Thermalright QC Issues

[Elcs]

When I got my Thermalright Ultima 90 at Christmas, I never checked the base despite all of the calls from the TRUE crowd complaining about non-flat bases. I lapped my Ultima 90 not because of a non-flat base but because I wanted to do a bit of DIY work. As I am sure some of you will appreciate, once you've put together a few PC's, fiddled around with some basics and found something you "obsess" over then you find more things you want to do and try.

Recently after replacing my 8800GT with an HD4870, I found a lot of heat buildup (due to the fact that the fan is loud and probably Powerplay issues where the idle temps are really high.

For a bit of fun and a concerted attempt to remedy my heat/noise issue, I purchased a Thermalright HR-03 GT. I looked at it and the Accelero S1 Rev 2 closely and despite the price difference, the Thermalright seemed to perform better and was smaller. The size of the Accelero S1 Rev 2 may have been an issue in my case so the Thermalright looked like the right option for me.

Upon opening the box, I was impressed with the light weight of this rather large heatsink. I had my credit card beside me and after reading some reviews I saw they showed the mate between the two surfaces so I tested. From my memory, seeing daylight is a bad thing.

I am a little frustrated that one of the biggest and best companies in the field has such a simple issue with quality. I do not mind this issue to be perfectly honest because I am quite confident in my skills to improve the mating surface but if I was not prepared to tackle this, it could be a rather odd RMA battle. I have also purchased from a non-recommended e-tailer from Thermalright which also puts me further down the queue.

I apologise for "ranting" but at the very least, its another little warning on Thermalright QC.

Thank you.

 

[JEDIYoda]

There is no heatsink manufacturer that produces perfectly flat heatsink bases.
I would never call that a quality control issue!

I have used many many heatsinks and have never lapped a one!
I have never had temp issues with any of the over 60 heatsinks that I own!

 

[RallyMaster]

Why do you have 60 heatsinks?

 

[Zap]

What computer enthusiast DOESN'T have too many heatsinks?

 

[imported_Scoop]

Originally posted by: Zap
What computer enthusiast DOESN'T have too many heatsinks?

A poor one?

 

[Zap]

Originally posted by: Zap
What computer enthusiast DOESN'T WANT TO have too many heatsinks?

Okay, fixed!

 

[The Odorous One]

Thermalright doesn't care, I know first hand. I had very good luck with an Ultra 120, nice flat base and great clamping mechanism. So, I decided for my second rig to go with a 120 Ultra 120 Extreme. Major disappointment. Although the sink itself was great, the base was convex. To make matters worse, they botched the clamping mechanism. Not only is the contact poor but the sink slides around on the cpu :| Not at all what I expect a high dollar out of box experience to be.

I contacted Thermalright and all I got was a good talking down to. Basically the guy told me the results were normal, and I wasn't qualified to know what I am talking about. My Q6600 had as much as a 10C delta between cores with the Extreme, re-affirmed with multiple mounts. The Ultra 120, 3C delta between cores.

I asked to upgrade to Ultra 120 and they said sorry, things are normal.....a-holes :|

The Extreme proved to be an expensively piss poor out of box experience. Convex base and clamping mechanism suckage.

The Ultra 120 is still available from SVC and is a hell of a sink for a Quad though.

 

[aigomorla]

Originally posted by: RallyMaster
Why do you have 60 heatsinks?

cuz yoda is an airsink whore.

Like how i am a waterblock whore.

 

[JEDIYoda]

Originally posted by: The Odorous One
Thermalright doesn't care, I know first hand. I had very good luck with an Ultra 120, nice flat base and great clamping mechanism. So, I decided for my second rig to go with a 120 Ultra 120 Extreme. Major disappointment. Although the sink itself was great, the base was convex. To make matters worse, they botched the clamping mechanism. Not only is the contact poor but the sink slides around on the cpu :| Not at all what I expect a high dollar out of box experience to be.

