Yup, I run SCSI Raid, and really do not want to corrupt my setup/shrug.
Larry
Printable View
Yup, I run SCSI Raid, and really do not want to corrupt my setup/shrug.
Larry
DDR2 1333 always been the bug for my "SPECIAL" board.
now at 3.8 gig but experiencing stability issue, failed orthos after less than 1 hour :(
Spec :
CPU : q9450 cooled by 9700nt - vid 1.25v
ram - patriot viper 1066 @ 1141 , 5-5-5-15 timing - 2.3v
NB - 1.4v
- rest of the voltages are set at their minimum spec
http://img.photobucket.com/albums/v6...n1973OC2-1.jpg
I think it is my memory controller dying - if I up the PL to 10 (was stable at 8 before) and give it 1.28v vNB I can pass 5 runs of linpack. I also have to run a little extra vdimm but the ram is stable at the old lower vdimm settings on my mates board. Just needs more voltage on my board for the same speed/timings.Quote:
Originally Posted by BeastNotro
I think my board may have degraded, or the bug that made me have the memory freq hole between 1200 - 1277 has now started to cause other issues.
I was at 1.26 vNB, 2.14v vdimm, 1.12v vFSB, 1.38v vcore. It was fine at these voltages, then when I retested it I'm getting errors all over the place. I used to pass 30 runs of linpack, now I pass 1 out of 5 runs. It's completely unstable, and I didn't change any settings at all.Quote:
Originally Posted by Mumid
I also do not run any pull-ins enabled (they never worked anyway they made no difference to the bandwidth or latency at all).
No problem mate, guess he was just typing it wrong.Quote:
Originally Posted by Ket
I´ve tried that and the result was the same, error in core#2 after a few minutes in the blend test :mad:Quote:
Originally Posted by Pntgrd
I don´t know what is making this happen, but i´m counting on you guys to help me out :D
Regards.
@ Blasar - if you are really at 1.24v vcore, you most likely need to raise it to get it stable at 3.8GHz. Try raising it so it says 1.28v in cpuz under load.
I need 1.28 vFSB to be stable with setup in sig. It was 1.22 but have needed to increase it after updating the bios and upgrading to 4 dimms. Have you tried higher volts here?Quote:
Originally Posted by CryptiK
No I haven't, as the max rated vFSB is 1.15v from intel's website, and I don't want to go over it. I really believe it is a memory controller problem, as my ram and cpu both test fine in another board at the same speeds & voltages that are unstable on the P5Q-Deluxe.
My board undervolts fsb by -0.08 last time i checked, could even be more now with new bios.Quote:
Originally Posted by CryptiK
Its not gonna do it any harm to test it
I have never measured my vFSB as my cooler is in the way. Can anyone else confirm the undervolting of vFSB from bios set value?
Well I have discovered something. I was completely unstable at PL 8, and was even more unstable at PL 7. I could get stability at PL 9 and PL 10, but obviously lost some performance.
I tried upping the vNB and retrying tests but it either made :banana::banana::banana::banana: all difference. I re-flashed 1201 just then, and the results were the same as before. I passed 1/5 tests in linpack at PL 8, 1.26v vNB. I could get it to pass 5/5 tests at PL 9 @ 1.26v vNB but obviously lost some performance.
Then I re-tried PL 7, at 1.26v vNB and passed 3/5 tests......weird, as I only passed 1/5 at PL 8 at the same voltage. I bumped up the vNB to 1.28v, and I just passed 5/5 tests at PL 7. I'm going to test further but this is strange behaviour.
**EDIT - passed 7/10 tests, then went to reboot and it wouldn't even post. Damn this board.
Seriously mate try upping vFsb and re-test, i think it might be able to get your stable again.
Guys I did a test today.... regarding the Different voltage when u flash to a different BIOS, at first I thought it was BIOS related, so I started from scratch, cleared CMOS flashed back to 0803 and measured with a VDIMM, so I did for 1201 and 1306. So My findings were through all 3 BIOSES the same oO, for example: 1.96vDIMM BIOS / IDLe: 1.95/ load: 1.96. Weird huh I thought, because when I bought the board and it came with BIOS 0803 and before I flashed to 1201 I check it out with a vdimm, and when i set VDIMM bios : 1.96 I get @ idle 2.04 load:2.05, so was @ BIOS 1201. So thought sumtin isnt right. SO I think just find the reason we got different voltage readings. My Corsair HX dont have the colored molex: red/black/black/yellow. They are all sleeved black. All the time I was testin the vDIMM. The black next to red would give me: 2.04(1.96v Bios set) Vdimm readout and when I stick the prob in the black near the yellow, I would get 1.95v ( 1.96 BIOS set). So obviously must be wrong, then I tried another molex, a extention that was from a fan. There both black would give me 2.04:S......... So which 1 is the proper ground...... Can some1 please duplicate my testing and confirm this.
