Requested proof of Thermalright quality control issues

[Goshawk]

I hadn't seen it mentioned, though please forgive me if it has...

What the NotLag could try is removing the top retention plate of the zif socket... removing the possibility the the height / thickness of the bracket from interfering with contact from the bottom of the heatsink to the top of the CPU... i guess it is possible that too much material was removed from the IHS in the lapping process.

I dunno, something else to try At least something like this wouldn't a permanent, warranty voiding experiment



~ Gos

 

[Notlag]

I hadn't seen it mentioned, though please forgive me if it has...

What the NotLag could try is removing the top retention plate of the zif socket... removing the possibility the the height / thickness of the bracket from interfering with contact from the bottom of the heatsink to the top of the CPU... i guess it is possible that too much material was removed from the IHS in the lapping process.

I dunno, something else to try At least something like this wouldn't a permanent, warranty voiding experiment
~ Gos

Thanks for the suggestion, Gos. How would I keep the CPU seated and held securely in place? Purely from the pressure of the heatsink? I would be a bit nervous about that, but don't see why it wouldn't work in theory.

Also, while it would be interesting, I tested my 1st heatsink -before- I lapped it, and the main reason I lapped it was because the 1st heatsink wasn't performing well. The lapping actually reduced temps by a few degrees, so I don't think it hurt. Admittadly that was all before the OP of this thread, which is actually the 2nd heatsink from TR, but yah - the temps were bad before the lapping and the lapping improved them slightly.

 

[Notlag]

Ok, so I've given this a bit of thought and come to the follow decisions:

I'm going to keep my existing CPU. That doesn't mean I won't test the Pentium D to gauge the performance of the U120X (still doesn't change the OP comparison w/ the Maxorb or the difference between the 2nd and 3rd U120X), but I'm not going to replace my E4300. There was some debate as to if the E4xxx's and E2xxx's had a less efficient die / IHS interface than the E6xxx's, but I'm not going to spend close to $200 to get an E6320 that runs at almost the same clock speed as my E4300 and is harder to OC given the lower multiplier simply so that I can benefit from extra cache and lower temps some. The lower temps would be nice, but are unlikely to result in a major additional overclock (ie - I'm at 3GHz now w/ a 1.8GHz chip w/ only 1.325v and I may be able to push it farther still before having to up the voltage - an E6320 won't give me that much more).

I'm also likely going to keep my current U120X. My temps are not great by any stretch, but this part at least is around 4C lower than the previous ones. I could spend another $50 on a Tuniq or other high performance heatsink and end up w/ temps that are little different - maybe I'd shave off a few degrees. That's not going to have a major impact on how far I can OC this chip or its lifespan (and its lifespan is a non-issue - it was a $100 chip and technology is usually obsolete in a few years). The other heatsink is still a consideration, but unlikely at this point (and if I keep the U120X I'm not going to send it for outside testing - using my Pentium D would be good enough for me to provide the same verification). Water cooling would lower temps farther as well, but would cost waaaaay more, and still would only give me a bit of additional overclocking headroom, so that's out for sure.

What I likely will do is buy a higher CFM fan and try to pipe outside air straight to it. That'll shave off a bit of temperature without having to spend a fortune.

 

[nd4spdbh2]

Seriously notlag... cmon stop playing around trying to defend your antics... simply put the ultra 120x and maxorb on the pentium d, or stop posting this nonsence.

 

[Notlag]

this is fawkin pethetic notlag.... STICK THE DAMN ULTRA120X AND MAXORB ON UR PENTIUM D AND TAKE SOME NUMBERS DOWN... untill then stop posting crap.

or take the damn IHS off your e4300... if im not mistaken they dont have their cores soldered to the IHS.

What's "pethetic" is your inability to read and understand my post about why the die / IHS interface doesn't explain the Maxorb tying with the 2nd heatsink.

I can test on the Pentium D to see if my 3rd heatsink is performing reasonably, as it is posting 4C better temps than the first two, but doing so will not change the fact that the first two tied w/ the Maxorb and were not as good as the third. The things that can explain that are:

1. The 3rd TR U120X has a solid bracket, which could have improved its interface with the IHS (versus the 1st two heatsinks w/ the scissors) and hence lowered temps. However, that is unlikely, as I have tried increasing mounting pressure on the first two heatsinks w/ washers on the spring screws and that did not appreciably lower temps.

