OCZ PC3200 Gold new BH5 die + DFI NF4 SLI-D Part2

Hmm I burned mine in for 24 hours, 3.2V @ 2-2-2-5, 200Mhz and so far I can get 240Mhz at 3.2V. I switched to 5V rail just for Memtest and could only get to 250Mhz. Had about 64 errors on 1st pass of #5 at 260Mhz, 3.5V. Upped voltage to 3.6V and have 5000+ errors! :confused: BH-5 don't like voltage, or maybe it's getting too hot? I have 92mm fan sucking air from RAM (case ceiling fan blows air towards RAM and mosfet)

2-5-2-7-16-2-2-1-2-3072-1-E-A-0-3-1-7-5-A-D-16-7-D

ninja_xxx: Isn't stock, 2.6V, 2-2-2-5?

Have you tried using different Drive Srength/Data Drive Strength values? Try:

trc = 9
tref = 3120, 2560 (not sure if 2560 is the correct one, something close to that)
Drive Strength = 5 or 8
Data Drive Strength = 2

Stock is 2.8V, 2-2-2-5, 200 MHz.

hey hovo73 thanks alot man.... i've tried all that but i got it to work with your first crazy :) time'ns at 230 at 3V's :)...seems it doesn't like 3.1 or 3.2 volts for some reason...anyways thanks alot man i'm happy now :banana: :banana:

i tried your second time'ns at 245 at 3.2v's didn't work but its all good :toast:

ps. sorry guys didn't mean to say stock should have said auto detect :slapass:

anyways what divider do i use to keep my ram at 230 and up my cpu to 2.5-2.7 :)

Nice to hear that.

You have CPU multi 10x right? This will work:

HTT=253, x10, CPU=2530, RAM=230, Divider=183
HTT=270, x10, CPU=2700, RAM=225, Divider=166
HTT=280, x09, CPU=2520, RAM=229, Divider=166

You have to play around to get a good match. Use the program in this Thread for easy calculations:

http://forums.extremeoverclocking.co...d.php?t=149717

Are typical BH-5/UTT 260+Mhz and 3.4+V achieved with the heatspreaders removed?

I'm willing to take them off for more Mhz.. ;)

Here are some screanies of my results:

OCZ EL Gold BH-5
DFI NF4 Ultra-D
Winchester 3000+
Using 510-2 BIOS, with 5.0V line with yellow DIMMs

245 MHz with 3.2V,
MyGenie Settings:
2-2-2-6-1T-7-14-2-3-1-2-3120-A-E-A-0-L5-L2-7.0-F-5.0-016-D-16x-7x-D

I am going to run another session of ~12 hour burn-in with this speed and push further.

http://www.xtremesystems.org/forums/...chmentid=32135

http://www.xtremesystems.org/forums/...chmentid=32136

hey hovo what psu do you have...thats very nice over clock for 3.2v's...is that the 3.3 rail.....

i think i found my problem...when i set my ram to 3.2v's in bios it says 3.17v's in bios but right now i'm at 3v's at 230(thanks to you :) ) and check'n smartguardian to see what the atx 3.3v is at and it says 3.24 :confused:

why is my ram gett'n undervolted is it my mb or psu or bios(5/10-2)

what does you ram volt in bios say?

nice going so far hovo. it takes me 3.5 to get 250 on the gold modules. nothing i do can get that down. but then again, everything in my board needs more juice, cpu & ram wise. oh well, what can u do..

ninja_xxx, mine is Enermax Noisetaker 485W and I am using 5V rail. I also get about 3.17V if I set it to 3.2V if I am using the 3.3V rail, but it goes to about 3.23V if I am using 5V rail. Basically the 3.2V is very unstable if you are using 3.3V, I basically almost get no increase in voltage from going to 3.1V to 3.2V.

Thanks pershoot, my target is 252MHz with as little voltage as possible, So I can run with 183 divider with my CPU at 9x280. I am also going to try some TREF settings that were discussed in another thread in this forum:
http://www.xtremesystems.org/forums/...ad.php?t=65151

Thats pretty sweet but how long have you been runnin ur ram like that? Hopefully it will not burn out like a lot of the UTT lately.

Quote:

---

Originally Posted by uscfan

Thats pretty sweet but how long have you been runnin ur ram like that? Hopefully it will not burn out like a lot of the UTT lately.

---

Less than 5 days :D I am no stranger to RAM death, both of my TwinMOS got canned on this board. I had been off of 5V line for a while, I just went back to it a week ago. I am crossing my fingers ... :D

hey hovo from what i read on dfi-street and here its not good to use less then 3.3v's on the 5v rail....anyways i'm not use'n the 5v rail till dfi says it ok to use and wont kill ram and mb.

ps. has anyone tried the jumper trick without an adjustable psu and see if we can get full 3.24 volt on the 3.3 rail. maybe you should try this hovo i'm too lazy :)

Quote:

---

Originally Posted by ninja_xxx

hey hovo from what i read on dfi-street and here its not good to use less then 3.3v's on the 5v rail....anyways i'm not use'n the 5v rail till dfi says it ok to use and wont kill ram and mb.

---

Yeah, but [3.2v on 3.3V rail] != [3.2V on 5.0V rail]. They are really two very different beasts. When my TwinMOS CH-5/UTT sticks were still alive, they could only go up to only 215-220 with 3.1V, changing to 3.2V made absolutely no difference, but when I changed to 5V rail, the same 3.2V took it all the way up to 255 MHz (memtest and p95 stable!).

Yeah I am aprehensive about using 5V also.

Quote:

---

ps. has anyone tried the jumper trick without an adjustable psu and see if we can get full 3.24 volt on the 3.3 rail. maybe you should try this hovo i'm too lazy :)

---

No haven't tried it, I really don't see how it can help without modding the 3.3V line, given the fact that going from 3.1 to 3.2 VDIMM doesn't make any difference.

this is the best i could do in win, so far w./ 3.45. ram is ~2 wks old. memtest #5 & #8 to 250 (30-40min. test each). in win, memtest win32 errors on 248, and 249-250 i get resets. guess ill let it sit for a few weeks and revisit. hopefully they will loosen up.

http://indabasement.com/oczgold.JPG

have you not tried trfc at 14 or 16 for more stability?

anything but 12 causes instability for me, from initial observation.

Quote:

---

Originally Posted by pershoot

anything but 12 causes instability for me, from initial observation.

---

i haven't tried 12 yet, i didn't think they would go that low.

its BH die; np w./ tRFC 12:)

Quote:

---

Originally Posted by pershoot

its BH die. np w./ tRFC 12:)

---

nice tip pershoot :toast:

Edit 12 is the magic number :banana:
i've never been able to pass test 7 in memtest without getting at least 1 or 2 errors, now with trfc at 12 i made 2 passes error free :woot:
after 3 weeks of banging my head against a brick wall im finally getting somewere.
Thanks again pershoot :toast:

Quote:

---

Originally Posted by ninja_xxx

ps. has anyone tried the jumper trick without an adjustable psu and see if we can get full 3.24 volt on the 3.3 rail. maybe you should try this hovo i'm too lazy :)

---

I tried the jumper trick, got the same results as with using the 5v rail.

wow this thread still going strong :)

I updated my guide at http://i4memory.com/showthread.php?t=327 in async section re- venice/sandy and using dividers for ram and some settings to try to stabilise memory clocks :D

yep im gonna use that since my OCZ Value series is BH-5

though im on a DFI UT NF3 w/ a 754 NC

Quote:

---

Originally Posted by perkam

Though, are you sure 3200 VX doesnt do 260+ on avg ????...I mean 4000VX isnt using any different chips...

Perkam

---

arent they?well i ahve an example right here that they arent, at least in this case.tried 2x512pc3200 and 2x512pc400 all togethher in this dfi=no boot. then i go to cpu-z and guess what the mems show up as having different SPDs.I then get a mates 2xpc4000 ans guess what? 4xpc4000 all work and have same SPDs but different to the PC3200s.maybe an OCZ guy would like to explain this?

P.S. also tried 1xpc3200 with 1xpc4000 and no boot, different ICs and thats with BIOS reset in case you think of timings or CPC etc.

@pershoot have you tried Idle Cycle Limit lower?With these 4xpc4000VX I get faster SuperPis the lower I got and just as stable at 0 than 16.

Quote:

---

Originally Posted by pumbertot

@pershoot have you tried Idle Cycle Limit lower?With these 4xpc4000VX I get faster SuperPis the lower I got and just as stable at 0 than 16.

---

yea. only 16 and 128 work well.