Requested proof of Thermalright quality control issues

[Notlag]

Ok, so on to the good stuff. Here's a quick recap of the temperatures (I've included the deltas in parenthesis - really the deltas are the real ones people should be comparing because my ambient temps were higher for the 3rd heatsink). The temps were generated using TAT. Based on other tests I've done I think TAT itself reads about 7C high for me across the board, but I haven't factored that in (as I don't think most people have done much to correct for their TAT / CoreTemp temperatures). For those that care, my CPU at the speed / voltage below w/ TAT at 100% produces around 85W of thermal power, giver or take a couple of watts(which I've confirmed by comparing wall-outlet amperage at various speeds).

Core Temperatures - E4300 @ 3GHz, 1.325v, 100% Load
(Difference Between Core Temp and Ambient Temp)
TT Maxorb: 67C (43C) [75F Ambient]
First two TR U120X's: 66C (42C) [75F Ambient - horizontal mounting]
3rd TR U120X: 65C (39C) [78F Ambient - horizontal mounting]
3rd TR U120X: 63C (37C) [78F Ambient - vertical mounting]

Interpret that how you will. The first two TR's barely beat what should have been a much less capable Maxorb. Even if you give them the 2C benefit of the doubt had they been vertically mounted, that's still a narrow win. The 3rd heatsink from Thermalright does do somewhat better, managing temperatures 3C better than the first two.

So were the first two defective? Was the third also but a bit less so (the base was much better, which might account for most of the difference)? Or are the current parts being produced by Thermalright not quite what they're cracked up to be? A Maxorb nearly tied with the first two U120X's and a Tuniq Tower 120 would likely perform about the same as my 3rd U120X (maybe even a touch better, but likely not by much of a margin either way). Bear in mind this is for a dual core system - quad cores might fare a bit better performance from the Thermalrights, but the numbers are hardly earth shattering.

(mounted same way as original tests - horizontally)

(rotated 90 degrees - mounted vertically)

 

[Notlag]

What PSU is that I'm liking the UV (or did you sleeve?)

Thx Yah, what you're seeing is sleeving. The PSU is an Ultra X2 modular (which I had from my previous PC), but in the P180B case the PSU is at the bottom of the case so I needed extensions to reach a lot of things. So what I did was bolt some brushed nickle drawer pulls in the case as cable tie points and then ran the Ultra power cables part way then the UV green extension the rest. The power cables came sleeved, but I bought sleeving for the fan cables and did it myself. Got most of it from http://www.performance-pcs.com

 

[Notlag]

Btw, for the record, at this point I'm happy enough w/ my temps. Not necessarilly pleased with the U120X's - I don't think they're living up to hype or potential - but if I lap it, throw a somewhat higher output fan on it, and provide some fresh air via a side inlet fan that should likely get me down another 5C or so. While 3GHz is not a fast clock speed, it's a 66% overclock for that CPU, so I'm probably burning as much power as some people running more in the 3.6GHz neighborhood who started w/ faster chips to begin with (ie - the temps aren't great at all given the reputation and potential of the U120X's, but they could be worse too).

That won't be stellar, but I can feel comfortable getting a bit more speed out of it and call it a day. I've ran some numbers on the Tuniq and I don't expect it will do much better than the 3rd heatsink (should be in the same ballpark). I'm not sure I want to burn another $45 on yet another heatsink to further prove a point (since I had to buy the Maxorb for this thread). Though I suppose it would give me an excuse to replace my girlfriend's HSF when all the testing is done - lol (since I'll have heatsinks coming out my arse).

 

[bing]

Glad to hear that your 3rd one is better !

Btw, just curious, how thick is that solid bracket plate compared to a quarter dollar ? Example comparison with my TT120 bracket HERE.

What is your CPU VID ?

 

[Karadhas]

and a Tuniq Tower 120 would likely perform about the same as my 3rd U120X (maybe even a touch better, but likely not by much of a margin either way).

I know they're hardly comparable but even OCed to 3ghz an E4300 shouldn't output as much heat as a 4ghz prescott, and at 4ghz and 1.45v I was under 55C priming. Ofc I may have God's Own Prescott 550 but it seems you're getting U120X's from a particularly bad factory.