2. TR's manufacturing has been inconsistant, and that has resulted in the difference between the first set of heatsinks and the 3rd heatsink.

3. The inconsistant quality has led to heatsinks w/ mediocre performance, including the first two, which were barely capable of topping the Maxorb.

THE CHIP DOES NOT EXPLAIN EITHER THE MAXORB VS U120X COMPARISON OR THE FACT THAT THE 3RD HEATSINK WAS NOTICABLY BETTER THAN THE FIRST TWO!

 

[nd4spdbh2]

What's "pethetic" is your inability to read and understand my post about why the die / IHS interface doesn't explain the Maxorb tying with the 2nd heatsink.

I can test on the Pentium D to see if my 3rd heatsink is performing reasonably, as it is posting 4C better temps than the first two, but doing so will not change the fact that the first two tied w/ the Maxorb and were not as good as the third. The things that can explain that are:

1. The 3rd TR U120X has a solid bracket, which could have improved its interface with the IHS (versus the 1st two heatsinks w/ the scissors) and hence lowered temps. However, that is unlikely, as I have tried increasing mounting pressure on the first two heatsinks w/ washers on the spring screws and that did not appreciably lower temps.

2. TR's manufacturing has been inconsistant, and that has resulted in the difference between the first set of heatsinks and the 3rd heatsink.

3. The inconsistant quality has led to heatsinks w/ mediocre performance, including the first two, which were barely capable of topping the Maxorb.

THE CHIP DOES NOT EXPLAIN EITHER THE MAXORB VS U120X COMPARISON OR THE FACT THAT THE 3RD HEATSINK WAS NOTICABLY BETTER THAN THE FIRST TWO!

OK you kno what... forget ANYTHING AND EVERYTHING about the e4300 forget EVERYTHING ABOUT THIS THREAD... START COMPLETLY OVER and put the ultra 120x and maxorb on your pentium d. and how do you consider noticibly better with 4*c... you could EASILY have a 4*c variance by just mounting the hs differently...

 

[Notlag]

Seriously notlag... cmon stop playing around trying to defend your antics... simply put the ultra 120x and maxorb on the pentium d, or stop posting this nonsence.

I've already jumped through some hoops to try and prove the point to you (and others who can't accept issues w/ TR). At this point, whether or not my CPU is hot is irrelevant. I've already shown there were problems w/ the first two heatsinks. I've already explained why the CPU isn't a factor in those tests. I don't need to prove it any further to myself, and when it boils down to it, there is some small clique, of which you are part, that will not accept any amount of evidence, so I'm not that concerned about continuing to try and prove it to that group. The only benefit to testing the Pentium D is to see if my E4300 has issues in addition to the issues w/ the first couple TR's, and that's only small benefit to me because I'm not planning on replacing my E4300 regardless.

 

[nd4spdbh2]

I've already jumped through some hoops to try and prove the point to you (and others who can't accept issues w/ TR). At this point, whether or not my CPU is hot is irrelevant. I've already shown there were problems w/ the first two heatsinks. I've already explained why the CPU isn't a factor in those tests. I don't need to prove it any further to myself, and when it boils down to it, there is some small clique, of which you are part, that will not accept any amount of evidence, so I'm not that concerned about continuing to try and prove it to that group. The only benefit to testing the Pentium D is to see if my E4300 has issues in addition to the issues w/ the first couple TR's, and that's only small benefit to me because I'm not planning on replacing my E4300 regardless.

HOW DO YOU KNO if the heatsink is the problem if you have not throughly tested them on different chips?????

 

[Notlag]

you could EASILY have a 4*c variance by just mounting the hs differently...

I guess I work more consistently than you do. I've mounted and re-mounted my first two TR's a good half dozen to dozen times each and encountered nowhere near that much variance.

 

[Notlag]

HOW DO YOU KNO if the heatsink is the problem if you have not throughly tested them on different chips?????

Read the earlier posts and try to follow the logic. I've explained it all quite clearly. The OP and my post about the die / IHS interface w/ the pretty ASCII diagram covers everything you need to know. Somewhere there's also a post w/ a nice summary of the test results and why the CPU doesn't explain them.

 

[nd4spdbh2]

Read the earlier posts and try to follow the logic. I've explained it all quite clearly. The OP and my post about the die / IHS interface w/ the pretty ASCII diagram covers everything you need to know.

ok ... serious, if the interface btween an ihs and core can only trasmit so much heat, it doesnt matter what hs you put on top of it THEY ALL ARE GOING TO SUCK..

just PLEASE PLEASE oh PLEASE take the 10 min to put the ultra 120x on the pentium D for me...

