostly under an 'alias': sinxp1394, a number of those had 'broken' links to the manual, and none I could find had the, now, pretty much 'standard' memory compatibility list, which I was hoping might tell me something via deduction - however, I did find an FAQ entry that implied, in another context, that it should support 4G of RAM...

The BIOS/chipset appear to be too early to have a RAM timings subsection, but I did find a few things you might try:

As saint mentioned, make sure the RAM voltage is correct for your modules... You may also need to 'bump it' a tenth to handle the extra RAM...

On the main BIOS page, try 'disabling' the "Top Performance" item - my guess, it'll give you 'looser' timings overall...

It appears to only support two RAM speeds, 266, and 333 - try both, but don't leave it in 'auto' - these older modules have really minimal 'on-DIMM' PPD info (like, but much inferior to, today's SPD)...

With "Top Performance" disabled, try first, the "Load Optimized Default" on the main page; if that fails to produce the desired results, try "Load Fail-Safe Defaults"...

 

kc27

Distinguished
Jun 24, 2010
20
0
18,510
Thanks for the ideas on this

My manual is here GA-8SQ800 Info That is where I got the memory specs.

Here is an excerpt from my manual:

DRAM Voltage Control
>> Normal Set DRAM Voltage Control to Normal. (Default value)
>> +0.1V Set DRAM Voltage Control to +0.1V

Am I interpreting this correctly in that if I set it to Normal it will be 2.5 v(which is the correct voltage for my ram), then use the second option to move it up .1v?

I do have the DRAM speed set to auto. I will change that and the performance settings and see how it goes

Thanks again for the advice.
 

bilbat

Splendid
DRAM Voltage Control
>> Normal Set DRAM Voltage Control to Normal. (Default value)
>> +0.1V Set DRAM Voltage Control to +0.1V

Am I interpreting this correctly in that if I set it to Normal it will be 2.5 v(which is the correct voltage for my ram), then use the second option to move it up .1v?

That would be my assumption, as well... In the interim, I want to dig a bit, to see if I can find out what was cantained for set-up info on the PPD in these modules - may have it as close as my JEDEC directory - there's literally hundreds of items there I haven't 'wandered into' yet!
 

bilbat

Splendid
Yup - handy! First JEDEC I looked at:
has only eight bits of PD data, combined with a couple ID bits; bits PD1 through 4 are 'organization' selects - # of rows, columns, parity, ECC, and refresh speeds; remainder:
0266y.jpg

So, not surprising there's not a lot of adjustment range - not much defined to begin with!
 

kc27

Distinguished
Jun 24, 2010
20
0
18,510
I didn't have any luck with the suggested changes. I was two versions behind the last stable bios release, so I updated the bios and will see how that goes. There is nothing in the release notes to suggest any memory handling changes were made, but I suppose it is worth a shot.
 

bilbat

Splendid
Still hopeful - here's the part I didn't 'plop in' last time:
0267l.jpg

Note the 'TBD's in the lower rows - JEDEC shorthand (to be determined) for "the technology's developing fastrn' the damned standard!" Hopefully, between the time your 'shipped' BIOS was done, and the most recent issue, some of those rows 'got filled in' - in both places, the DIMM's PPD, and the new BIOS!

Good luck...


Bill
 

kc27

Distinguished
Jun 24, 2010
20
0
18,510
The updated BIOS didn't resolve the problem. I'm going to try a clean install of Win XP. If that doesn't do it, then I'll just cut my losses and stick to RAM sticks that are 512mb or smaller.
 

bilbat

Splendid
I can only think of one more suggestion: compare the manual to the actual, new BIOS, and let me know if you see any settings that exist in the BIOS, but are not in the manual.... You might also try doing a <CTRL><F1> at the main page, before looking through the rest - that's GB's 'secret' key combo for 'unlocking' hideen menus and settings!
 

kc27

Distinguished
Jun 24, 2010
20
0
18,510
I did the CTRL + F! key combination at the BIOS main menu It added an "Advanced Chipset Features" section to the BIOS menu.