I contacted Thermalright and all I got was a good talking down to. Basically the guy told me the results were normal, and I wasn't qualified to know what I am talking about. My Q6600 had as much as a 10C delta between cores with the Extreme, re-affirmed with multiple mounts. The Ultra 120, 3C delta between cores.

I asked to upgrade to Ultra 120 and they said sorry, things are normal.....a-holes :|

The Extreme proved to be an expensively piss poor out of box experience. Convex base and clamping mechanism suckage.

The Ultra 120 is still available from SVC and is a hell of a sink for a Quad though.

Suffice to say as I stated earlier I have never lapped any of my heatsinks whether i thought it was needed or not!
I have not ever been disappointed with my temps. Of course I have always had good temps and when it comes to heatsinks I would like to think I know abit more than the average joe...except for maybe my friend Zepper and aigomorla or BonzaiDuck.
Every heatsink I own I have used for at least a week or two.
You expect that when you purchase a heatsink that it should be absolutely flat. That is a very unrealistic expectation!
In fact I have found no issues regarding the base of the heatsink unless the middle portion is convex inward.
Then I would definately RMA that heatsink.
But some heatsinks actually work netter if the are mildly convex outward.

No lets address the cllamping mechanism issue.
To use your own words-- " To make matters worse, they botched the clamping mechanism. Not only is the contact poor but the sink slides around on the cpu Not at all what I expect a high dollar out of box experience to be. "

I doubt they botched the clamping mechanism it was most likely user error.....possibly not following directions or for that matter thinking you knew what you were doing without stopping and mentally installing the heatsink before you actually did install it.

Sounds to me like you contacted the good people at ThermalRight with an attitude to begin with.
Why didn`t you just contact the place where you purchased the heatsink from and tell them you are not happy?

Good Luck@!!

 

[PliotronX]

It is surprising that the HR-03 suffers the same intentional base imperfection as the TRUE, but most tests I've seen are only affected within 2 degrees by the base of the heatsink, the shape of the IHS being much more important. G92 variants and your 4870 do not feature the IHS so the temperatures can only suffer more noticeably I would imagine. What is further surprising is the clamping pressure. This has been a rather large issue with the TRUE, but I'm not sure if the penny mod can be applied to the HR-03. My only guess at TR's issue with improving mounting pressure is attempting to avoid shall we say butterfingers who will damage their hardware with too much uneven pressure. All that aside I can understand your frustration, as their official stance is that endusers who complain are wrong and "resistant to change." They are still exceptional heatsinks though.

 

[The Odorous One]

Originally posted by: JEDIYoda

Suffice to say as I stated earlier I have never lapped any of my heatsinks whether i thought it was needed or not!
I have not ever been disappointed with my temps. Of course I have always had good temps and when it comes to heatsinks I would like to think I know abit more than the average joe...except for maybe my friend Zepper and aigomorla.
Every heatsink I own I have used for at least a week or two.
You expect that when you purchase a heatsink that it should be absolutely flat. That is a very unrealistic expectation!

Huh? Why is the Ultra 120 perfectly flat and the Extreme convex? And I am talking many users with a similar situation. Being the smart guy you profess to be, you do know they make flat base sinks.

In fact I have found no issues regarding the base of the heatsink unless the middle portion is convex inward. <---CONCAVE would be the term for that
Then I would definately RMA that heatsink.
But some heatsinks actually work netter if the are mildly convex outward.

Maybe so, but a 10C delta between cores is a bit much huh

No lets address the cllamping mechanism issue.
To use your own words-- " To make matters worse, they botched the clamping mechanism. Not only is the contact poor but the sink slides around on the cpu Not at all what I expect a high dollar out of box experience to be."

I doubt they botched the clamping mechanism it was most likely user error.....possibly not following directions or for that matter thinking you knew what you were doing without stopping and mentally installing the heatsink before you actually did install it.

User error? You have to be kidding right? There is only one way to put it on. A monkey or even you Yoda could do it. Do you really think after multiple installs I did it wrong every time Nice trolling on that one or it could just be your arrogance. :::SIGH:::

Sounds to me like you contacted the good people at ThermalRight with an attitude to begin with.

