Thermalright Speaks on the Ultra-120 Extreme Problems!

WhiteZero

WhiteZero

2[H]4U
Joined
Oct 21, 2004
Messages
3,638
I decided to send Thermalright an email concerning some of the issues with the widly popular Ultra-120 Extreme.
I'll be sure to update this thread with any future EMails.

My EMail to them:
I'm very interested in your Ultra-120 Extreme product. My concern, though, is that pretty much everyone I've heard of getting this heatsink says that the base of it, where the heatsink meets the CPU, is never very flat. Many people are displeased with the current build quality of the Ultra-120 Extreme and usually have to manually lap the surface so it's flat.

I'd like to know if Thermalright is actively doing anything to improve the build quality of this otherwise top-tier cooler? The competition (Thermaltake, Zalman, etc) dont seem to have nearly as big of a problem with this issue. And surely Thermalright's reputation could only improve in the enthusiast market if you were to work on your build quality of these heatsinks.

Thank you for your time!

-"WhiteZero"
-[H]ardForum Member
Click to expand...

Thermalright's Reply (which was very speedy and on the weekend no less!):
Hi,

Thermalright has stated and continue to state that the design of the base is meant to be convex, because the heat spreaders on most Intel CPU's is concave, thus matching both at the most important point. The only other company that does this is Swiftech on a few of their CPU water blocks, but they also use a seal around the block that tends to hide this fact.

People are used to the old days when cooler bottoms had to be flat and polished because the application was directly on the core of the CPU, but that is no longer the case today, due to the use of heat spreaders, but most people resist change and still want it the way it was.

As I look at just about every heatsink that hits the market, none of them, at least those that end-users get, consistently match the quality of Thermalright products, but you would expect me to say that, so look for yourself.

Now the real problem that's bothers people is that you can move the heatsink once it is installed and until now that has never been the case, but Thermalright will soon have an accessory that will act as the seal on the Swiftech water block, preventing the ability to move the installed heatsink.

Your own site had this review posted http://enthusiast.hardocp.com/article.html?art=MTQwMiwxMSwsaGVudGh1c2lhc3Q=

Thermalright Support,

Bob
Click to expand...

Great, non-automated response. Very cool. To which I've replied:
Thank you for the quick and personalized response, Bob. I appreciate it greatly, considering I work for a company that responds to emails mainly with canned responses. ;)

I do see the logic behind Thermalrights reason for having the base of the heatsink to be convex. However, in practice, it seems to have been proven that when one laps the base of the heatsink, you seem to get better results, as seen in the following popular example: http://www.hardforum.com/showthread.php?t=1188093

I suppose this could just be a fluke since this users considerable temperature drop occurred when he laped the CPU's heat spreader after the heatsink. But his 2C drop in temperature is quite noticeable. However, I guess I shouldn't argue too much with Thermalright's designers, they probably have a better grasp on thermal dynamics than I do.

I'm also very very pleased to hear that Thermalright is working to fix the flimsy mount of the Ultra-120 Ex. That was one of the main things that was scaring me away from that product.

Again, thank you for your personal attention to this matter. The community as a whole appreciates it.

-"WhiteZero"
-[H]ardForum Member
Click to expand...

Thermalright's responce:
Hi, WhiteZero,

All our tests showed negative results, that is if we didn't lap the heat spreader as well. Lapping both did show anywhere from .5 to 2C improvement and that had allot to do with the systems the both of them were running in. housekeeping, non-disrupted airflow and ambient air temp had as much bearing as the lapping did, in other words, with good case housekeeping, good to better than average airflow and ambient air temp of no more than say 24C, I was able to achieve the same type of improvement without lapping.

As I said earlier and with lapping of both heatsink and CPU, the user was able to nullify the movement of the heatsink, which I'm sure made him feel better and could have been something that added to his success. Now the devise our engineers have come up with will nullify the movement and I'm sure that once it hits the streets many will be claiming improved performance.

I should have a few samples early this week to test on my new rig, it should be interesting, to say the least!!

Thermalright Support,

Bob
Click to expand...

-----------------------------------------------------------------------------------------------------------------------------------------------------------------

My next EMail:
Can you comment on whether or not this new accessory for the mounting will come with standard with new shipments of the Ultra-120 Ex. or will it only be available for separate purchase?

Thanks!


-"WhiteZero"
-[H]ardForum Member
Click to expand...

Reply:
Hi WhiteZero,

That I can't tell you at this time because I simply don't know.

Thermalright Support,

Bob
Click to expand...
 
So they're resorting to it being intentionally *bowed* by design. Interesting. Now I'm curious to see temps with it lapped-vs-straight out of the box to see if their claims are found.
 
from what I can see having 2 different true 120's....

my first one SUCKED, it was WEIRD, the mount was loose as hell and I dunno why.
it wasnt any better than my tuniq tower.