 

[Notlag]

ok ... serious, if the interface btween an ihs and core can only trasmit so much heat, it doesnt matter what hs you put on top of it THEY ALL ARE GOING TO SUCK..

just PLEASE PLEASE oh PLEASE take the 10 min to put the ultra 120x on the pentium D for me...

But they are not going to suck equally... Read the post w/ the ASCII art. I'm not full of crap. Heat transfers much like electricity. Everything it passes through has a certain amount of thermal resistance, and those resistances work very similarly to the way resistors work w/ electricity. I'm not going to explain it any further in this thread unless other people want to see it too. I may post a guide or how-to on it as a seperate thread.

I'm getting a little sick of tinkering w/ this PC. I have had to futz w/ this for like a month now. I just want to get the thing finished. I still have to lap the heatsink, figure out my max OC that I'm comfortable with, install a higher CFM fan, cut the side panel window, install a side inlet fan in the side panel window, possibly make some duct work, etc... I don't really feel like dragging it out much farther and doing the Pentium D doesn't prove anything useful about the TR heatsink. What will end up happening is I will test w/ the Pentium D, then somebody will argue some other point and ask me to test something else, then someone will say send the heatsink off to someone, then someone will say buy a different CPU, etc... And in the end all it does is drag things out even farther to prove something I have already given sufficient proof on.

Consider this. If there was 4C variation going from the 1st and 2nd heatsink to the 3rd, then that actually says Thermalright's QC is even WORSE if my temps are high because of the CPU. A 4C variation with deltas around 35C from ambient to core is around 11%. If my high temps were from the CPU being bad then my deltas would be lower, let's imagine around 25C instead (with a "working" CPU), but the variation between the heatsinks would still be 4C based on the difference in thermal performance of the heatsink and the fact that the same amount of thermal power still has to be dissipated. That would be more like 16% instead of 11%. That means the spread in quality between parts is even wider if indeed my CPU is the cause of the overall high temps. And again, the CPU also can't explain the Maxorb tying the first two heatsinks.

 

[Automata]

But they are not going to suck equally... Read the post w/ the ASCII art. I'm not full of crap. Heat transfers much like electricity. Everything it passes through has a certain amount of thermal resistance, and those resistances work very similarly to the way resistors work w/ electricity. I'm not going to explain it any further in this thread unless other people want to see it too. I may post a guide or how-to on it as a seperate thread.

Well, what about this.

The IHS's on the e2xxx series are not soldered on, they just use a paste. The first two heatsinks had the scissor bracket that supposedly does not have as much pressure as the solid one. The third one had the hard bracket and got lower temps. You showed pics of the max orb mounting (really strange ) and it looks like it has alot of pressure too.

What if the first two heatsinks you had, performed like they did because it wasn't putting enough pressure on the IHS? If pressure was the issue, then that explains why the third one did better with the solid bracket and why the maxorb is keeping up

I'd say try it on a different processor if you can.

 

[nd4spdbh2]

But they are not going to suck equally... Read the post w/ the ASCII art. I'm not full of crap. Heat transfers much like electricity. Everything it passes through has a certain amount of thermal resistance, and those resistances work very similarly to the way resistors work w/ electricity. I'm not going to explain it any further in this thread unless other people want to see it too. I may post a guide or how-to on it as a seperate thread.

I'm getting a little sick of tinkering w/ this PC. I have had to futz w/ this for like a month now. I just want to get the thing finished. I still have to lap the heatsink, figure out my max OC that I'm comfortable with, install a higher CFM fan, cut the side panel window, install a side inlet fan in the side panel window, possibly make some duct work, etc... I don't really feel like dragging it out much farther and doing the Pentium D doesn't prove anything useful about the TR heatsink. What will end up happening is I will test w/ the Pentium D, then somebody will argue some other point and ask me to test something else, then someone will say send the heatsink off to someone, then someone will say buy a different CPU, etc... And in the end all it does is drag things out even farther to prove something I have already given sufficient proof on.

