BakaNeko Posted August 12, 2009 Share Posted August 12, 2009 I have been using two sets of the TWINX4096-8500C5D RAM (8GB total) in my Gigabyte GA-EP45-UD3P motherboard for about 8 months with no problems. In the past month, I started receiving assorted blue screens with stop errors, etc. They never seem to be the same. If I run Memtest right after one of the stop errors, I get a number of RAM errors. If I turn off the system, wiggle the RAM, let it cool down (or any of a number of other "remedies") it will usually come back up just fine. I can then run memtest again and it reports no problems. It will generally fail again within a day or two, however. It will fail no matter what OS is being used, including a boot from a Bart PE disk, so I know it's not a software issue. I installed it with SPD with all memory settings at default or auto. I did NOT remember to up the standard 1.8v to the 2.1v shown for the memory. I know overvoltage can fry the sticks, but can prolonged use at undervoltage do any damage? Since the problems started, I upped the voltage to 2.1, but it seems to make no difference at this point. Coincidently, once I purchased this combo for my home system and liked it so much, I purchased 10 similar systems for my office, and recently three of those systems have started having similar problems. I'm going around and resetting the voltage on the ones that are still working, but even after resetting it on the three dying ones, I still have problems. Any suggestions? Is the memory fried at this point? Link to comment Share on other sites More sharing options...
Corsair Employee RAM GUY Posted August 12, 2009 Corsair Employee Share Posted August 12, 2009 That might suggest a thermal or voltage problem, does the MB you have use a fan on the North Bridge? If not you might want to add our memory fan and see if that helps part number CMXAF1. Link to comment Share on other sites More sharing options...
BakaNeko Posted August 13, 2009 Author Share Posted August 13, 2009 Forgot to mention - I am using the memory fan that came with my Dominator sticks. Link to comment Share on other sites More sharing options...
Corsair Employee RAM GUY Posted August 13, 2009 Corsair Employee Share Posted August 13, 2009 Do you have access to another system you can use to test the memory? Link to comment Share on other sites More sharing options...
BakaNeko Posted August 13, 2009 Author Share Posted August 13, 2009 I have tried the RAM in my server as well (same model motherboard and same Corsair RAM - as I said, I initially liked the configuration so I used it in my home systems as well as at work). In all combinations, it eventually fails (PC RAM in server, server RAM in PC, server RAM in server, PC RAM in PC). Am I missing something fundamental? As I said, I have them set to auto for everything except the voltage, which is now set to 2.1v. Should I NOT trust the SPD values and set them all manually? By the way, my original two sticks were revision 1.1 - everything since then is revision 2.1. I've made sure to pair them equally, even putting four version 2.1 in at a time, but it makes no difference. Bill Link to comment Share on other sites More sharing options...
Corsair Employee RAM GUY Posted August 13, 2009 Corsair Employee Share Posted August 13, 2009 What are the exact settings you have set for both CPU and memory, just using Auto settings with 4 modules will almost never be stable? With 4 modules I would suggest setting the memory frequency at DDR667 or DDR800 and set the memory Voltage to 2.1 Volts and set the NB/MCH/SPP Voltage to +.2 Volts as well and test the system with http://www.memtest.org. In addition, with some MB's (Mostly ASUS) you have to disable legacy USB in the bios when running any memory test. Link to comment Share on other sites More sharing options...
BakaNeko Posted August 13, 2009 Author Share Posted August 13, 2009 I just saw this advice in another of your posts. Can't say what I have at home, but the systems here at work are set to (all greek to me, by the way): CPU Clock Ratio 8x cpu vcore 1.2125 cpu termination 1.2 cpu pll 1.5 cpu reference .76 mch core 1.1 mch reference .76 mch/dram ref .9 ich i/o 1.5 ich core 1.1 dram voltage 2.1 dram termination .9 channel a reference .9 channel b reference .9 auto by spd selects: cas lat 5 trcd 5 trp 5 tras 15 trrd 4 twtr 4 twr 8 trfc 68 trtp 4 cmd rate 0 memory freq 1066 Basically I appear to be having the exact same problems as listed in http://forum.corsair.com/forums/showthread.php?t=80448. The memory fails, I boot memtest without powering down and it's bad, then I power down and run memtest and it's good. I'll try the advice you posted and give it a shot. When you say "set the NB/MCH/SPP Voltage to +.2 Volts" you mean increase whatever they are currently set to by .2 volts, right? Thanks, Bill Link to comment Share on other sites More sharing options...
Corsair Employee RAM GUY Posted August 13, 2009 Corsair Employee Share Posted August 13, 2009 The MCH Voltage should be set to +.2 Volts and with 4 modules DDR800 would be suggested. Link to comment Share on other sites More sharing options...