How in the hell could you possibly know my attitude. I asked with a very professional question and was belittled.

Why didn`t you just contact the place where you purchased the heatsink from and tell them you are not happy?

Pay for their mistake? uh-uh...

I love this precious statement of yours:

"I would like to think I know abit more than the average joe..."

Sadly, I would like to think you do to, but so far not too impressed big fella. You certainly think quite a bit about yourself.

Good Luck@!!

:lips:

 

[aigomorla]

Originally posted by: The Odorous One

The Extreme proved to be an expensively piss poor out of box experience. Convex base and clamping mechanism suckage.
.

LOL!

the IFX-14 had the same problems too.

You'd think there flag ship would be flat?


The HR-05's are not flat as well, i had to lap those as well. :T

What there trying to do is bow the sink to put more pressure on the middle of the IHS and have a better core to ihs contact.

This was proven to be very very good on water blocks, but VERY VERY BAD on air sinks. Why? torque.

You cant get a good amount of = pressure with a sink that large. That is why theres a very bad delta. :T

 

[Elcs]

A heatsink should be relatively flat. In my opinion, no daylight should be visible in a simple "credit card test" and the only "imperfections" on the surface of the sink should be minor manufacturing grooves.

A heatsink which has visible daylight in the credit card test is of unacceptable quality and should be eligible for RMA.

I wasnt expecting to have to but I am willing to lap my HR-03 GT and I know a lot of people here who wouldnt even consider it.

Thermalright always offer heatsinks which are in the top performers of a generation. Ever since my Thermalright SLK900A which I had on my AMD XP 1700+ DLT3C JIUHB, I have enjoyed the benefits of a high-end heatsink. Thermalright have been chosen for my CPU and GPU this time around because they were the best solutions on the marketplace for me and my situation.

I may remove them from the running next time I choose to upgrade my cooling solutions until they can remedy this, what I see as a quality control issue.

 

[JEDIYoda]

Originally posted by: The Odorous One

Originally posted by: JEDIYoda

Suffice to say as I stated earlier I have never lapped any of my heatsinks whether i thought it was needed or not!
I have not ever been disappointed with my temps. Of course I have always had good temps and when it comes to heatsinks I would like to think I know abit more than the average joe...except for maybe my friend Zepper and aigomorla.
Every heatsink I own I have used for at least a week or two.
You expect that when you purchase a heatsink that it should be absolutely flat. That is a very unrealistic expectation!

Huh? Why is the Ultra 120 perfectly flat and the Extreme convex? And I am talking many users with a similar situation. Being the smart guy you profess to be, you do know they make flat base sinks.

In fact I have found no issues regarding the base of the heatsink unless the middle portion is convex inward. <---CONCAVE would be the term for that
Then I would definately RMA that heatsink.
But some heatsinks actually work netter if the are mildly convex outward.

Maybe so, but a 10C delta between cores is a bit much huh

No lets address the cllamping mechanism issue.
To use your own words-- " To make matters worse, they botched the clamping mechanism. Not only is the contact poor but the sink slides around on the cpu Not at all what I expect a high dollar out of box experience to be."

I doubt they botched the clamping mechanism it was most likely user error.....possibly not following directions or for that matter thinking you knew what you were doing without stopping and mentally installing the heatsink before you actually did install it.

User error? You have to be kidding right? There is only one way to put it on. A monkey or even you Yoda could do it. Do you really think after multiple installs I did it wrong every time Nice trolling on that one or it could just be your arrogance. :::SIGH:::

Sounds to me like you contacted the good people at ThermalRight with an attitude to begin with.

How in the hell could you possibly know my attitude. I asked with a very professional question and was belittled.

Why didn`t you just contact the place where you purchased the heatsink from and tell them you are not happy?

Pay for their mistake? uh-uh...

I love this precious statement of yours:

"I would like to think I know abit more than the average joe..."

Sadly, I would like to think you do to, but so far not too impressed big fella. You certainly think quite a bit about yourself.

Good Luck@!!

:lips:

You have a good day!



