Requested proof of TR quality control problems - TR U120X vs TT Maxorb
EDIT: Please read my disclaimer at the end - this is not a recommendation for a specific heatsink, but a discussion on apparent Thermalright quality control problems.
I posted previously about major quality control issues at Thermalright, warning people to be very wary of their products until the issues were addressed. Despite being very thorough with my testing (even validating my methodology w/ the stock heatsink), I was flamed for (heaven forbid) saying that Thermalright has been turning out some pretty poorly constructed parts lately. Many people also requested proof in the form of screen caps and asked me to test my TR U120X against another cooler at overclocked speeds, and there were some people of the opinion that I should post absolute temps rather than calculating the thermal resistance of the solution.
People have insisted they've had great TR U120X's because it has reduced their temperature by a considerable margin versus the stock HSF - but has it reduced temps for them as far as it should? Have most people went to any length to validate that their TR U120X performs the way it should? When most people have warmer temperatures than they should, they don't think the heatsink is defective, they lap the base, try to improve their airflow, accept the improvement those two things offer, and figure any other deficiencies are the result of something else in their system.
So... I purchased a Thermaltake MaxOrb. It's not a bad cooler, but common sense tells us it shouldn't cool anywhere near as well as the TR U120X - it has fewer heatpipes, has less surface area, weighs much less, and is not a tower cooler. Besides, we all "know" it should perform nowhere close to the TR U120X because the many reviews tell us so (which is precisely why I selected it - I estimated it would perform as well as my defective U120X's even though it should perform much worse than a working one). Just take a look at AnandTech:
http://www.anandtech.com/showdoc.aspx?i=3001&p=6
Before people start running on about how reviews don't tell the story and how manufacturers send the review sites "golden samples" that are picked from the production runs because they have better than normal performance, remember that the same thing can be said for Thermaltake. The above review was performed using a standard test platform, performed by the same person, and Thermaltake should be just as likely to send them a "golden sample" for review as Thermalright. While it is arguable that ther reviewers would receive better performance from their parts than we will from retail parts, if the two companies manufacture their retail parts with similar quality controls, then the retail TR U120X should generally outperform the retail TT Maxorb by a similar margin.
Unfortunately, that's not the way it worked - with either TR U120X I received. When tested side by side, they performed nearly identically. I will not argue that the U120X is by far the better design, but it is clearly not manufactured with the same attention to quality. Even the best engineering can be crippled by poor manufacturing.
Don't take my word for it though, here are the test results:
Test Platform
-------------
E4300 @ 9 x 333 MHz, 1.325v set in BIOS
Gigabyte GA-P35-DS3P
Antec P180B case w/ side removed
Antec Tri-cool fans throughout
Rear case fan set to low
Top, bottom, and front case fans set to high
Room temperature was consistently 75F throughout testing
TAT was used to load the cores (TAT provides the highest possible load on the cores - more so than Orthos - while some may argue that it is unrealistically high load, it is fair to compare load temps between heat sinks provided the same application is used to create the load, and many others would argue TAT is the better tool for testing cooling than Orthos anyways)
Summary of results
------------------
Thermalright U120X Idle: 27C (Speedfan CPU temp) / 39C-40C (TAT core temp)
Thermalright U120X 100% Load: 54C (Speedfan CPU temp) / 66C (TAT core temp)
Thermaltake Maxorb Idle: 27C (Speedfan CPU temp) / 39C-40C (TAT core temp)
Thermaltake Maxorb 100% Load: 54C (Speedfan CPU temp) / 67C (TAT core temp)
Thermalright Ultra 120 eXtreme
------------------------------
Thermalright Ultra 120 eXtreme (as mounted in my PC)
The heatsink fan is an Antec Tri-cool fan at high speed (79CFM). The Antec Tri-cool series fans have been well reviewed on SPCR:
http://www.silentpcreview.com/article695-page3.html#tricool
People have criticized the horizontal mounting because they say the video card obstructs air flow, but if you look at the picture, there's around 3" of clearance between the video card and the fan inlet. In addition, the top of the video card is well below the top of the fan, so the upper portion of the fan is completely unobstructed. If I mounted the heatsink vertically then the optical drives and cables would provide as much obstruction as the video card.
