Jump to content
Corsair Community

Overclocking TwinX 3200LL ver1.2


Recommended Posts

I read in the forums that my ram (TwinX Corsair XMS 3200LL ver1.2) is CH-5 and doe not like Volts. So I tested it. 2.6v good 2.7v good 2.8v Prime fails after 51minutes 2.9v Prime fails after 9 minutes 3.0v Prime fails 3.1 Prime fails after 9 minutes 3.2v Prime is liking it 2hrs and counting. 2.6v good to 210 in Memtest86 3.2v good to at lease 230FSB Memtest86 I found other users with similiar results. So whats the real story.. DFI NF II Ultra Infinity 1/21 2500XP-M (IQYHA 0351 MPMW) 11x219FSB 1.85V core 3.2Vdim 1.5AGP Still testing Prime95 Corsair XMS 3200LL @11-3-3--2
Link to comment
Share on other sites

  • Corsair Employees
Max supported voltage is 2.8 Volts, and since there is a VR in the memory IC's, with XMS3200XX Rev 1.2, it will make little difference to the memory what you set the voltage to, but that might suggest a issue or limitation with the MB it self. Can you test the modules on another MB?
Link to comment
Share on other sites

On my Abit NF7 r2, My XMS 3200LL v1.2 is currently working at 221 fsb stable on 2.7v Bios (2.8v Winbond monitor) at 11-3-2-2 timmings. strangely upping fsb higher causes the odd memtest error and loosening timmings changes nothing . 233 fsb is almost atainable but i dont want to push the vdimm right upto 3v. The default 2.6v (actual 2.7v) gets me to 215 without issues.
Link to comment
Share on other sites

[quote name='mad-matt']On my Abit NF7 r2, My XMS 3200LL v1.2 is currently working at 221 fsb stable on 2.7v Bios (2.8v Winbond monitor) at 11-3-2-2 timmings. strangely upping fsb higher causes the odd memtest error and loosening timmings changes nothing . 233 fsb is almost atainable but i dont want to push the vdimm right upto 3v. The default 2.6v (actual 2.7v) gets me to 215 without issues.[/QUOTE] what multiplier are you using? At the same memory settings and the same mobo, ram with xp mobile 2600 cpu (IQYHA 0348 SPMW)and a multiplier >9 gives errors. At 2.6v dimm, memtest errors on the first pass and freezes, 2.7 and 2.8 vdimm allow memtest to run but give many errors. A multi of 9 is error free but this gives a 'lame' cpu of 1980 mhz. i can also achieve error free memtest and prime95 with a multi of 11.5 and an fsb of 211 which gives a 2426 mhz cpu.
Link to comment
Share on other sites

[quote name='bwana']what multiplier are you using? At the same memory settings and the same mobo, ram with xp mobile 2600 cpu (IQYHA 0348 SPMW)and a multiplier >9 gives errors. At 2.6v dimm, memtest errors on the first pass and freezes, 2.7 and 2.8 vdimm allow memtest to run but give many errors. A multi of 9 is error free but this gives a 'lame' cpu of 1980 mhz. i can also achieve error free memtest and prime95 with a multi of 11.5 and an fsb of 211 which gives a 2426 mhz cpu.[/QUOTE] I dont have brilliant cpu cooling yet so cant push too far and my psu cant seemt to push the 1.775vcore i need to get my 2500 anywhere near stable at 2200. So at this moment in time im running at 2100 (9.5x221). Now ill describe an odity with this board for me . Running memtest86+ 1.11 it shows 1550MB/s with 11-3-2-2 and an Auto or 5/5 6/6 divider. I get the odd error, usualy 2 in test 5. It is error free at 11-3-3-2 which shows as 1521 MB/s. Now the oddity is that if i set a 3/3 or 4/4 divider i can be stable at 11-3-2-2, but the mem speed is now 1521, equal to the memspeed at 6/6 11-3-3-2. So for stabilit i am runnign the 3/3 divider which allows me the 11-3-2-2 latancy error free at 221 fsb with 2.7volt setting. I can only go higher fsb 222 when using the 8/8.5 multis and maybe 11, but i can test due to cpu cooling atm. some reason with the 9/9.5/10/10.5 multis, system just wont post. 8.5x233 posts, but i think thats expecting abit much from this memory, , i get a fair amount of errors in test 5 and 6although they are alot less with 2.8v setting(2.9v actual - breaking warrenty conditions). Changing Lantancy values seems to not effect it in any way, same amount of errors with 11-4-4-3 as 11-3-2-2.
Link to comment
Share on other sites

your modules in dual channel (1 gig), command rate 1T, cl2T, will yield: asus a7n8x rev. 2 - 2600+ tbred 333 - DDR ~446 (223) 3(tRP)-3(tRCD)-11(tRAS) - 2.6vdimm Abit NF7-S Rev. 2 - mobile 2500+ 266 - DDR ~444 (222) - 3-3-11 - 2.6vdimm some batches require 2.7v to get there. mine cant take any more then 2.6. there is no way you need 3.2v to get to 219fsb (as per your sig) with this ram. anything past 2.6 or 2.7, and the rest is given off in heat. the ram wont accept anymore voltage (OVP).
Link to comment
Share on other sites

