Requested proof of Thermalright quality control issues

[Notlag]

Last question, how do you remove the springs on the bolts, I thought there is something that holds it on....I'll have to tear it apart later

Lol - I discovered it by accident when one of my screws fell loose in my case then got wedged in a spot and I had to tug on it a little. It may have been a bit looser to begin with, but there's just a ridge on the bold that holts the spring on. If you use a pair of pliers or something the springs pop off with a bit of a pull. When you put them back on (say you didn't use washers after all) they'll generally always be a bit looser than before and may fall off the bolt more easily, but it has no effect on their performance (just that when you handle them, if you have a loose one you may have to hold it to keep it from falling off while you position the screw to tighten it down).

 

[Automata]

Lol - I discovered it by accident when one of my screws fell loose in my case then got wedged in a spot and I had to tug on it a little. It may have been a bit looser to begin with, but there's just a ridge on the bold that holts the spring on. If you use a pair of pliers or something the springs pop off with a bit of a pull. When you put them back on (say you didn't use washers after all) they'll generally always be a bit looser than before and may fall off the bolt more easily, but it has no effect on their performance (just that when you handle them, if you have a loose one you may have to hold it to keep it from falling off while you position the screw to tighten it down).

Ok, will do! Thanks!!

It may be awhile before I do the tests.

 

[Notlag]

Ok, will do! Thanks!!

It may be awhile before I do the tests.

Just don't go too crazy w/ too much pressure Unless you have something to measure the pressure created by those springs, it will be kinda' hard to tell when you're exceeding Intel's spec's and by how much. There's a couple things you could do to estimate it, but... I've never heard of somebody tightening down a heat sink so far they crushed their socket or CPU (at least not since they started including IHS's), but I wouldn't want you to be the first

 

[Drinkyoghurt]

Could you get temp readings with CoreTemp? I have trust issues with TAT.

stock temps look pretty.

here you go Ink

idle temp:



and load temp

 

[Automata]

Just don't go too crazy w/ too much pressure Unless you have something to measure the pressure created by those springs, it will be kinda' hard to tell when you're exceeding Intel's spec's and by how much. There's a couple things you could do to estimate it, but... I've never heard of somebody tightening down a heat sink so far they crushed their socket or CPU (at least not since they started including IHS's), but I wouldn't want you to be the first

Hehe, I won't put that much on

I've had a bit of experience with bolting stuff down. I create a bolt-through it for my waterblock since it had those acorn nuts....and with a lapped processor, that was a no-go. This socket/processor can take some pressure

 

[Notlag]

here you go Ink

idle temp:



and load temp

Out of curiousity, what is your room ambient temps, are those temps w/ the side panel on or off, and can you run it w/ SpeedFan (or something else that reports the CPU temp given by the motherboard - the single CPU temp as opposed to the core temps).

I'm wondering if the Tjunction used by CoreTemp is right. The 21C / 22C idle temps have got to be darn close to ambient temps (that's only 69F or 70F) which would be almost impossible on air cooling - especially not overclocked (idle your processor still uses, very roughly, around 1/4 to 1/5 the power it does at load). If the temps are accurate though then that is amazing. If they're not accurate then odds are you need the 100C TJunction, in which case you'd have to add 15C to those, but that's still very good.

If you can, read my thread on verifying your Core Temp / TAT accuracy:

http://www.ocforums.com/showthread.php?t=532020

It includes some procedures for verifying whether you should use 85C or 100C Tjunction and how to estimate temps more closely yet.

EDIT: If you search around on 'E6420 Tjunction B2' on Yahoo / Google and read a number of posts you'll probably come across some talking about later B2's posting really wierd, even impossibly low, temps.

 

[Notlag]

Hehe, I won't put that much on

I've had a bit of experience with bolting stuff down. I create a bolt-through it for my waterblock since it had those acorn nuts....and with a lapped processor, that was a no-go. This socket/processor can take some pressure

lol - Yah, figured you'd be ok, but didn't want to say that w/o some disclaimer lol

 

[Automata]

lol - Yah, figured you'd be ok, but didn't want to say that w/o some disclaimer lol

At the very least, I'd have pictures!

 

[Drinkyoghurt]

Out of curiousity, what is your room ambient temps, are those temps w/ the side panel on or off, and can you run it w/ SpeedFan (or something else that reports the CPU temp given by the motherboard - the single CPU temp as opposed to the core temps).

I'm wondering if the Tjunction used by CoreTemp is right. The 21C / 22C idle temps have got to be darn close to ambient temps (that's only 69F or 70F) which would be almost impossible on air cooling - especially not overclocked (idle your processor still uses, very roughly, around 1/4 to 1/5 the power it does at load). If the temps are accurate though then that is amazing. If they're not accurate then odds are you need the 100C TJunction, in which case you'd have to add 15C to those, but that's still very good.

If you can, read my thread on verifying your Core Temp / TAT accuracy:

http://www.ocforums.com/showthread.php?t=532020

It includes some procedures for verifying whether you should use 85C or 100C Tjunction and how to estimate temps more closely yet.

EDIT: If you search around on 'E6420 Tjunction B2' on Yahoo / Google and read a number of posts you'll probably come across some talking about later B2's posting really wierd, even impossibly low, temps.

i think room ambient is around 18-20c, and yes those temps are with the side panel off...i think it reports just right. why? when i had a stock cooler on it at 3.4ghz it gave me a 59c temp reading, that sounds just about right with stock cooler, now it gives me around 36c load with 3.2ghz(45c with stock)

 

[Notlag]

i think room ambient is around 18-20c, and yes those temps are with the side panel off...i think it reports just right. why? when i had a stock cooler on it at 3.4ghz it gave me a 59c temp reading, that sounds just about right with stock cooler, now it gives me around 36c load with 3.2ghz(45c with stock)

Core temperatures aren't measured as absolute temps. They're reported as the difference between the thermal diode temperature and the temperature where thermal management kicks in (Tjunction max). So the real reading might be like -40C or something like that. For CoreTemp or TAT to report the absolute temp they have to add the value of Tjunction to that number. The problem is that Tjunction varies from chip to chip and can't be read from the chip, so based on the stepping and such they try to determine whether to use 85C or 100C. Not all versions choose the right Tjunction. On top of it, the real Tjunction is not necessarilly exactly either 85C or 100C on desktop processors (but tend to be in the ballpark of one of the two).

Core temps being only a couple degrees over ambient at idle is pretty cool, and I've seen other threads about E6420 B2 steppings reading off. The one way to confirm it is to crank down your voltage and speed to the lowest you can boot at (I can get down to 600MHz and 0.85v). At that low of a setting your processor should be running at very very near ambient temps, but always has to be over ambient temp (unless you're using a TEC or phase change), so if your core temps are below ambient at that speed then CoreTemp / TAT is reading low.

The thread I referred you to discusses this phenomenon and the process of validating those temps.