And the screenshots of the test results:
Thermaltake Maxorb
-------------------
Thermaltake Maxorb (as mounted in my PC)
Air flow may be somewhat less restricted than the Thermalright, and the air is not coming from near the northbridge and video card, so it may be slightly cooler. I verified that by measuring the fan inlet temps on both heatsinks, and indeed the Maxorb's fan inlet air temperature was about 1.5C lower than the Thermalright U120X. However, any test of the heatsink installed in a real PC is likely going to have similar issues. Vertical orientation might reduce the difference a bit, but even if it made up the entire difference, 1.5C doesn't impact the results of these tests enough to invalidate them. At the clock speed of the processor as tested, the Maxorb should have run around 8C to 12C hotter than the Thermalright U120X (based on Anandtech and other reviews).
It didn't though. In fact, it ran nearly the same exact temperature. The CPU package temp was the same while the cores were only 1C hotter:
Other Things Worth Noting
-------------------------
Bear in mind this is not an isolated incident. This is the 2nd TR U120X in a row that has performed this way. There've also been quite a few other posts lately by people asking if their warm U120X temps are normal. Nevermind that the bases have also had really bad quality:
That said, for those with defective heat sinks, if the defect is with your heat pipes, lapping the heat sink will NOT fix the problem. I should know - I lapped my 1st heat sink, and I had to fight like hell to get it replaced when it became obvious that there was more wrong than just the base:
Yes, the lapping of the 1st heat sink improved temperatures by 3C or so versus what is in this review, but you can safely figure the review sites did not lap their heat sinks, nor should someone have to lap the heat sink just to get the performance of it closer to what it should be without lapping. Even counting that 3C improvement (some of which resulted from the CPU lapping), it isn't as good as the heatsink is "supposed" to perform. If you want to lap your TR U120X, go ahead - but wait until you know you have nothing else wrong with it.
What was Thermalright's response to all this when I first brought it to them (before I argued like hell on the 2nd heat sink)? Read it for yourself:
So, now that I've provided the proof I was asked for, hopefully this thread won't be deleted, and maybe at some point Thermalright will take note of the fact that they can't swing crappy quality control passed us forever before we stop buying their produts. I'm currently waiting on my 3rd heat sink from them. Maybe they'll get it right on the 3rd try. If it works I'll post the working heat sink results. But until they can state that they've taken steps to improve their manufacturing controls, I'll be reluctant to buy any other products from them. While there are a lot of people here that have been happy with their Thermalright heat sinks, somebody purchasing from them today runs the same chance I did in ending up w/ something that looks fine but works poorly. If I had to do it all over again, I'd have probably water cooled (which would've ran $300+ for the performance I wanted) or went with some other model. Even the mediocre Maxorb I just tested performed nearly as well as the U120X.
Disclaimer
For the record, this is not a recommendation for a TT Maxorb or any other specific cooler. The fact is, Thermalright's engineering is really top-notch and the TR U120X is, at the time of this post, by far the best heatsink design on the market. However, the best engineering can still be crippled by bad manufacturing, and I'm not the first person to get a dud from them and complain about it. Two in a row is pretty bad though, and their customer service's initial response was awful (their over-seas support at least - their US support has tried to be helpful, but has at times had their hands tied by over-seas).
Don't buy a TT Maxorb thinking you'll get TR U120X like performance - they're in an entirely different class and a working TR U120X will blow a TT Maxorb away. I compared the two to show just how bad a defective TR U120X could get (and prove to the doubters that both of mine were indeed defective). If you have qualms about purchasing from Thermalright (I do now), then consider some other high performance tower heatpipe coolers like the Tuniq Tower, Scythe Ninja Plus B, Noctua NH-U12, Zalman CNPS9700, or 3RSystems Iceage 120. Water cooling is another option, but plan on spending $300 to beat the above coolers.