Consider this. If there was 4C variation going from the 1st and 2nd heatsink to the 3rd, then that actually says Thermalright's QC is even WORSE if my temps are high because of the CPU. A 4C variation with deltas around 35C from ambient to core is around 11%. If my high temps were from the CPU then my deltas might be 25C (or less) instead, but the variation would still be 4C based on the difference in thermal performance of the heatsink, and that would be more like 16%. That means the spread in quality between parts was even wider. And again, the CPU also can't explain the Maxorb tying the first two heatsinks.

HOW do you kno that they will all suck on the pentium d... YOU HAVNT EVEN TRIED IT... for goodness sake i swear your trying all you can to badmouth thermalright.

 

[Automata]

HOW do you kno that they will all suck on the pentium d... YOU HAVNT EVEN TRIED IT... for goodness sake i swear your trying all you can to badmouth thermalright.

You might want to stop yelling

He gets your point, it is just the way your transmitting it. Which is....to be honest....pretty rude.

 

[nd4spdbh2]

You might want to stop yelling

He gets your point, it is just the way your transmitting it. Which is....to be honest....pretty rude.

well... after 12 pages he hasnt listened to anyone... i guess its really a lost cause.

 

[Automata]

well... after 12 pages he hasnt listened to anyone... i guess its really a lost cause.

12???

It is on page 4....

Looking forward to test results if he agrees to do it

 

[Notlag]

Well, what about this.

The IHS's on the e2xxx series are not soldered on, they just use a paste. The first two heatsinks had the scissor bracket that supposedly does not have as much pressure as the solid one. The third one had the hard bracket and got lower temps. You showed pics of the max orb mounting (really strange ) and it looks like it has alot of pressure too.

What if the first two heatsinks you had, performed like they did because it wasn't putting enough pressure on the IHS? If pressure was the issue, then that explains why the third one did better with the solid bracket and why the maxorb is keeping up

I'd say try it on a different processor if you can.

Well, the other proc won't change the mounting variation though. It's too bad I didn't keep one of the scissor brackets I never thought TR would send me a heatsink w/ the old solid bracket. A few quick points though:

1. I tried using washer w/ the old scissor brackets to increase mounting pressure and that didn't change temps. Admittadly, I only used one flat washer (approx 1mm thick?), so that may have been a minor change.

2. The Maxorb mounting pressure is actually not created by the frame. The whole point of that frame is to just stick to the mainboard and not move. Tension is applied by a spring-like lever that latches onto the frame. As long as the frame doesn't move and is held at the proper height then it doesn't matter whether it is bolted on, clipped on, etc... I wish I could explain it better w/o you actually owning a Maxorb. Maybe if I get some pics of the Maxorb's tension clip?

3. The reason the heatsink doesn't rotate in the solid bracket has nothing to do w/ the mounting pressure. The heatsink can't be rotated because the bracket wraps over the sides of the heatsink and prevents rotation. However, that does not mean the solid bracket isn't resulting in increased mounting pressure for other reasons. It could very well be, and that could contribute to the 4C better performance (though I think the issue is perhaps w/ the base - the 3rd heatsink had a WAY better base than the 2nd).

4. If the difference between the 3rd heatsink and the 1st / 2nd is because of the bracket, then it just means that the 3rd is also no better than the Maxorb had it been mounted w/ the scissor clip. That, imho, is just as bad for TR as the 3rd heatsink being better. Either we have fairly large variation in TR's manufacturing resulting in fairly large variations in performance, or we have TR producing many defective heatsinks lately that are barely better than Maxorbs.

5. The CPU also can't explain the Maxorb comparison w/ the first two heatsinks. The only thing that can explain that would be if the Maxorb's slightly modified mounting did result in greater mounting pressure for it, but I explained that as best as possible in point 2 I'll need pics to explain it better.

So... The only thing I can think of out of all that which might explain it other than the TR quality itself... Maybe if the frame for the Maxorb was held marginally closer to the board using the screws than it would have been w/ the plastic push-pins (the push pins have a little spacer that raises them a certain height from the board). The difference though is certainly no more than the same height as the washers I used w/ the TR's to try to increase mounting pressure, and that made no difference. And.... even if that were the case, it doesn't explain other people who have posted complaining about their issues w/ TR (there've been a few other posters in this thread who've said their latest purchases from TR haven't been too great).

The only reason for me to try the other proc is to get certain people (not you) to be quiet, but in the end I know those people will latch onto something else, so I'm not too worried about it. I'd end up having to test iteration after iteration of stuff and eventually they wouldn't come to the same conclusions no matter the results, they'd just stop posting and let it drop, unable to determine what I'm doing wrong.