Under Advanced Chipset Features I found the following:

Configure DRAM Timing [Auto]
CAS Latency Setting [Auto]
DRAM RAS Active Time [6T]
DRAM RAS Precharge Time [3T]
DRAM RAS to CAS Delay {3T]
Dual Channel Mode [Enabled]
AGP Aperture size [64MB]
AGP Fast Write Support [Auto]


Not sure what I should be doing with any of these settings, other than leaving them at "Auto"

 

bilbat

Splendid
AhhA! I've got the data sheet - will have to do a crapper-load of calculation, and get back. Also want to find out if there's some old program floating around on the web that will allow us to 'read' the PPD data, in cycles, rather than nanoseconds...
 

bilbat

Splendid
Ok - having a bit of difficulty here; need some more info... As usual, items 'hidden' under the <CTRL><F1> are not documented!

Change the "Configure DRAM Timing" item to 'manual'
highlight each one of these, and tell me what choices are available:
CAS Latency Setting - particularly interested if this one has 'fractional' choices - RAM spec sheet says it's CAS '2.5' - set to 3 if 2.5 not available
DRAM RAS Active Time [6T] this one I'm pretty sure should be 7: 42ns / 6ns per cycle = 7
DRAM RAS Precharge Time [3T] (not on spec at all, but usually equal to CAS...)
DRAM RAS to CAS Delay [3T] (same as above...)
Dual Channel Mode [Enabled] (leave)
AGP Aperture size [64MB]
AGP Fast Write Support [Auto] (what's available here?)
 

kc27

Distinguished
Jun 24, 2010
20
0
18,510
bilbat - Thanks for your efforts on this. I will go through this tonight and reply with the options. I did pop in a different hard drive and do a clean install of Win XP Home, Service Pack 3. It did not solve the problem.

I had a problem similar to this in the past that turned out to be video card related. A few months ago I switched to a "HIS H26XQT512ANP Radeon HD 2600XT 512MB 128-bit GDDR3 AGP 4X/8X HDCP Ready IceQ Turbo Video Card" I am thinking I should dig up my old AGP card and see if that quiets things down. Several of the user reviews of this card on newegg.com outlined similar problems.

 

kc27

Distinguished
Jun 24, 2010
20
0
18,510
Thanks again for the help.

Here are the additional settings and their options available in the BIOS. Also, below I included links to screen-shots of my BIOS screens.

This is a screen-shot of a summary of the system memory taken with the "System Information for Windows" program.

System Memory


Here are the BIOS memory settings

CAS Latency Setting
Auto 3T 2.5T 2T


DRAM RAS Active Time
6T 9T 8T 4T 5T 7T


DRAM RAS Precharge Time
3T 4T 2T


DRAM MS to CAS Delay
3T 4T 2T


Dual Channel Mode
Enabled Disabled


AGP Aperture Size
64MB 33MB 16MB 8MB 4MB 256MB 128MB


AGP Fast Write Support
Auto Enabled


Here are links to screen-shots of my BIOS screens
:
Standard CMOS Features

Advanced BIOS Featurs

Advanced Chipset Features

Integrated Peripherals

Power Management Setup

PNP PCI Configurations

PC Health Status

Frequency Voltage Control
 

bilbat

Splendid
Definitely, try 'er with the other card...

Only thing I can see that's obvious is that RAS active at six; try switching 'Configure DRAM timing' to 'manual', and set that one to seven; I am disappointed that 'AGP Fast Write' doesn't seem to have a 'disable' setting - that's what I would have tried' also, on Advanced BIOS page, might try switching 'Flexible AGP 8X' to '$X' to 'lock 'er up'...
 

kc27

Distinguished
Jun 24, 2010
20
0
18,510
I may have this resolved.

found the same model motherboard that I have (GA-8SQ800) for sale, so I purchased and installed it, and had trouble booting. I then swapped out the power supply, and that got the computer going. It's been nearly a week, and so far I have not experienced any of the random reboots, where before I was getting them multiple times a day,
sometimes multiple times an hour.