EDIT: Please read my disclaimer at the end - this is not a recommendation for a specific heatsink, but a discussion on apparent Thermalright quality control problems.
I posted previously about major quality control issues at Thermalright, warning people to be very wary of their products until the issues were addressed. Despite being very thorough with my testing (even validating my methodology w/ the stock heatsink), I was flamed for (heaven forbid) saying that Thermalright has been turning out some pretty poorly constructed parts lately. Many people also requested proof in the form of screen caps and asked me to test my TR U120X against another cooler at overclocked speeds, and there were some people of the opinion that I should post absolute temps rather than calculating the thermal resistance of the solution.
People have insisted they've had great TR U120X's because it has reduced their temperature by a considerable margin versus the stock HSF - but has it reduced temps for them as far as it should? Have most people went to any length to validate that their TR U120X performs the way it should? When most people have warmer temperatures than they should, they don't think the heatsink is defective, they lap the base, try to improve their airflow, accept the improvement those two things offer, and figure any other deficiencies are the result of something else in their system.
So... I purchased a Thermaltake MaxOrb. It's not a bad cooler, but common sense tells us it shouldn't cool anywhere near as well as the TR U120X - it has fewer heatpipes, has less surface area, weighs much less, and is not a tower cooler. Besides, we all "know" it should perform nowhere close to the TR U120X because the many reviews tell us so (which is precisely why I selected it - I estimated it would perform as well as my defective U120X's even though it should perform much worse than a working one). Just take a look at AnandTech:
http://www.anandtech.com/showdoc.aspx?i=3001&p=6
Before people start running on about how reviews don't tell the story and how manufacturers send the review sites "golden samples" that are picked from the production runs because they have better than normal performance, remember that the same thing can be said for Thermaltake. The above review was performed using a standard test platform, performed by the same person, and Thermaltake should be just as likely to send them a "golden sample" for review as Thermalright. While it is arguable that ther reviewers would receive better performance from their parts than we will from retail parts, if the two companies manufacture their retail parts with similar quality controls, then the retail TR U120X should generally outperform the retail TT Maxorb by a similar margin.
Unfortunately, that's not the way it worked - with either TR U120X I received. When tested side by side, they performed nearly identically. I will not argue that the U120X is by far the better design, but it is clearly not manufactured with the same attention to quality. Even the best engineering can be crippled by poor manufacturing.
Don't take my word for it though, here are the test results:
Test Platform
-------------
E4300 @ 9 x 333 MHz, 1.325v set in BIOS
Gigabyte GA-P35-DS3P
Antec P180B case w/ side removed
Antec Tri-cool fans throughout
Rear case fan set to low
Top, bottom, and front case fans set to high
Room temperature was consistently 75F throughout testing
TAT was used to load the cores (TAT provides the highest possible load on the cores - more so than Orthos - while some may argue that it is unrealistically high load, it is fair to compare load temps between heat sinks provided the same application is used to create the load, and many others would argue TAT is the better tool for testing cooling than Orthos anyways)
Summary of results
------------------
Thermalright U120X Idle: 27C (Speedfan CPU temp) / 39C-40C (TAT core temp)
Thermalright U120X 100% Load: 54C (Speedfan CPU temp) / 66C (TAT core temp)
Thermaltake Maxorb Idle: 27C (Speedfan CPU temp) / 39C-40C (TAT core temp)
Thermaltake Maxorb 100% Load: 54C (Speedfan CPU temp) / 67C (TAT core temp)
Thermalright Ultra 120 eXtreme
------------------------------
Thermalright Ultra 120 eXtreme (as mounted in my PC)
The heatsink fan is an Antec Tri-cool fan at high speed (79CFM). The Antec Tri-cool series fans have been well reviewed on SPCR:
http://www.silentpcreview.com/article695-page3.html#tricool
People have criticized the horizontal mounting because they say the video card obstructs air flow, but if you look at the picture, there's around 3" of clearance between the video card and the fan inlet. In addition, the top of the video card is well below the top of the fan, so the upper portion of the fan is completely unobstructed. If I mounted the heatsink vertically then the optical drives and cables would provide as much obstruction as the video card.