BakaNeko Posted August 14, 2009 Author Share Posted August 14, 2009 Pardon my ignorance, but once again I assume you mean ADD .2v to the current value? Also which MCH voltage are you referring to, the core (currently at 1.1)? Also, when setting the multiplier I get various similar options on this board. There's a 2.00A (gives DDR667), 2.50A (DDR800), etc. but there are also similar values with a different suffix - 2.00B, 2.00D, etc. Do you know what the difference is between the A, B, D, and others is? I've set it to 2.50A, but don't know if that's the right one. Thanks. Link to comment Share on other sites More sharing options...
Corsair Employee RAM GUY Posted August 14, 2009 Corsair Employee Share Posted August 14, 2009 Pardon my ignorance, but once again I assume you mean ADD .2v to the current value? Also which MCH voltage are you referring to, the core (currently at 1.1)? A: Yes the MCH Voltage never mess with the reference Voltage that may over volt the system. Just use 2.50A (DDR800) Link to comment Share on other sites More sharing options...
Plexor Posted August 15, 2009 Share Posted August 15, 2009 I have been using two sets of the TWINX4096-8500C5D RAM (8GB total) in my Gigabyte GA-EP45-UD3P motherboard for about 8 months with no problems. In the past month, I started receiving assorted blue screens with stop errors, etc. They never seem to be the same. If I run Memtest right after one of the stop errors, I get a number of RAM errors. If I turn off the system, wiggle the RAM, let it cool down (or any of a number of other "remedies") it will usually come back up just fine. I can then run memtest again and it reports no problems. It will generally fail again within a day or two, however. It will fail no matter what OS is being used, including a boot from a Bart PE disk, so I know it's not a software issue. I installed it with SPD with all memory settings at default or auto. I did NOT remember to up the standard 1.8v to the 2.1v shown for the memory. I know overvoltage can fry the sticks, but can prolonged use at undervoltage do any damage? Since the problems started, I upped the voltage to 2.1, but it seems to make no difference at this point. Coincidently, once I purchased this combo for my home system and liked it so much, I purchased 10 similar systems for my office, and recently three of those systems have started having similar problems. I'm going around and resetting the voltage on the ones that are still working, but even after resetting it on the three dying ones, I still have problems. Any suggestions? Is the memory fried at this point? I have the exact same problem. I have two sets of the same RAM. I have tried using one set and even one stick. The problem shows up on both my GA-EP45-UD3P 1.1 and GA-EP45-UD3R 1.1. Seems the problem shows up less often when running one stick in single channel mode. The problem shows up regardless of what bios settings I use. Increasing MCH voltage or reducing ram frequency to 667/800 sometimes helps temporarily. As the problem shows up on both of my boards, I suspect it is faulty ram. Link to comment Share on other sites More sharing options...
Corsair Employee RAM GUY Posted August 19, 2009 Corsair Employee Share Posted August 19, 2009 Let's get them replaced, please use the On Line RMA Request Form and we will be happy to replace them. Link to comment Share on other sites More sharing options...
Han Posted August 19, 2009 Share Posted August 19, 2009 Just for the record I'm yet another frustrated user with exactly the same issues. Vista x64 ran fine on Gigabyte GA-EP45-DS3R with two TWINX4096-8500C5D sets (8 GB total) for more than half a year... Link to comment Share on other sites More sharing options...
Corsair Employee RAM GUY Posted August 19, 2009 Corsair Employee Share Posted August 19, 2009 Have you tested the modules one at a time to be sure one is not failing and were you running them at DDR667 or DDR800? Link to comment Share on other sites More sharing options...
Han Posted August 20, 2009 Share Posted August 20, 2009 I was running them at full speed for over 6 months without issue. Memtest did report a few errors at 8 GB setup, but Vista and XP seemed to work fine, even under load. When I tested modules in pairs (4 GB setup), Memtest went thru error free at full speed. When random restarts started to occur three days ago, I lowered the frequency to 800 MHz, set latencies at 5-5-5-18, and left the voltage at 2.1V. Memtest almost got it thru - there was only one error in test 7. But when I repeated the test, the error was gone. However, Vista keeps restarting, without blue screens - like I'd press the reset button. A quick peek at the Event Viewer shows many system errors, though. The one most interesting is "The hardware has reported an uncorrectable memory error." Today I'll test the modules in pairs (2×2 GB) and see how they do.... Link to comment Share on other sites More sharing options...