Peace!!

 

[BonzaiDuck]

To suggest that the ThermalRight U-120-Extreme suffers from "Quality Control" issues shows an understanding of manufacturing processes and process-improvement, but it is just a little off the mark.

If there were a QC "issue," it was in the design process. TR deliberately manufactured the sink so that the base had a convex surface. Several things entered into this decision, including how pressure is distributed and the location of the cores, but it became obvious that lapping the surface to flat, bare copper improved load temperatures.

If they wanted to manufacture the U-120-Extreme with a perfectly flat base, they could do it.

Another drawback in TR "design" is the nickel-plating of the base. And again, this represents a deliberate trade-off choice: According to TR, the nickel-plating "insures" the integrity of the solder or weld joints between the heatpipes and the base. Other logic suggests that it reduces or eliminates corrosion.

But the thing to remember is this: copper has a higher thermal conductivity and lower thermal resistance than does nickel. And just about everyone I've spoken to, including SVC tech-support pertaining to warranties and their custom-lapped U-120-Extreme offering, disagrees that there is any threat to "structural integrity" or risk of corrosion for just lapping the heatsink-base to bare copper.

That's why many obsessive "enthusiasts" (including me) report as much as a 10C improvement in load temperatures with lapping. I thought I noticed a 5C improvement just from lapping the IHS, and another 5C improvement with a lapped heatsink-base -- provided that the lapping effort removes most of the nickel-plating.

If you plan to lap a TR heatsink with a convex base, I suggest that you build a simple C-clamp and wood-block jig to hold it perfectly perpendicular to the sandpaper until you see a flat surface develop on the convex "ridge."

AFTERTHOUGHT EDIT: TR probably voids warranty for lapped heatsink bases because they've made a worst-case assumption that the DIY'er would hold the heatsink by the fins and pipes while lapping. This would certainly put those solder-joints or welds at risk. But just for the record, you don't need to hold the pipes and fins to do the lapping.

 

[The Odorous One]

Originally posted by: aigomorla

Originally posted by: The Odorous One

The Extreme proved to be an expensively piss poor out of box experience. Convex base and clamping mechanism suckage.
.

LOL!

the IFX-14 had the same problems too.

You'd think there flag ship would be flat?


The HR-05's are not flat as well, i had to lap those as well. :T

What there trying to do is bow the sink to put more pressure on the middle of the IHS and have a better core to ihs contact.

This was proven to be very very good on water blocks, but VERY VERY BAD on air sinks. Why? torque.

You cant get a good amount of = pressure with a sink that large. That is why theres a very bad delta. :T

:thumbsup:

 

[The Odorous One]

Originally posted by: BonzaiDuck
To suggest that the ThermalRight U-120-Extreme suffers from "Quality Control" issues shows an understanding of manufacturing processes and process-improvement, but it is just a little off the mark.

If there were a QC "issue," it was in the design process. TR deliberately manufactured the sink so that the base had a convex surface. Several things entered into this decision, including how pressure is distributed and the location of the cores, but it became obvious that lapping the surface to flat, bare copper improved load temperatures.

If they wanted to manufacture the U-120-Extreme with a perfectly flat base, they could do it.

Another drawback in TR "design" is the nickel-plating of the base. And again, this represents a deliberate trade-off choice: According to TR, the nickel-plating "insures" the integrity of the solder or weld joints between the heatpipes and the base. Other logic suggests that it reduces or eliminates corrosion.

But the thing to remember is this: copper has a higher thermal conductivity and lower thermal resistance than does nickel. And just about everyone I've spoken to, including SVC tech-support pertaining to warranties and their custom-lapped U-120-Extreme offering, disagrees that there is any threat to "structural integrity" or risk of corrosion for just lapping the heatsink-base to bare copper.

That's why many obsessive "enthusiasts" (including me) report as much as a 10C improvement in load temperatures with lapping. I thought I noticed a 5C improvement just from lapping the IHS, and another 5C improvement with a lapped heatsink-base -- provided that the lapping effort removes most of the nickel-plating.