And the screenshots of the test results:
Thermaltake Maxorb
-------------------
Thermaltake Maxorb (as mounted in my PC)
Air flow may be somewhat less restricted than the Thermalright, and the air is not coming from near the northbridge and video card, so it may be slightly cooler. I verified that by measuring the fan inlet temps on both heatsinks, and indeed the Maxorb's fan inlet air temperature was about 1.5C lower than the Thermalright U120X. However, any test of the heatsink installed in a real PC is likely going to have similar issues. Vertical orientation might reduce the difference a bit, but even if it made up the entire difference, 1.5C doesn't impact the results of these tests enough to invalidate them. At the clock speed of the processor as tested, the Maxorb should have run around 8C to 12C hotter than the Thermalright U120X (based on Anandtech and other reviews).
It didn't though. In fact, it ran nearly the same exact temperature. The CPU package temp was the same while the cores were only 1C hotter:
Other Things Worth Noting
-------------------------
Bear in mind this is not an isolated incident. This is the 2nd TR U120X in a row that has performed this way. There've also been quite a few other posts lately by people asking if their warm U120X temps are normal. Nevermind that the bases have also had really bad quality:
That said, for those with defective heat sinks, if the defect is with your heat pipes, lapping the heat sink will NOT fix the problem. I should know - I lapped my 1st heat sink, and I had to fight like hell to get it replaced when it became obvious that there was more wrong than just the base:
Yes, the lapping of the 1st heat sink improved temperatures by 3C or so versus what is in this review, but you can safely figure the review sites did not lap their heat sinks, nor should someone have to lap the heat sink just to get the performance of it closer to what it should be without lapping. Even counting that 3C improvement (some of which resulted from the CPU lapping), it isn't as good as the heatsink is "supposed" to perform. If you want to lap your TR U120X, go ahead - but wait until you know you have nothing else wrong with it.
What was Thermalright's response to all this when I first brought it to them (before I argued like hell on the 2nd heat sink)? Read it for yourself:
So, now that I've provided the proof I was asked for, hopefully this thread won't be deleted, and maybe at some point Thermalright will take note of the fact that they can't swing crappy quality control passed us forever before we stop buying their produts. I'm currently waiting on my 3rd heat sink from them. Maybe they'll get it right on the 3rd try. If it works I'll post the working heat sink results. But until they can state that they've taken steps to improve their manufacturing controls, I'll be reluctant to buy any other products from them. While there are a lot of people here that have been happy with their Thermalright heat sinks, somebody purchasing from them today runs the same chance I did in ending up w/ something that looks fine but works poorly. If I had to do it all over again, I'd have probably water cooled (which would've ran $300+ for the performance I wanted) or went with some other model. Even the mediocre Maxorb I just tested performed nearly as well as the U120X.
Disclaimer
For the record, this is not a recommendation for a TT Maxorb or any other specific cooler. The fact is, Thermalright's engineering is really top-notch and the TR U120X is, at the time of this post, by far the best heatsink design on the market. However, the best engineering can still be crippled by bad manufacturing, and I'm not the first person to get a dud from them and complain about it. Two in a row is pretty bad though, and their customer service's initial response was awful (their over-seas support at least - their US support has tried to be helpful, but has at times had their hands tied by over-seas).
Don't buy a TT Maxorb thinking you'll get TR U120X like performance - they're in an entirely different class and a working TR U120X will blow a TT Maxorb away. I compared the two to show just how bad a defective TR U120X could get (and prove to the doubters that both of mine were indeed defective). If you have qualms about purchasing from Thermalright (I do now), then consider some other high performance tower heatpipe coolers like the Tuniq Tower, Scythe Ninja Plus B, Noctua NH-U12, Zalman CNPS9700, or 3RSystems Iceage 120. Water cooling is another option, but plan on spending $300 to beat the above coolers.
Last edited:
