increase your VCore gradually until you have no more freezing.Originally Posted by Orangeblast
increase your VCore gradually until you have no more freezing.Originally Posted by Orangeblast
CEO Swiftech
The random freezing is because your CPU is are probably voltage starved. Increase voltage two increments at a time in the bios, until you no longer get any crashes. once you have reached that, then try to drop back one increment, and see if you are still stable.Originally Posted by Orangeblast
I'd also try to add a smidgen of voltage on the memory.. like 2.3 instead of 2.2, but this really depends on your frequency and timings (Eric is MUCH better than me on memory matters, so take this with a grain of salt). To see if your memory is stable try to run the 4M super pi. If it passes without errors, its a reasonable test.
There is really no need to run orthos or TAT stability tests for 8 hours, unless your real life applications run for 8 hours straight. Testing your computer for day to day use is about stability with the applications that you actually run. If you play a game for several hours in a row, then by all means run that game in demo mode, or autopilot for the same amount of time you'd play it in real life, and see if it crashes.
CEO Swiftech
Thanks for the tips, I will try them out.
I also find it a bit strange that when setting it to 3.5ghz all the way to 3.75 the system does not POST properly. It will freeze at POST no matter what VCore voltage I set.
My memory frequency at this point is about 755mhz (a far cry from its standard 1066 stock working frequency)
The dominator 8500 set is to run at 5-5-5-12 @ 2.2 @ 1066mhz.
I have it at 5-5-5-12 @ 2.2 @ 755
I have the 680i linked and synced and this is why I do not have 1066. I found dramatic stability at this setting. Unlinked and 1066 causes all sorts of instability EVEN THOUGH Memtest passes for 4 hours+
evga motherboard forum has all sorts of complaints about RAM and this board.
none the less I will try out the tips. It is getting a bit frustrating though. Especially considering the HUGE gap of 3.5-3.75 not posting. I find it funny that at 3.775ghz, Vista will load and I can do some low level stuff in Vista at that speed. Yet, the system refuses to post at even 3.5ghz.
???? you got me!?!? *shrugs*
------------------------------------------
Just a bit of an Update:
Okay, I am 12 hours orthos stable at 3.4ghz
funny thing is, check out my voltages:
VCORE: 1.4875 (before Vdrop)
FSB: 1.2
VDIMM: 2.2
SPP: 1.40
MCP: 1.50
MCP-SPP 1.20
As you can see from the above, My voltages are extremely low, Infact they are below "auto" settings for that ghz frequency (forget the Vcore, I'm talking about the other voltages) and....I'm 100% stable at 3.4ghz including NO random freezing.
Its the most curious thing about this chip, any increase in ghz from 3.4 results in freezing. Example, I went to 3.475ghz and went all the way up to 1.675 Vcore (before vdrop) And still Vista would freeze at the loading screen. Its as if 3.4 is the maximum this chip will do!
I find it hard to believe given my temperature thanks to the TEC and alot of people getting 3.6 on air alone! I must have a crappy overclocker..?
would it really justify me going out and getting another E6600? damn........I was hoping to hit at least 3.7 with my TEC setup.
Another thing I've noticed is that increasing the other voltages (spp, mcp etc..) seem to have litte to NO effect on the stability of the system. I mean just look at my voltages and how I got stable at 3.4ghz. See how low those other voltages are!?!?! They might as well not even exist!
Last edited by Orangeblast; 02-28-2007 at 12:54 AM.
reflash your bios? maybe the bios integrity is bad.?
tried last night. went to p25 bios on evga 680i. still the same and now i'm getting freezing again at my supposed stable OC of 3.4.
how do you know when your CPU is taking a dive?
So is this new TEC waterblock worth it or not? Also does it need a specific PS or a normal PS (a second powerstream 520W) do the job for the TEC?
Okay, got an update:
Well, I decided to go out and get another E6600 C2D. I NEEDED TO SEE if there was a difference compared to my chip. It was killing me. So I bit the bullet and went out to my local beloved Mwave.com and picked up my new E6600. Of course one of the first things I checked was the Batch.