Then again Tuniq claims to test every heatsink they make before they ship them, so who knows.

 

[Notlag]

I know they're hardly comparable but even OCed to 3ghz an E4300 shouldn't output as much heat as a 4ghz prescott, and at 4ghz and 1.45v I was under 55C priming. Ofc I may have God's Own Prescott 550 but it seems you're getting U120X's from a particularly bad factory.

Then again Tuniq claims to test every heatsink they make before they ship them, so who knows.

Yah, if you're getting 55C core temp's priming on a Prescott that does say something about these U120X's, doesn't it? At most I'm pulling down about 85W to 90W. Your 550 is going to be way way over that number and I doubt you operate your PC in a meat locker

I am curious how the Tuniq's live up to their claim (says it right on their website). Just having a hard time justifying another $45. Admittadly, if the Tuniq does much better I can yell at TR some more until they send me a U120X that beats the Tuniq or give me some line like how certain heatsinks will perform better in certain systems and there's too many variables

 

[Notlag]

Glad to hear that your 3rd one is better !

Btw, just curious, how thick is that solid bracket plate compared to a quarter dollar ? Example comparison with my TT120 bracket HERE.

What is your CPU VID ?

CPU VID is 1.213v.

I don't have the bracket in front of me, but it was about as thick as the AM2 bracket that comes w/ all the U120X's. Around quarter thickness would be a fair guess - though I might put it more like a dime

 

[Karadhas]

Yah, if you're getting 55C core temp's priming on a Prescott that does say something about these U120X's, doesn't it? At most I'm pulling down about 85W to 90W. Your 550 is going to be way way over that number and I doubt you operate your PC in a meat locker

I am curious how the Tuniq's live up to their claim (says it right on their website). Just having a hard time justifying another $45. Admittadly, if the Tuniq does much better I can yell at TR some more until they send me a U120X that beats the Tuniq or give me some line like how certain heatsinks will perform better in certain systems and there's too many variables

Well it's not all good. Anecdotal evidence and scattered reports point to them performing absolutely horribly for quadcores despite being the previous air heatsink of choice for the C2D 4ghz club right here on OCF. They're a bit of a mystery really. Kept a 4ghz prescott cool, but couldn't keep a 1ghz 1.1v q6600 B3 under 55c when idle in a 70 degree room.

 

[Notlag]

Well it's not all good. Anecdotal evidence and scattered reports point to them performing absolutely horribly for quadcores despite being the previous air heatsink of choice for the C2D 4ghz club right here on OCF. They're a bit of a mystery really. Kept a 4ghz prescott cool, but couldn't keep a 1ghz 1.1v q6600 B3 under 55c when idle in a 70 degree room.

That's really bizare. Was it the exact same heatsink (ie - same exact one moved from one machine to another or processor swapped or what-not)? Or a different part altogether (maybe defective)? All four cores, or were just a couple of them affected?

If the base was convex on one axis and only makes contact w/ the middle of the heat spreader then it could do worse on quads than duals, or even worse on duals when rotated 90 degrees. Or it might mate well with one CPU IHS, but a CPU w/ an IHS that was shaped differently might not mate as well with it. Did lapping resolve the issue?

In my case I'm running just a dual-core, so the Tuniq would probably be my next best bet if I wanted to avoid Thermalright. Some other possibilities, like the Scythe Ninja Plus B or 3RSystems Iceage 120, won't clear my northbridge or tall memory w/ the fan attached.

 

[nd4spdbh2]

so the 3rd one didnt even work!... i dont think its the heatsink AT ALL... ever thought the tim and contact between the IHS and the 2 cores could be bad? i could see that easily...

so have we even tried these "pos thermalright ultra 120 extreems, that everone says are great" on another cpu? maybe like a cpu with its lid poped?

 

[d4m0n]

so the 3rd one didnt even work!... i dont think its the heatsink AT ALL... ever thought the tim and contact between the IHS and the 2 cores could be bad? i could see that easily...

so have we even tried these "pos thermalright ultra 120 extreems, that everone says are great" on another cpu? maybe like a cpu with its lid poped?

If that were the case, how do you explain the performance of the MaxOrb? I don't rule out the possibility that he got 2 duds in a row.