Han Posted August 20, 2009 Share Posted August 20, 2009 As expected, two modules pass the Memtest+ without a single error, at full speed (1066 MHz @ 5-5-5-15 @ 2.1V). Is there anything we can do in order to run four 2 GB modules without issues? Who's to balme? Intel P45 chipset? Mobo manufacturers? Something's fishy here! :evil: Link to comment Share on other sites More sharing options...
Corsair Employee RAM GUY Posted August 20, 2009 Corsair Employee Share Posted August 20, 2009 Did you increase the NB MCH Voltage as well when you had all 4 modules installed? Link to comment Share on other sites More sharing options...
Han Posted August 20, 2009 Share Posted August 20, 2009 I did, but it didn't help. Link to comment Share on other sites More sharing options...
Corsair Employee RAM GUY Posted August 20, 2009 Corsair Employee Share Posted August 20, 2009 What were the exact settings you had set for both CPU and memory, and on what test are you getting errors? Link to comment Share on other sites More sharing options...
Han Posted August 20, 2009 Share Posted August 20, 2009 To be honest I was pretty much fed up with random OS restarts so I didn't test the memory at defferent settings that extensively. After 10 hours of operation at 800 MHz (5-5-5-18, 2.1V) Memtest showed errors in test 4, so I aborted and increased MCH by 0.1V. That made it even worse (errors started to show in test 2), because the memory was actually sped up by about 200 MB/s (just like I'd turn Turbo mode on). Looks like BIOS screws up with memory speed on its own... In general I got errors in tests 2, 3, 4, 6, 7 at random MBs. Sometimes at 7000+ MB, sometimes at 150 MB. Optimal defaults (Turbo mode) work best for 2 modules. No errors whatsoever. Note that you must enable Standard mode when you use 4 modules, otherwise the memory bandwidth is set too high. I left all the CPU settings at default (10x 333 MHz). Link to comment Share on other sites More sharing options...
BakaNeko Posted August 21, 2009 Author Share Posted August 21, 2009 After doing some more testing, I'm in the same boat as Ham. Two modules seem to work pretty well, but four don't. My server is still experieincing random lockups and event viewer continues to show memory errors. I'm going to be pulling the Corsair RAM from my server and RMAing it to see if that will make a difference. Once I check that, I may be RMAing my RAM from my working system as well. If all goes well after that, I might be RMAing the RAM on the 3-5 systems I have at work that also fail... Link to comment Share on other sites More sharing options...
Corsair Employee RAM GUY Posted August 21, 2009 Corsair Employee Share Posted August 21, 2009 This has been posted previously with running two sets but some MB's will just not run 4 up and be stable, can you test them in another system? With 4 modules I would suggest setting the memory frequency at DDR667 and set the memory Voltage to 2.1 Volts and set the NB/MCH/SPP Voltage to +.2 Volts as well and test the system with http://www.memtest.org. In addition, with some MB's (Mostly ASUS) you have to disable legacy USB in the bios when running any memory test. Link to comment Share on other sites More sharing options...
Han Posted August 21, 2009 Share Posted August 21, 2009 That might work, but I didn't buy 1066 MHz Dominator RAM to run it at half the rated speed... Could you please explain what's going on here? Why this RAM fails to operate at 4× 2 GB configuration (verified @800 and @1066 MHz)? Could you please contact Gigabyte and let them know about this issue? I'm sure they'll listen to you so that maybe we could expect a fixed BIOS... By the way, I have 4x 1 GB Corsair CM2X1024-6400C4 in my other computer (Gigabyte GA-P35-DS3L). This memory runs rock stable at 800 MHz (4-4-4-12 @ 2.0V)! Link to comment Share on other sites More sharing options...
Corsair Employee RAM GUY Posted August 22, 2009 Corsair Employee Share Posted August 22, 2009 Please search for that I have explained it many times, but in a nut shell there is more loading as you add more memory and its common to have to slow the memory down as you add more memory. Not to mention we do not test Twin sets in 4 up configurations and why the next lower speed grade is suggested. If you want 4 modules I would suggest using Q28G6400C4DHX as that is the only 8 Gig DDR2 kit we offer at the moment. Link to comment Share on other sites More sharing options...
Han Posted September 22, 2009 Share Posted September 22, 2009 Update: two days ago (after a month since I went back to 4 GB total) one pair of Dominator failed. Vista and applications began to crash, Memtest86+ shows lots of errors. Today I'll put in operation my second pair, but I'm seriously considering to RMA both of them and buy a different high speed brand that offers 4×2 GB or 2×4 GB at 1066 MHz. P.S. I'm fully satisfied with 800 MHz XMS memory (4×1 GB @ 4-4-4-12 @ 2.0V), but the 1066 MHz Dominator is a huge disappointment! Link to comment Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.