Overclock.net › Forums › Components › Memory › Interesting problem with Ballistix DDR2 1066
New Posts  All Forums:Forum Nav:

Interesting problem with Ballistix DDR2 1066

post #1 of 5
Thread Starter 
Yo memory noob here,




I've been on a mac for the last 5 years and I haven't had to tinker with my three year old gaming rig till now. I'm a little rusty on this stuff.




My rig has been ROCK solid till now even if it is getting a bit dated. Originally I had a 2GB (two 1GB) Ballistix kit running at 4-4-4-12 at 800mhz. I I decided the other day to pull the trigger on another set of used Ballistix 1066mhz RAM after I found out that the 1066 stuff is really just 800mhz overclocked. I read about a lot of people complaining that the 1066 wouldn't get up to speed correctly, etc. I figured I wouldn't have an issue because I just wanted to run the RAM at 800mhz, same as my old. So, I slapped the RAM in my system, fired it up and immediately found out that windows xp 32bit only supports 3.25gb of memory after I troubleshot for an hour. OK, that's fine. Beside that I was having other issues as well. Lots of BSODs, freezes when I would get past the BIOS screen to the windows screen, etc. I was running all my RAM at that point at 1.8v and 4-4-4-12 @ 800mhz. I'm not interested in getting to 1066 because I understand AMD memory controllers can only handle 1 DIMM per channel at that speed and I have four. (at least mine can't). I use memtest, all 4gb of memory comes back clean with a preliminary hour long test (I stopped at test 7).




Still getting issues. BSODs, unable to POST, freezes while booting windows with memory errors, etc. Classic memory errors. Long story short, I eventually run 3GB of RAM(including my old 2.0v kit which I thought was 1.8v) at 2.2v and with the BIOS auto settings of 5-5-5-18 to eliminate the possibility that 32bit winxp was not interfacing correctly with the oversized amount of RAM. Witcher 2 ran like a champ. Civilization 5 ran like a champ. BUT I was playing Civ 5 the other day for maybe 3 hours and the system crashed to a BSOD with the "stable" 3GB config with loose timing and high voltage.




I know it is the new memory. Maybe one of the DIMMs isn't flawed but I'm concerned about the fact that I haven't yet been able to get these things stable where I want them, which is 800mhz, 4-4-4-12, identically timed all the way through. Having said that, memtest comes back clean.




The RAM is not identical. My BIOS wants to run the new 1066 stuff at 125.7ns trfc (i hope that is the right number). The old RAM runs at 105ns. Since the new stuff is rated at a higher voltage, speed, etc, shouldn't I be able to run it at 4-4-4-12 at 800mhz with identical settings to my old ram? Could the fact that part of my ram is running at 125.7ns vs 105ns trfc be the source of the intermitent memory errors? Can someone help me troubleshoot this stuff? I have a feeling that it isn't just a dud DIMM because the ram is capable of being stable under load, even with win xp only recognizing 3.25 gb and the new stuff being undervoltaged. It just doesn't last. I attached some CPUID screens. I hope you can see them. Oh, I didn't include GPU but I am running a XFX Radeon 4850 512mb




375

post #2 of 5
Thread Starter 
OK I overformatted that post a bit and my screens are way too small. I tried to upload the original BMP again but it was just as small. How do I get them bigger?
post #3 of 5
Try running something looser, like 5-5-5-18 or even 6-6-6-20 @ 800 MHz and 2.2V. I think your 4-4-4-12 is stressing the memory controller. You will hardly notice the difference.

Also consider upping the voltage to your NB a notch. You're only at 1.8 GHz, should be able to get 2.2 or 2.4? +0.1 or +0.2V there to stabilize the mem.
 
