• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

First time tweaking voltages, advice requested.

L

Little_4nic8er

Gawd
Joined
Oct 24, 2009
Messages
603
Hello everyone. :)

I have played with my FSB in a very mild manner in the past with varying chips but never went as far as to get a 3rd party cooler, or up the voltages at all. I managed to get a great deal on a Q9550, (C1 revision unfortunately.) and replaced my E6850. I noticed the 9550 runs a lot warmer than my 6850 did, on both load and idle. The 6850 would idle at 24C, the 9550 40C. The 6850 would hit 40C, and the 9550 60C on load at default clocks. I used AC MX-2. I also reapplied it twice to make sure it wasn't a compound issue. I bought a Noctua NH-U12P SE2 (and a 5870) and received the cooler yesterday. (The GPU was back ordered before my payment went through. Tis a popular card. Sold out in less than 2 days.)

After spending a good part of the day completely dismantling my PC I now have it all up and running. (My cables are much neater than before to, though they were not terrible.) It dropped my temps quite nicely.

I have started to OC, and have the 9550 at 3.63Ghz. It was Prime95 stable at less voltage, but I like to push things so I ran a 5 pass Intelburn test. It failed shortly after the second test would start. So I pushed all voltages up. (Incrementally.) They are now as such. The vcore is set to 1.375000. FSB OverVoltage +0.3, DDR2/DDR3 OverVoltage +0.3, and MCH OverVotlage +0.2.

My temps hit 68C max with an Ambient of 34C (Core0, the rest were up to 5C cooler during fluctuations.) during a "Very High" Intelburn test scan, but I passed the test with no BSOD. (I've seen so many over the last 48 hours.)

It (Core0, always the hottest.) is currently idling at 44C with an ambient of 35C. (core1-3 vary between 37-40C.)

So what I am wondering is does anything here set off alarm bells for those of you who have done this for some time? Did I need to tweak the DDR /FSB/etc overvoltages so high? I kinda just bumped them up here and there when just a vcore bump didn't work. Is trial and error the only way to find the weak spot, or is there something that can help point me to a specific area?

I know I'll never get close to pushing it with normal use. (Gaming.) So should I be safe? Should I maybe aim for just Prime95 stable, at less voltages?


Other relevant components:

Gigabyte P35C-DS3R (Rev. 1.1)
6GB DDR2 800Mhz OCZ Reaper @ 5-6-6-20 (1:1)
Antec 900
Antec Truepower 650W

Apologies for the short novel but I'm hoping I covered what needed covering. If I didn't, please feel free to ask for whatever I may have missed.

Thanks.

edit: Added and corrected info.
 
Last edited:
You are good, all of those voltages should pose no issue long term. Temps nothing to crow about but are perfectly fine too. I will leave it to others to pose possible improvements and just say that getting everything "tweaked" to perfection is something that takes a lot of time to do if you really want to fine tune it as sharp as a tack. All in all - good job !

(quick snap observation, with 6GB I doubt you will be able to reduce MCH or FSB voltage much if at all and maintain stability - but no worries they are not alarmingly high. A fan glued to the MCH/Northbridge would be my main suggestion, do a forum search on Northbridge Fan but it can be as easy and cheap as using Silicone RTV adhesive and just sticking a 40mm fan to the MCH heatsink.)
 
Thank you, I appreciate the confirmation and suggestion. Sadly I went back and tried to run the Intelburn test again but it crashed this time, even Prime95 crashed. I backed it off to 3.57Ghz and it's passing the tests again.

I should also correct my earlier statement of "Prime95 stable", I should have said "Real Temp with Prime95 stable", or basically a 10 minute test using them. (Small FFT's) I had run Prime95 over night but it was at a lower clock then. My memory is terrible at times, apologies.

I guess I am going to have to play around some more to find the sweet spot.

Would you have any suggestions on what is the max vcore I should use with a 9550? (Though I think temps would kick in before I crossed that threshold.)

P.S.