If you plan to lap a TR heatsink with a convex base, I suggest that you build a simple C-clamp and wood-block jig to hold it perfectly perpendicular to the sandpaper until you see a flat surface develop on the convex "ridge."

AFTERTHOUGHT EDIT: TR probably voids warranty for lapped heatsink bases because they've made a worst-case assumption that the DIY'er would hold the heatsink by the fins and pipes while lapping. This would certainly put those solder-joints or welds at risk. But just for the record, you don't need to hold the pipes and fins to do the lapping.

I didn't mention QC, it is widely known that TR designed the sink this way. Furthermore, lapping should not be required of a flagship product IMO. Buy it so you can get to work fixing it? i don't think so.

The flagship out-of-box should work better than its' predecessor (Ultra 120 in this case) out-of-box, wouldn't you agree? Adding the extra pipes to the Extreme was a fantastic idea, but forced a clamp re-design which is clearly inferior to the Ultra 120 clamp. The pressure is simply not enough for the convex base, making a potentially killer sink a lame duck. The sink sliding around on the IHS is a major disappointment and the temps very obviously point out the deficiency.

What is mind blowing is that TR doesn't ackknowledge there is a problem. One problem (convex base) is bad enough, inducing a second (under pressured clamp) makes me wonder if they really know what they are doing.

 

[Elcs]

Having lapped both my Ultima 90 and my HR-03 GT, I can safely say both bases were Concave.

During the lapping process, the nickel plating wore off around the edges first and the machine marks were still left in the middle of the sink.

 

[God Mode]

I've lapped a few TR sinks. A 65 year old alcoholic could do a better job with a belt sander than what TR does.

If some of you bought several of the same model TR sinks, you would notice that some are more flat than others. I find that the finish and quality of the older TR sinks (sk6 era +) were better than what they release now.

My pet theory is that they probably use some kind of flash heating method to bond the heatpipes/base/fins together. The rapid heat/cooling phase probably warps it. You could prevent this by finishing the base after all this is done but along with the nickel coating and such, it would be more costly and time consuming.

 

[PliotronX]

Originally posted by: BonzaiDuck
Another drawback in TR "design" is the nickel-plating of the base. And again, this represents a deliberate trade-off choice: According to TR, the nickel-plating "insures" the integrity of the solder or weld joints between the heatpipes and the base. Other logic suggests that it reduces or eliminates corrosion.

I would take their word for it because nickel could very well facilitate welding with non-eutectic solder sweat between the base and pipes. It is difficult to solder bare heatpipes, the heat flux travels very quickly away from the joint. I don't have much experience there, but I would venture to guess that the resulting solder joint aids in performance more than the nickel plating on the base inhibits it. Good post and points. I agree that straight up QC issues would be unintentional. What we possibly have here is some cost cutting going on.

Originally posted by: God Mode
I've lapped a few TR sinks. A 65 year old alcoholic could do a better job with a belt sander than what TR does.[/i]

I can't think of a better way to describe the bases of TR's most recent work. Sigged :beer:

If some of you bought several of the same model TR sinks, you would notice that some are more flat than others. I find that the finish and quality of the older TR sinks (sk6 era +) were better than what they release now.

My pet theory is that they probably use some kind of flash heating method to bond the heatpipes/base/fins together. The rapid heat/cooling phase probably warps it. You could prevent this by finishing the base after all this is done but along with the nickel coating and such, it would be more costly and time consuming.

Yep, my first was the followup to the SK6, the AX-7 still running on a Tualeron. That one did have the nasty machining marks similar to what is representative of TR's latest works, but it was very nearly flat. Lapping didn't make a detectable difference on that one (then on a 700E @ 1GHz).

 

[Tullphan]

Originally posted by: BonzaiDuck


If you plan to lap a TR heatsink with a convex base, I suggest that you build a simple C-clamp and wood-block jig to hold it perfectly perpendicular to the sandpaper until you see a flat surface develop on the convex "ridge."

I'm a little dumb and can understand better with a picture. Got one hooked to a C-clamp?