• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

0437 Winchester, bad week?

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

JonEapples

Member
Joined
Dec 22, 2004
Location
Granite Bay, California
I pulled off my heatsink yesterday to reapply the AS5, and noted the stepping on my 3500 Winchester. It is a week 37 and after browsing the web a little I read some consider this a bad batch. I sure hope not, I bought it in December from NewEgg, boxed. I have had trouble getting this thing stable beyong 2.4 Ghz, but I thought it was because I'm limited with my RAM and I'm new to overclocking and still have alot to learn. Have you guys heard or experience any info on this week's batch? I looked in the steppings thread but there were only two listed 3500 Winchesters and they didn't have thier info on their steppings.
screenhunter0024gc.jpg
 
Last edited:
Looking at your sig, the more probable reason for your lack of a good overclock is that power supply. I hate to do this yet again, but there's nothing spectacular about a Thermaltake 420W Purepower. On top of that, you're running a 6800 GT, which is also known to be a bit on the power hungry side. If I were you, I'd try to get my hands on another unit (perhaps borrow from a friend) and see if that nets you a higher clock.

deception``
 
I wouldnt say that about the PSU...since I am running the same PSU ;)...

Some people at XS have had the same issues of not being able to get over 2.4 with that stepping...but this is might not be your case...

A few things that might help...

Get a XP-90 heatsink and a good 92mm fan...

Also...if you have not already...crack open the 420 and crank the +5 and +12 rails up...turn them all the way up...with them all the way up you will get a nice rail improvement and odds are will help the OC...

My Tt 420w is almost 3yrs old and runs

6x Tt2 smart fans
1x 92mm tornado
2x YS Tech 60mm fans (Pro cooling fans)
9800 Pro
1x 160 IDE HDD
1x 80 IDE HDD
1x 60 SATA HDD
1x Plextor 12x DVD-RW
1x 16/48x DVD/CD-RW combo

Along with the OC in my Sig

Also...what settings are you running for 2.4Ghz...(FSB, multiplier, core voltage, memory voltage and HT multiplier)
 
sorry off topic but you dont mind me adding your stepping to my sticky do you ?
 
glock19owner said:
My Tt 420w is almost 3yrs old and runs

6x Tt2 smart fans
1x 92mm tornado
2x YS Tech 60mm fans (Pro cooling fans)
9800 Pro
1x 160 IDE HDD
1x 80 IDE HDD
1x 60 SATA HDD
1x Plextor 12x DVD-RW
1x 16/48x DVD/CD-RW combo

The clincher here is that you're only on a 9800 Pro. Keep in mind that, on average, 6800 GT's consume a good deal of extra power. And let's not forget that both the processor and video card draw from the +12 volt, too.

deception``
 
Im with glock on this. I doubt its the PSU. If you recall deception i thought getting a powerstream would help my overclock....didnt do squat.

____

Mine tops off at the same speed bro, its just a bad week (i have a 0433)
 
deception`` said:
The clincher here is that you're only on a 9800 Pro. Keep in mind that, on average, 6800 GT's consume a good deal of extra power. And let's not forget that both the processor and video card draw from the +12 volt, too.

deception``

I fully understand this...this is why I mentioned to full adjust the hidden pots...

But also...look at my setup...especially cooling...odds are I am still pushing my Tt more then his complete setup ;)...

And with my OC and Pro flashed to XT...my +12 is 12.05 at the lowest and 12.22 at the highest...

Either way...I honestly would try to adjust the pots first before spending more money on a PSU...worse comes to worse...they spent 15 or 20 minutes adjusting the pots and are right back in the same situation...but if adjusting the pots work...then they just saved some cash...;)
 
Sentential said:
Im with glock on this. I doubt its the PSU. If you recall deception i thought getting a powerstream would help my overclock....didnt do squat.

____

Mine tops off at the same speed bro, its just a bad week (i have a 0433)

Very true, but he can barely pull off more than 200 Mhz from a 3500+, whereas you were at least able to squeeze out 500 Mhz on your 3200+. Given this noteworthy fact, I'd say that it's the power supply. I highly doubt that his 3500+ is that poor of an overclocker even with it being a week 37 chip.

deception``
 
Did anyone notice the heatsink they were using...which is why I stated about the XP - 90 also ;)

Something else also...we really need to know what settings they are tying also...which is why I asked this in my first post...never know...we all might be going the wrong way to find out they forgot to change the HT multiplier or some other small setting...
 