 

[Notlag]

12???

It is on page 4....

Looking forward to test results if he agrees to do it

I'll consider it, but I'm really getting pretty much through w/ wanting to futz with this anymore. My CPU isn't cooking, I've presented my evidence, and I don't think my computer is going to melt-down anytime soon. Can you blame me for wanting to be done w/ this?

Maybe while I'm doing the other stuff I'll have a good opportunity to run some tests, but for nd4spdbh2 to say I can do all the swaps, run all the trials, screen cap everything, etc... in 10 minutes is a bit absurd. I know that as I do that it's just going to open more questions (since it is introducing more variables) that I'm going to have to do other stuff to disprove, eliminate, etc... Plus unless I can find time to do it all in one sitting then I have to put my E4300 in my other PC or leave it down for a while (which stinks because I have people over a lot that use that PC when we're gaming). We'll see - I'll play it a bit by ear, but odds are I won't end up doing the swap and test unless I end up buying a different heatsink, which is becoming less likely.

 

[Automata]

Well, the other proc won't change the mounting variation though. It's too bad I didn't keep one of the scissor brackets I never thought TR would send me a heatsink w/ the old solid bracket. A few quick points though:

1. I tried using washer w/ the old scissor brackets to increase mounting pressure and that didn't change temps. Admittadly, I only used one flat washer (approx 1mm thick?), so that may have been a minor change.

2. The Maxorb mounting pressure is actually not created by the frame. The whole point of that frame is to just stick to the mainboard and not move. Tension is applied by a spring-like lever that latches onto the frame. As long as the frame doesn't move and is held at the proper height then it doesn't matter whether it is bolted on, clipped on, etc... I wish I could explain it better w/o you actually owning a Maxorb. Maybe if I get some pics of the Maxorb's tension clip?

3. The reason the heatsink doesn't rotate in the solid bracket has nothing to do w/ the mounting pressure. The heatsink can't be rotated because the bracket wraps over the sides of the heatsink and prevents rotation. However, that does not mean the solid bracket isn't resulting in increased mounting pressure for other reasons. It could very well be, and that could contribute to the 4C better performance (though I think the issue is perhaps w/ the base - the 3rd heatsink had a WAY better base than the 2nd).

4. If the difference between the 3rd heatsink and the 1st / 2nd is because of the bracket, then it just means that the 3rd is also no better than the Maxorb had it been mounted w/ the scissor clip. That, imho, is just as bad for TR as the 3rd heatsink being better. Either we have fairly large variation in TR's manufacturing resulting in fairly large variations in performance, or we have TR producing many defective heatsinks lately that are barely better than Maxorbs.

5. The CPU also can't explain the Maxorb comparison w/ the first two heatsinks. The only thing that can explain that would be if the Maxorb's slightly modified mounting did result in greater mounting pressure for it, but I explained that as best as possible in point 2 I'll need pics to explain it better.

So... The only thing I can think of out of all that which might explain it other than the TR quality itself... Maybe if the frame for the Maxorb was held marginally closer to the board using the screws than it would have been w/ the plastic push-pins (the push pins have a little spacer that raises them a certain height from the board). The difference though is certainly no more than the same height as the washers I used w/ the TR's to try to increase mounting pressure, and that made no difference. And.... even if that were the case, it doesn't explain other people who have posted complaining about their issues w/ TR (there've been a few other posters in this thread who've said their latest purchases from TR haven't been too great).

The only reason for me to try the other proc is to get certain people (not you) to be quiet, but in the end I know those people will latch onto something else, so I'm not too worried about it. I'd end up having to test iteration after iteration of stuff and eventually they wouldn't come to the same conclusions no matter the results, they'd just stop posting and let it drop, unable to determine what I'm doing wrong.

I agree with alot of things you said and see where you are coming from on the others. It is just hard to compare my heatsink to yours since we are running different processors and other variables could be throwing it off; meaning, my temps would mean nothing to you. Looking at it from your point of view, I can see why you are drawing these conclusions, and instead of testing this further (which may/may not produce different results), we could focus on ways to improve the performance (like washers).

1. That may be the cause, I need to add washers to mine as my processor is lapped and I can move my heatsink fairly easily (scissor bracket). I'll let you know if my temps improve at all when I do so (I'll pm you).

2. I'll google it

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