*edit I tried another corsair molex connector both black would give me 2.04 readout, so probbly I had 1 dodgy cable. Also tested with my external HX520W to ground it , also both black would give me 2.04 readout....
Just got the motherboard yesterday. I really like the chipset cooling but I did find that the start vid was lower than the DFI's.
Quad 6600 G0:
DFI P35 = 1.28v
Pq5 Deluxe = 1.19v
This is with all throttling turned off (C1 etc..). After work today I'm going to install Vista x64 and hopefully get some slight overclocking done. I'll post some results either end of today or tommorrow. :D
Question, when you guys are measuring say vdimm, what ground are you using? In other words, the mobo tray/case, ground in a molex...?
Larry
I use the black ground right beside the yellow 12v wire in any 4 pin molex.
Ground is ground and even the one beside the RED 5v Wire is still ground and same readings for me.
Thanks :) I just took a reading off 1 side and gota scrambled # weird, as my meter reads the molex voltage fine.Quote:
Originally Posted by BeastNotro
Larry
http://www.xtremesystems.org/forums/...&postcount=472
I took the reading off the right side.
So volt meter is set to 20 DCV?
And you put black probe in Ground wire . which one beside red wire or yellow wire as both are ground?
And then you put red probe tip onto the Soldered Part of the Vdimm Chip .. either or chip no matter as both should read the same.
Make sure your touching the proper ones.
EDIT:
I just tested mine again and i get 1.84v on the left chip and 1.88v on the right chip . weird as both should be same or very close like it is for me.. i never seen them different for me till now . maybe its Kets Bios.
I am at STOCK settings as i flashed Bios to 1201 modded one Ket made to test it out.
I'm on the correct 1, I tried ground to both grounds in the molex, individually of course. Maybe I wasn't getting a good contact or will try another molex later. Not knowing my ram voltage is just crazy. I'll try it again later as I need to run now. Thank you for the reply :)
Larry
I just started getting some instabilities in an otherwise rock solid OC. It started after installing 1306, so I reverted back to 1201, then 1103, and finally back to 803. The problem persists in all these BIOS versions.
Basically, the problem is that I hang when XP boots. It goes from the blue welcome screen, pauses for 10-15 secs, loads up my wallpaper, then that's it. I have full mouse / keyboard control, but can't ctrl-alt-del. BTW, I did manage to get into XP once and threw every stress test I could at it, and I passed with flying colors.
To fix the problem, I started playing with voltages, and lo and behold, by upping my FSB term voltage from 1.36 to 1.40, everything's fine again.
But here's what I don't get....why all of sudden do I need more FSB term voltage to boot? Did installing 1306 screw something up, even though when I reverted to previous versions that problem still persists? Is my XP install corrupt?
Trust me i can't do 3.6ghz stable ever since putting 1306 bios and now that you posted i thought i was nuts and mayeb CPU degraded or mobo is screwed... it still can be CPU degradtion...... but since you posted your issue.
I also went to 1201 Official . then to 0803 and same frigging thing .. can't pass linpack 1.4 at 3.6 no more.
So right now i am tryinh Kets modded 1201 bios to see how it is......
We as in you and a selected few .. may have to shut PC down... Unplug Power Supply Wire (after turning PSU off) and then clear CMOS with jumper and pull battery otu at same time and when done that with battery out and jumper crossed. push in the PC's power button a few times holding it about 20 secs each time ..... then let it sit for 10 min or 20 min and the uncross jumper . install battery and then plug PSU in and turn it on .... boot PC into bios load defaults .... then boot again and then set it and see how it goes.
Maybe there is a Quirk in Bios 1201 and 1306 as i could pass linpack 1.4 1201 then i could not so i upgraded to 1306 and same thing .. so for me it can be BIOS. CPU degration and or Mobo issue.
I am still tetsing it.
Anyhoo i am waiting for my E8600 to arrive and i hope all is fine with my MOBO and other parts.
I guess I need a different volt meter/I can read the molex voltages fine, 5.17 on the 5v, 12.18 on the 12v but when I take a reading on 1 side of the vdimm I get 3.16 with volts jumping and the other side nothing, go figure. Even went into bios and lowered my ram volts to 1.95 from 2.02 and get the same reading after a reboot. sheesh. My fluke probes are too darn big to be probing around in the box with. Just my luck.
I figure if I were actually getting 3.1 on the ram they woulda been toast a few days ago :D
Larry
Get an analog meter and see if the needle fluctuates.
I forgot I also have an Ideal, buried in the truck somewhere, will find it tmm and see what gives, probes are smaller and I can tape the ends... like you, getting tired of throwing more $$ in the pit :DQuote:
Originally Posted by gillbot
Larry
OK, found it in the truck, calibrated it and took my measurement, Bios 2.0v vdimm= 2.1v Idle. 0.1v above bios setpoint in 803.
Larry
Nevermind, my problem is just a bad stick of ram.