I did order a second one and for whatever reason it seems much better, the mount is tighter without modification on more than 3 mobos I ahve had it on. Also it runs cooler then the first one on the same cpu. Same scissor bracket also.

Dont ask me....
 
when I lap I've always gone all the way, both heatsink and cpu as I've found both of them to be lacking in mating as evidenced by the uneven spread patterns when you first do test mounts. It's just the nature of things, for the most part, I've gone through 5 generations of Thermalright CPU heatsinks and everyone of them has performed above and beyond my expectations of air cooling. All of those heatsinks are still in use today, 1 went to a friend, 1 to a family member and 3 of them are still in systems around the house.
 
That gets a gold star in my book, thermalright actually being in touch with the enthusiast community and not just another company out money as has been seen all too much in the past.

I don't have a u120 extreme but unless theirs something when i decide to get another case/computer, ill definitely be looking at one.

I would think the 2c difference is just because when you lap the base, you remove the nickel plating and get to to the copper base underneath (well to get it flat, but theirs copper under there) which conducts heat better than the coat of nickel.
 
If that were true then why wouldn't the whole base be copper? (Just curious, am N00b.) Aka. why nickel plate it?
 
"The process of oxidation depends on the amount of oxygen present in the air and the nature of the material it touches."

If I am not mistaken, there is no oxygen in contact with the base? (This could, of course, explain why I didn't think of oxidation off the bat.)
 
Anomili said:
"The process of oxidation depends on the amount of oxygen present in the air and the nature of the material it touches."

If I am not mistaken, there is no oxygen in contact with the base? (This could, of course, explain why I didn't think of oxidation off the bat.)
Click to expand...

Air is everywhere ;)
 
I believe they also nickel-plate it so it fits in better with the aluminum fins.
 
The real reason why plate the entire thing (AFAIK) is because it's quite hard to selectively plate a material. The entire thing is probably immsersed in the plating solution, which prevents control over which surfaces get coated and which do not. If you've ever done your own anodizing, you'll know what I mean!

It would not really be economical to re-lap each heat-sink after plating, I'm sure...

Sphere said:
Air is everywhere ;)
Click to expand...

It damn well better not be between the HSF and the IHS... If it is, you've got *major* problems.

Zero82z said:
I believe they also nickel-plate it so it fits in better with the aluminum fins.
Click to expand...

Yep, pretty much why they do it.
 
Very cool that TR responded to community queries like that.
However the information about the bowed bases being intentional on their part, and their reasoning for this, has been in circulation for a while now. Its been restated in reviews, forums, and posts like this repeatedly. Its like people are ignoring it on purpose. I dont get it.

My guess as to why the lapping of both produces superior results is that that is the best possible surface you can obtain, as you're matching both up to be flat manually.(duh) However, TR has no control over the quality of Intel's product so they alter the things they DO have control over, namely matching their sinks shape to the mainstream processors. This makes it harder for the hardcore to lap, but produces superior results for the majority market. (As compared to if they didnt bow it.) I'm sure if they could somehow cause the Intel IHS to be flat, they would have made the u120 flat. Its not like they couldn't have machined it flat; TR isn't so stupid that they would do something like this for no reason. So basically:
[IHS Lapped + HSF Lapped] > [IHS Stock + HSF Bowed] > [IHS Stock + HSF Flat]

Cool that they are coming up with an official fix for the mount, though a 5 cent rubber/neoprene washer will still do the trick. ;)
 
I've had this stock AMD 4 heatpipe heatsink with lapped copper base (lapped it when I got it) and it doesn't really show much if any oxidation (a little dirty and unpolished looking on the bottom), all i would need to do is to just rerun it with 1000/1500/2000 grit sand paper again real quick to get the oxidation off.
 
Rattle said:
from what I can see having 2 different true 120's....

my first one SUCKED, it was WEIRD, the mount was loose as hell and I dunno why.
it wasnt any better than my tuniq tower.

I did order a second one and for whatever reason it seems much better, the mount is tighter without modification on more than 3 mobos I ahve had it on. Also it runs cooler then the first one on the same cpu. Same scissor bracket also.

Dont ask me....
Click to expand...

What where the temperatures between the two Ultra-120extremes?
 
Arcygenical said:
The real reason why plate the entire thing (AFAIK) is because it's quite hard to selectively plate a material. The entire thing is probably immsersed in the plating solution, which prevents control over which surfaces get coated and which do not. If you've ever done your own anodizing, you'll know what I mean!
Click to expand...