Get rid of your memory using the the pathetic CH-5 modules. Corsair make great memory, but they cant polish dog poop. Get on ebay and get some 3500 or 3200ll v1.1 modules. I get 250 MHz with timings of 5,2,2,2. You just cannot buy memory like this from corsair or anyone else anymore. Bring back BH-5 Winbond! your new stuff may be cheaper to make but its about asw good as EDO ram! :mad:
Link to comment
Share on other sites

I am fairly new to the OC scene, but I too am using TwinX 3200LLPT, but ver1.1 memory modules (2x512). With an MSI K7N2 Delta ILSR board can not get above 220 FSB speeds (2-4-4-11), voltage tops out @ 2.7v. Using an AMD XP-M 2600+ mobile processor swiftech 463 heatsink. What do I need to do to achieve higher FSB? Thanks in advance! mld45 :confused:
Link to comment
Share on other sites

many mobile users are stuck within that fsb range. for some, the L12 mod (forcing the chipset to read the cpu as a different fsb model then what it really is) has worked. it is exclusive to an Abit NF7 though. it probably wont work in your board though... ive seen DFI users hit 240 range with a mobile, and one hit 235 on an NF7 (L12 mod), so it is possible. if your sticks are revision 1.1, then they are Winbond-BH5 based. dont be afraid to give them up to the warrantied voltage (and slightly higher for even better results (but you'll void your warranty). your timings can be 2(tRP)-2(tRCD)-11(tRAS) (tight) or moderate 2-3-11, or loose 3-3-11, cl2T, command rate 1T. they should accept more than 2.7 and should be able to do a RAS to CAS of 2 if they are revision 1.1. if you have 1.2, then youll probably not be able to go past 2.7, 2-3-11 (tight), 3-3-11 (moderate), 3-4-11 (loose). your 4-4 timings are hurting your performance i bet. if you can tighten them up a bit and settle for a slighty less fsb possibly, youll benefit more. you can get a modified bios for your board to give you more options (vdimm, etc.). search through [url]http://nforcershq.com/forum[/url] , under your boards heading.. and you should be able to find it. the modified bios will void your warranty as well on the board. good luck.
Link to comment
Share on other sites

  • 4 months later...
Have the same Twinx 3200LL sticks. Have an Abit AN7. I also discovered that strange thing with the voltage. 3.2 Volts gives it a chance to make it to 230MHz. Anything other and it will hardly make it to 220MHz. A little while ago i bought two sticks of Crucial PC3200 512 sticks. I made it all the way to 235MHz 2.5, 3, 3, 11 timings with these sticks. Fairly good if I may say so. And it didn't need a lot of voltage either. And best of all.... They are half the prize of Corsair 3200LL Twinx! I'm sure there are other Corsair memory that is better than these, and that the Rev 1.1 sticks where much better. Corsair which promote their RAM as overclocking RAM and at those prize levels, they should NOT sell memory like the 3200LL Rev 1.2!
Link to comment
Share on other sites

  • 2 weeks later...
I'm having the opposite sort of problem...I just got 2 sticks of 3200 512LLpro, complete with flashy lights and all that nonsense. Anyways, I whacked them in my NF-7 straight at 2.8 and set the fsb to 220. Prime failed within minutes. Went to 2.9, same story. Tried all the way down to 200 and prime failed every time. 200 was ok, couldn't get above it at all. Tried all the different flavours of bios out there...no joy. Still couldn't get above 200 without prime failing. Downloaded memtest86+, ok at stock, but anything above 200 gave errors. Nice, 2 sticks that don't overclock at all.!! I'd never heard of such a thing on decent memory, I couldn't believe it. This was all at 2.8 or 2.9 Vdimm. On a whim (and a prayer) I set the volts to default and whadya know....know the damn things are running at 217 ok. Go figure, they just didn't like the volts at all it seems. I've yet to prime more than an hour so it's not set yet, but I couldn't get more than 30 mins before. Damn computers!!! :confused:
Link to comment
Share on other sites

  • 3 weeks later...
Since overcloking the CH-5 seems to be today's fashion, I (carefully) gave a try, and the results are very positive. 2,9 and 3.0 V don't lead to any improvement, but at 3,1V I can run my system at 236 FSB (2-3-2-11). This is memtest and prime stable. I first tried with a fan blowing on the RAM, but it stays very cool without. 238 Mhz gives errors in memtest. 2-2-2-11 timings seem unstable whatever the speed over 220Mhz. I'd like to know if anyone fried his sticks at 3.3V before testing it, since I'm very pleased with my results so far. Skogkatt
Link to comment
Share on other sites

drop your tRAS to 6, and your ram will fly! in that DFI @ 2-2-2 (flying hard with 3.1-3.2 of course, hehe). corsair PCB is funny like that. 6 also seems to be more balanced then 11 is past a certain mark on this PCB. oskar wu's 12/18 seems to be good (what ive heard.. dont own a dfi board myself). i havent heard of anyone frieing their sticks (but im sure they exist).. so long as you keep them cool. i wouldnt go beyond 3.2 though. def. stick a fan or 2 up there. the voltage limiter normally gets beaten out at around 3.1 where the ram scales with voltage. i am using a piece of dfi's implementation of bpl 3.19 in the bios im in, and the ram doesnt seem to have that voltage limitation on it (it scales with voltage less then 3.1 (so that 3.1 rule doesnt apply to me). you shouldnt have a limit either with a nice modded bios with bpl 3.19. good luck:)
Link to comment
Share on other sites

Archived

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

×
×
  • Create New...