The replacement board has its own peculiarities, It won't POST with more than 3.5gb of ram (it's supposed to be able to handle up to 4gb). To keep all the RAM the same size and type, I just put in three 1gb sticks. This differs from the original board, which would post with 4gb of ram, but would also give the random reboots. Maybe the power supply issue was causing problems. When I get another power supply I will do some testing with the original board, since I swapped motherboards before I hit on replacing the power supply.

I used a Thermaltake Dr. Power #A2358 ATX power supply tester as a double check of my reasoning that the power supply was an issue. The LED voltage indicators light up to show the voltage is being output, but the tester also supposed to beep with the power supply is switched on, and it does not beep when testing the original power supply. The replacement power supply does elicit the beep. The power supply documentation was minimal on the product package and the website, so I don't really know what the beeps are supposed to signify.
 

bilbat

Splendid
Did/does the beep occur rapidly - say, within a half-second of power-up? May be the test of the PWR_OK signal. It's fuction is covered in the last bit of the 'sticky', under Power Supply - Basic. There are troubles that the average tester won't find - the obvious one is testing rails under full load; the not-so-obvious one is that capacitors age (and just sometimes randomly go to hell!), and if the rail's filtering has become defunct, you can be getting enough power-line harmonics 'leaking through' to cause problems for the board's voltage regulators...

documentation was minimal on the product package and the website, so I don't really know what the beeps are supposed to signify.
...don'tcha just love what passes for 'documentation' these days :kaola:
 

kc27

Distinguished
Jun 24, 2010
20
0
18,510
Thanks for the link on power supplies. Yes, the beep occurs nearly simultaneously with power up.

I wrote to Thermaltake asking what the beeps indicate about the power supply, and specifically what it means if your power supply test shows lights but no beep. I did test with another power supply just to confirm that the tester was capable of beeping, and that one did produce a beep. Below are the product description and user instructions.


Thermaltake Dr. Power A2358 ATX Power Supply Tester
For ATX 12V 2.0 power supply. (backward support earlier version power supplies)

Quick and easy to test +3.3V, +5V, and +12V voltage output

Compatible for Molex connector, P4, Floppy, S-ATA(+5V/+12V), 20-pin or 24-pin connector

Green LEDs and beeps allows user to know testing result. If the power output is working, the LED will light and you will hear a sound, if the power output fails, you won't see the LED light & no sound

Built-in load generates steady output
 

kc27

Distinguished
Jun 24, 2010
20
0
18,510
It's not just their documentation, their support on the item is pretty weak, too. I asked twice what the beep indicates, and what it means if the tester beeps when testing power supply A, but doesn't beep when testing power supply B. They don't have an answer, but they do want me to return it for an exchange.

For all I know it is working correctly, and that the lack of a beep on the one power supply ndicates it has detected an error.
 

bilbat

Splendid
You know - we asked for this! This is exactly what you get when you give a hundred-and-thirty-seven assorted bureaucrats power over each manufacturing concern - gotta kiss the EPA'a behind, gotta kiss the SEC's behind, gotta suck up to OSHA, gotta file twelve various environmental impact statements, gotta fight with the union to keep your essentially skilless help ("stand here, put the part in there, flip the yellow lever to clamp it, push the blue button"), pretty soon, the manufacturers figure out it makes more sense to just go elsewhere... And we lose sixteen million manufacturing jobs, and wind up with Elbonian mud-weasels doing tech support, and 'Mandenglish' semi-translated manuals!

I recently had a garage sale, to get rid of a lot of the junk accumulated around here, and while I was sitting in the garage, and looking at it all, I realized that nearly all of it had something that's become rarer and rarer in this day and age - a "Made in the US" sticker!