Yeah that makes more sense. Thanks for clearing it up :)
 
The TRUE is a superb heatsink, no question about it. I used one for a year, great results. But I have replaced it with a Noctua NH-U12P. Temps are about the same, but the mounting system and base are far superior:cool:
 
ty1er said:
What where the temperatures between the two Ultra-120extremes?
Click to expand...

From AnandTech's latest cooler review:
loadmf0.png

(The Extreme is using an S-Flex fan as well)
 
Despite reading this thread I ordered the Ultra 120 Extreme. Kind of thinking I should have just gotten the Ultima 90.

Oh well.
 
The Ultra 120 extreme really shines when you start adding in more heat, at lower loads, it 's on par with even it's predecessor.

What some people need to realize is that chart is also based on an ideal type processor, lower voltage and high overclocks. Most of us that get into overclocking will find that we will be producing heat akin to those results in the 3.7 - 3.9 range.

My regular 120 was hitting avg 65C with near 80C core temps at 3.4 Ghz, I'm running 3.6 with the 120 extreme with more voltage and subsequently more heat, and the highest I've seen a core hit is 73C, while the CPU is content to stay around 60C or less...
 
I must might have to play with an overclock to get more of my money's worth out of it. If not, in the end, the extra money I spend won't be missed.

Still seems silly spending nearly $70 after tax and shipping for a heatsink to go with a processor that cost around $200 :eek: I got the E6750 so that $70 extra dollars for the Q6600 is cooling my C2D lol
 
The part I find weird is that to go with a Q6600, people will spend around an average of $1500 for their setup, yet balk at $70 for a heatsink...

Reminds me of the people who buy a sports car capable of 150mph yet buy tires only rated to 115mph... or buying a slurpee to quench your thirst and to pick up one of the dinky straws... weird mentality huh?
 
CRXican said:
Still seems silly spending nearly $70 after tax and shipping for a heatsink to go with a processor that cost around $200 :eek: I got the E6750 so that $70 extra dollars for the Q6600 is cooling my C2D lol
Click to expand...

Yeah, the reason you spend $70 on cooling is to make that $270 processor (eg.q6600) perform 40% faster than a $540 processor (Q6700)
:p
Seems like good economy to me.
 
I like economics... though figuring that 3.2-3.6 intel quads don't exist yet in terms of retail, just how much more you are saving by going with a quality $70 cooler combination which is proven in the real world.
 
WhiteZero said:
Yeah, the reason you spend $70 on cooling is to make that $270 processor (eg.q6600) perform 40% faster than a $540 processor (Q6700)
:p
Seems like good economy to me.
Click to expand...

Ya, that makes sense. I admit I'm not into overclocking. I buy the badass heatsinks to keep things cool and long lasting at stock speeds. Never have to worry about temps or turning of the computer even in the summer time.
 
blackwinterday said:
May i come to a conclusion that TR are dismissing AMD cpu's with this product?
Click to expand...

Nope!

I've got an earlier version of this HSF sitting on a 165 Opty right now. It got to 2.8GHz dual P95 stable for 24 hours on air. Just couldn't hit the big 3.0


The only difference is the mounting hardware - In fact, I've got mine rotated 180 degrees from stock to clear my RAM fans on my DFI Lanparty UT nF4 socket 939.
 
I installed this heatsink this weekend. It felt pretty darn sturdy to me but I didn't try and spin it intentionally. The contact seems very good and my idle temps in Windows on my E6750 were at or below 30C. The base looked flat but I just eyeballed it.

I haven't done any load testing though. Oh and this thing is a freaking beast, it barely fits inside my CM Stacker. Can't argue with the silence of a 120mm Yate Loon at full speed though
 
CRXican said:
I installed this heatsink this weekend. It felt pretty darn sturdy to me but I didn't try and spin it intentionally. The contact seems very good and my idle temps in Windows on my E6750 were at or below 30C. The base looked flat but I just eyeballed it.

I haven't done any load testing though. Oh and this thing is a freaking beast, it barely fits inside my CM Stacker. Can't argue with the silence of a 120mm Yate Loon at full speed though
Click to expand...

I'm sticking these things in Lian-Li PC-70s, and if the fit were any tighter I'd need vaseline.

That being said, I didn't bother lapping and at stock speeds I've dropped 20 degrees off my Q6600 at fulll load. The cooler itself doesn't rotate - the retention bracket seems to be doing it's job, and I'm idling at 30 degrees.

Today, I'll be turning the screws on this thing till it squeals, though. I'm looking forward to seeing how much I can get out of it...
 
It's pretty cool in my room but just for fun:

idletemps.jpg


As you can see winamp is playing and I have rivatuner running. Pretty much idle though.
 
You must log in or register to reply here.
Back
Top