13" beast
(6 items)
 
HTPC
(8 items)
 
CPUMotherboardGraphicsRAM
Lapped X4 940BE @ 3.5Ghz / NB @ 2.4GHz Biostar TA790GX 128M 1GB HD6950 @ 965/1360 (1.26V) Kingston Hyper-X 2GB x 4, 1066 MHz 5-5-5-15-22 ... 
Hard DriveOptical DriveCoolingOS
Crucial M4 64GB SSD, WD 640GB x 2 none Scythe Katana 3 Win7 64bit 
MonitorKeyboardPowerCase
Dell 24", Samsung 17" x 2 Logitech Illuminated Antec Earthwatts 750W Antec 300 
CPUGraphicsRAMHard Drive
Intel Core i5 2410M @ 2.3 Ghz stock Intel HD Graphics 3000 4GB 500 GB 
OSMonitor
Windows 7 WXGA (1366x768) 
CPUMotherboardGraphicsRAM
Athlon II X2 245 GA-MA785GM-UD2H Onboard HD4200 Kingston 2GB 
Hard DriveOSKeyboardCase
WD 2TB x 2 Windows 7 Ultimate Benq Wireless Antec NSK1480B 
  hide details  
Reply
 
13" beast
(6 items)
 
HTPC
(8 items)
 
CPUMotherboardGraphicsRAM
Lapped X4 940BE @ 3.5Ghz / NB @ 2.4GHz Biostar TA790GX 128M 1GB HD6950 @ 965/1360 (1.26V) Kingston Hyper-X 2GB x 4, 1066 MHz 5-5-5-15-22 ... 
Hard DriveOptical DriveCoolingOS
Crucial M4 64GB SSD, WD 640GB x 2 none Scythe Katana 3 Win7 64bit 
MonitorKeyboardPowerCase
Dell 24", Samsung 17" x 2 Logitech Illuminated Antec Earthwatts 750W Antec 300 
CPUGraphicsRAMHard Drive
Intel Core i5 2410M @ 2.3 Ghz stock Intel HD Graphics 3000 4GB 500 GB 
OSMonitor
Windows 7 WXGA (1366x768) 
CPUMotherboardGraphicsRAM
Athlon II X2 245 GA-MA785GM-UD2H Onboard HD4200 Kingston 2GB 
Hard DriveOSKeyboardCase
WD 2TB x 2 Windows 7 Ultimate Benq Wireless Antec NSK1480B 
  hide details  
Reply
post #4 of 5
Thread Starter 
OK. Thanks for the reply Oh, now one of the new sticks that was a bit temperamental at 1.8v when I first stuck them in has apparently completely failed. Memtest runs fine through the first stick and then nothing but errors on the second stick.

I took out that stick and ran the machine at 4-4-4-12, 2.1v (.1v over for old stuff and .1v under for new stuff), I equalized all the TRFC values at 105ns and 800mhz and it ran games fine for a couple hours, etc. I stuck memtest on it last night and this morning it had got to test 10 I think (random number was the "title") and it had nothing but errors again, plus the screen was corrupted with weird characters and stuff. Very weird. This morning I ran memtest for half an hour or so, long enough for it to say "all passed, hit esc to exit". No errors.

I will try again tonight with 3GB and loose, loose timings, 2.2v, I might adjust the NB as well but I don't want to throw too many variables in there. I understand changing timings isn't that noticeable, I just thought this stuff would run very easily at those timings at 800mhz when they are rated at 1066 and my old RAM has never given me a problem (could be that I never set the BIOS to tight timings, I can't remember, I will check if it recognizes the faster timings with just my old stuff).That stick that went bad never booted when it was the sole DIMM, but memtest checked it out fine at one point. I don't understand. Did I ruin that stick?
post #5 of 5
Thread Starter 
I don't understand. Errors: I get nonstop errors in memtest at 6-6-6-18 (my MC/MB won't go to 6-6-6-20) on the new stick. Old stuff holds up fine. I tighten timing to 5-5-5-18 and now I get intermittent errors, some for the same address others not. I get 50+ errors from two passes in memtest on the new stick. Is this memory bunk? Or I'm I just not hitting the sweet spot for them yet? I don't understand why everything tests clean one day and then the next, with slower timing (i also upped NB), I get crazy errors. I will try backing the NB down a tad to 2.0ghz or so. Right now it is at 2.4 ghz or 6x multiplier, +.2v
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Memory
Overclock.net › Forums › Components › Memory › Interesting problem with Ballistix DDR2 1066