Its says:
BATCH: L639F039
...and of course I'm thinking, an F batch eh? So before I opened it up I did some research via good ol'e google and looked up my batch. Found that with this F batch, peeps are hitting 3.5 and it runs some what cooler than the other batches. Probably an improved IHS on this batch?
So I popped that beeyotch in and so far, I'm Super-PI 32 stable at 1.4 VCORE (before Vdrop) @ 3.4ghz. Something I deffenitly couldn't do with my other C2D.
So, so far things are looking good. I'm sure I could get alot higher with this chip vs. my other one. It IS running cooler. I'm hitting about 6 to 8 degrees celcius at load in Super Pi 32 (but thats most likely becuase I'm only at 1.4 Vcore, Before Vdrop)
I will keep this thread updated.
Anyone interested in buying a Lapped E6600?
Yeah, it seems as though it's a decent block IMO.Originally Posted by Charloz24
I would highly recommend a dedicated psu (ie. nonATX) for running the TEC. An example would be a Meanwell. Cheaper alternatives are ham/cb psu's from ebay. Or wire up your own (though, switching psu's are a bit fun to DIY).
When we going to see photos of it all naked and apart etc
I have a feeling someone "Xtreme" saw this and bought it already.Originally Posted by Pete
Lenovo Thinkpad X220 - Core i5 2410m, 4gb
waiting on 28nm video cards...
I measures the edge to the edge and got 58mm square. Maybe the top is bigger. If you try to put a 62mm x 62mm tec it will fail and take out your beloved CPU with it. Sorry, I'll run with the tec it came with.Originally Posted by m0da
Ugh?Originally Posted by serialk11r
Wait I'm not quite understanding this.. where exactly did you measure it I guess? Could we see a photo please, if it would be conv. for you? I had the impression that the base of the waterblock was quite similar to the base of the 5002, except that the diamond pins are shorter. Is that true? Because the 5002 can hold a 62mm TEC...Originally Posted by Philly_Boy
I haven't seen this discussed anywhere, but check out the pressure drop/flow rate graphs for the MCW6500-T, Storm, and MCW6002
Am I reading things correctly - at 1GPM, the MCW6500-T has a higher pressure drop than the Storm (around 5.7PSI compared to around 1.7PSI) which makes it even more restrictive? Assuming the test methodology in both graphs is the same, which is not a given - but still, what would account for that huge discrepancy, considering the Storm is one of the more restrictive designs around?
The 5002 cannot hold a TEC, at least properly. The cooling area is too small.
If you have a cooling question or concern feel free to contact me.
Yeah, you are reading it correctly. I am pretty sure it was posted somewhere that the mcw6500-t had 2-3 times the pressure drop. The block makes the water pass through the little pins creating a massive pressure drop.
http://www.xtremesystems.org/forums/...hlight=mcw6500
Last edited by thecoldanddark; 03-12-2007 at 02:22 PM.
Stock MCW5002 series came with a 50x50mm TEC, which fit since the cold plate is 63x63mm.
Thanks for posting that link, serves me right for skipping right to the charts on the Swiftech site. After looking at the inside pics of the block it came to me why the drop is so high, explained succinctly in this post:
http://www.xtremesystems.org/forums/...3&postcount=33
Low height top forces water through the pins, not on top.
Got mine today
. ss phase . nf11fx compressor . chilly1 evap . r507 . chilly1 condenser . freezepack controller . 965xe 4.63ghz || evga x58 classified || 2x evga gtx 295 || 6gb g.skill 1.6ghz cl8 || corsair ssd x64 || tt 1.2kw w/ 12 awg || win7 x64
Sorry AFI I meant a 62mm TEC, though the block wasn't so good to begin with. The new one looks very good, though a few more mm of pins wouldnt have hurt. (vertically)
If you have a cooling question or concern feel free to contact me.
Bookmarks