I just went back to do some more tweaking and realized I ran the last 3.57Ghz test with a vcore of 1.35000 not 1.37500.

Also, I think I should have had the fans blowing from front to back instead of bottom to top. I figured being an Antec 900 case, I'd take advantage of the the large top fan. Not thinking that the chipset heatsink right below the CPU is now passing warm air directly into the CPU fan. Would it make a big difference to switch? My thinking is heat rises no matter what, so it will get sucked in, but is it better to have it "dilute" with the ambient air for that extra second or so? I would think so.
 
Last edited:
max vcore iirc on intel is 1.45 but that might just be recommended, being an amd guy i say its 1.55, but it might be different for intel
 
I think you are correct, the newer 45nm cores have Intel datasheet max of 1.45 (65nm cores 1.55) trouble is I dont remember which tech that cpu is. But either way 1.375 (and it is very likely the cpu is actually seeing a bit less) set in the bios is more than safe. I would even give 1.3825 a shot and see if stability come back and would not be too awful concerned up to 1.40 which is where I would call it quits.

There is voltage drop/droop that reduces the voltage the cpu sees compared to the bios setting so you never quite get what you ask for with your board (and most other boards) that give you a kinda built in extra saftey margin. But it is your call and sometimes a few 100MHz vs piece of mind - piece of mind wins because a few MHz is not going to really make much if any difference in your game/apps. /shrug.
 
Seems it's 1.3625 on paper at least. Most OCing discussions I can find seem to imply 1.45 is doable but at a cost. (Life)

I have good news. I upped the MCH overvoltage and lowered the vcore to 1.3675 and just passed a 5 pass, Very High Intelburn test with temps at 65C or less @3.63. (I turned my Antec 900 onto "vacuum" or high mode, but I don't think it's why the temps were lower because idle/ambient temps never change. Just the CPU cores under load.)

Looks like it's time to up the FSB some more.

Thanks for both your replies. I'm feeling pretty good about this OCing business. I'll have to look into the chipset heatsink fans,and possibly water cooling down the road. I think I'll be dabbling with GPU OCing to. My 8800GTX OC is pretty hot with it's default fan, but with the 5870 on the way, I'm hoping I can get more from it.

P.S.

CPU-Z Validation
 
Last edited:
I would get it fully stable at 3.6 with as low voltages as possible and leave it there if this is a production (i.e. general use/work/gaming and not for benchmarking/suicide runs).
 
LittlePud said:
I would get it fully stable at 3.6 with as low voltages as possible and leave it there if this is a production (i.e. general use/work/gaming and not for benchmarking/suicide runs).
Click to expand...

Agreed, that is what I think I'll do......after I finish playing. *poop eating grin*

I managed to pass RealTemp+P95 small FFTs at 3.82, but I had to bump the voltage up. I jumped from 1.375 to 1.400 on the vcore. I wont go beyond 1.4 with the vcore. Now I just need to see what I can push the FSB to before my next BSOD. The max temp RT recorded was 63C, ambient is 33-34. I'm currently idling at 41C like I was yesterday morning at 3.63.


I wish I could get some of the OCs I see others mentioning with their 9550s. Others are getting similar Ghz but at lower vcores. (Much in some cases.)
 
Last edited:
Well I am back hoping you folks can shed some light on some observations and issues.

Ok, so, my previous "stable" OC has gone to pot all of a sudden. It's stable while gaming and what not, but I can no longer pass IBT or OCCT running it @ 3.65, with a vcore of 1.375 through 1.40V. It just passed OCCT with a vcore of 1.45 @ 3.65.

Is this normal, or does this indicate possible damage? Basically any idea why it would do that?

Also, is it normal that on 2 separate occasions, I have upped the vcore, but my temps went down. The ambient was the same in both tests. Other times it either stays the same, or goes up. I believe the first time it did this was when I went from 1.35 to 1.375 and just now going from 1.40 to 1.45. I just found it odd.
 
You must log in or register to reply here.
Back
Top