 

[Spawne32]

so the 3rd one didnt even work!... i dont think its the heatsink AT ALL... ever thought the tim and contact between the IHS and the 2 cores could be bad? i could see that easily...

so have we even tried these "pos thermalright ultra 120 extreems, that everone says are great" on another cpu? maybe like a cpu with its lid poped?

Theres no way to tell until he gets the 4th one.

 

[WonderingSoul]

If that were the case, how do you explain the performance of the MaxOrb? I don't rule out the possibility that he got 2 duds in a row.

It explains it because his mounting could be wrong or TIM not spread correctly, or proper pressure. The MaxOrb doesn't explain much other than it can cool his CPU.


In my opinion I still say it's the fan. Get a different fan, or some other kind of fan configuration. Something from Delta, Panaflo, or Yate Loon.

 

[Spawne32]

It explains it because his mounting could be wrong or TIM not spread correctly, or proper pressure. The MaxOrb doesn't explain much other than it can cool his CPU.


In my opinion I still say it's the fan. Get a different fan, or some other kind of fan configuration. Something from Delta, Panaflo, or Yate Loon.

well this picture leads me to believe that more then likely he just lapped his processor wrong. All this time i was assuming he was using an unlapped processor. Now you have to factor in the error margin of that also.

 

[d4m0n]

My point is: Wouldn't a poorly lapped CPU or bad heat transfer from the cores to the IHS yield higher than expected temps with the MaxOrb as well? I'm certainly not a ThermalRight critic, but are they above making the occasional dud?

 

[Spawne32]

My point is: Wouldn't a poorly lapped CPU or bad heat transfer from the cores to the IHS yield higher than expected temps with the MaxOrb as well? I'm certainly not a ThermalRight critic, but are they above making the occasional dud?

there are so many variations in the testing here now i cant even wrap my head around it lol

 

[UglyChild]

well this picture leads me to believe that more then likely he just lapped his processor wrong. All this time i was assuming he was using an unlapped processor. Now you have to factor in the error margin of that also.

Its not lapped properly because its not shiny?? Common now, it doesnt have to be shiny to perform well. And no one can tell from the pic that the IHS was not lapped properly.


If you want to know if it was Lapped properly, then ask him to show you a pic with a knife edge on the IHS instead of assuming all the time.

 

[Notlag]

so the 3rd one didnt even work!... i dont think its the heatsink AT ALL... ever thought the tim and contact between the IHS and the 2 cores could be bad? i could see that easily...

so have we even tried these "pos thermalright ultra 120 extreems, that everone says are great" on another cpu? maybe like a cpu with its lid poped?

The overall high temps could in theory be the thermal interface between the die and IHS, BUT that would affect the Maxorb equally, so it doesn't explain the Maxorb being a near tie w/ the first two heatsinks. The third was about 3C to 4C better than the first two, so about 4C to 5C better than the Maxorb.

 

[Notlag]

It explains it because his mounting could be wrong or TIM not spread correctly, or proper pressure. The MaxOrb doesn't explain much other than it can cool his CPU.


In my opinion I still say it's the fan. Get a different fan, or some other kind of fan configuration. Something from Delta, Panaflo, or Yate Loon.

The Antec Tri-cools are really well reviewed fans. See the review from SPCR:

http://www.silentpcreview.com/article695-page3.html#tricool

I shouldn't need to put a 150CFM Delta beast on there to get the job done. I've tried the fans push, push-pull, and pull. I've also now rotated the heatsink and it made only a 2C improvement.

As for the TIM, I've tried many different amounts and methods of application and I've unmounted and mounted these things many times. There's no issue w/ the mounting or TIM. Even if there were, it would have to be a grossly bad application of it for it to affect temperatures enough to explain these results.

 

[Spawne32]

Its not lapped properly because its not shiny?? Common now, it doesnt have to be shiny to perform well. And no one can tell from the pic that the IHS was not lapped properly.


If you want to know if it was Lapped properly, then ask him to show you a pic with a knife edge on the IHS instead of assuming all the time.

Did i say that? No. All im saying is, we have all these bad temps, and here all of the sudden, we find out were dealing with a lapped processor..... 2+2 does = 4 right? Id like to see the razerblade/flashlight test.