Last edited:
deception`` said:
at least able to squeeze out 500 Mhz on your 3200+.

deception``
Actually just as a heads up it was 400. My 3200+ wont do much over 2.4 fully stable on P95. 2.5ghz has always been a crapshoot but ive never been willing to admit it :cry:
 
Sentential said:
Actually just as a heads up it was 400. My 3200+ wont do much over 2.4 fully stable on P95. 2.5ghz has always been a crapshoot but ive never been willing to admit it :cry:

In your case, it is obvious that your processor is simply to blame :( But I highly doubt that 200 Mhz is all that his 3500+ is good for. Of course, it could also be due to a number of other factors, such as his operating bios for example.

deception``
 
deception`` said:
In your case, it is obvious that your processor is simply to blame :( But I highly doubt that 200 Mhz is all that his 3500+ is good for. Of course, it could also be due to a number of other factors, such as his operating bios for example.

deception``
Vrykyry's 3500+ wouldnt do 2.5 either. Just ask him. The old CBBFDs are *really* that aweful :cry:
 
ps2cho said:
sorry off topic but you dont mind me adding your stepping to my sticky do you ?
Yeah, go ahead and put it in the sticky. I don't know, my PSU seems to do pretty well, I don't get a BSOD when trying to run at 2.6GHz, I just fail SuperPi, Prime, 3DMark, ect. I'm confused about the way the memory clock is generated. For example right now I'm at 241x10, 166 memory divider, 4x HT. My memory clock is listed as 200.85 MHz in clockgen. Now if I switch the multiplyer to 10.5, I'm at 2.5 GHz but the memory clock goes down to 194.67 MHz. Why does it go down, when I just bumped the clock up? This is just an example of how much I don't know what I'm doing. So I hope it's probably more user error then a bad PSU or a bad week's batch from the AMD factory. :rolleyes:
 
JonEapples said:
I just fail SuperPi, Prime, 3DMark, ect.

This is a sign that your memory timings are too tight or not enough voltage to the memory...

Are you using NF3 tweaker to adjust your OC...I have seen the de-throttling when using the NF3 clockgen program...
 
O.K I just reset my rig and in the BIOS set this for overclock of 241x11. I'm at 39C CPU and 32C sys temps and typing this right know with these settings:
241x11- 2651MHz
166 memory divider 3-3-3-8 1T
4x HTT
1.55 VCore
2.70 Memory voltage
I failed SuperPi at 16K with a "not covergent in SQR05"
Should I loosen the memory timings? I always overclock in the BIOS, I just use ClockGen to see what adjustments would do to the different clocks.
 
JonEapples said:
I failed SuperPi at 16K with a "not covergent in SQR05"

A few things...clockgen is the reason for the de-throttle...which is why you probably got the errors using that...

Why are you running PC3200 at a divider? XMS should be able to run at 1:1 timings at 240...to me 8-3-3-3 is kind of loose for XMS...and it just might be too loose for Super Pi and such...running ram at too loose of timings can be just as bad as tring too tight of timings...

See if they will run at 240 1:1/1T with 8-3-3-2.5 (just to make sure it will run at 1:1 timings) and 2.75 or 2.8v...if the memory will run at those settings with 1:1 timings...then retry super pi, pifast, etc...and if they pass then...then try upping the FSB to around 250 or so...

Sidenote...some memory do not like cas3 that well...this might be the issue...
 
There is very little difference between any winchester except the ones amd speedbins. I have been told that because demand for low end is so much more than high end, amd has no choice but underclock most of them that otherwise are capable of higher. My 3000+ could just as easily be a 3500+ to be honest I never tried for its max stable overclock due to ondie controller limit. I did try 2.48GHz with my ram async and the cpu wont do it, failed superPI in seconds. I did get 3dmark to pass as this benchmark is less picky but my performance went down. 1:1 is stable to about 250fsb, maybe 255 tops, but performance is just as good and its stable too. If your unhappy with your 3500+ why not just sell it and get a 3200+ youll get both extra money and higher overclocks?
 
glock19owner said:
A few things...clockgen is the reason for the de-throttle...which is why you probably got the errors using that...

Why are you running PC3200 at a divider? XMS should be able to run at 1:1 timings at 240...to me 8-3-3-3 is kind of loose for XMS...and it just might be too loose for Super Pi and such...running ram at too loose of timings can be just as bad as tring too tight of timings...

See if they will run at 240 1:1/1T with 8-3-3-2.5 (just to make sure it will run at 1:1 timings) and 2.75 or 2.8v...if the memory will run at those settings with 1:1 timings...then retry super pi, pifast, etc...and if they pass then...then try upping the FSB to around 250 or so...

Sidenote...some memory do not like cas3 that well...this might be the issue...
I tried running at these settings, still wont pass SuperPi! I bumped up the VCore to 1.60 as well. Even now I went back to 240 x10 and with the memory at 1:1, I won't pass SuperPi. Should my memory be able to do 240MHz? :bang head
 
Back