Cooler testing criteria discussion

[doyll49]

I started hi-jacking another thread and Micrornd replied, so instead of continuing I started this thread.

My original post

Silver Arrow (all varients) are among the best coolers made. There are minor differences in cooling between original Silver Arrow and newer SB-E & IB-E variants Original cooled slightly better at low rpm and not quite as good at full speed. I did not find any difference between SB-E and IB-E. Extreme versions have same cooling and noise figures at same rpm as normal versions, but latch the hatches before running them on up to their 2500rpm limits. The most huge amounts of air and are also quite loud at full speed.

NH-D14 is similar as is NH-D15. Main Reason NH-D15 cools a little better is it's fans run 2-300rpm faster. s for the Silver Arrow fans, do a side by side comparison of NH-A14 and TY-140, 143 and 147 fans. They are near identical except for color and housing shape. Performance is also near identical in pressure, cfm and noise levels when running at same rpm.

But to the issue of H110 vs Silver Arrow, cooling is similar but H110 is louder at same temps

Reason review tests vary so radically is most use room ambient for baseline temerature instead of the cooler / radiator intake air temp. While room temp has an effect on cooler intake temp, cooler intake has many other variable affecting it and is always warmer than room .. and can easily be 10, 15 or even 20c warmer under full load. Of course the CPU will also run similar amounts hotter, but the room ambient will not change. End results are we see this huge 20c variance in results between reviews even when using same CPU & mobo.

Honestly, using room ambient, especially when testing in a case, is like going into the kitchen to look at thermometer to know how warm your lounge is. Which bring me to all this hype about "real world enviroment" when testing in a case and using room ambient. They are not testing cooler performance. They are testing their systems' performance with different coolers. And the only "real world environment" is their system in their room .. unless you happen to have the exact same system in the same room. I know I don't.

Micrornd's reply

I understand why that makes sense from your point of view and although I agree to a point, I would say that may be a little too narrow.

In my particular case, room ambient temp is the same as cooler intake temps, as my rads are the sole intake for the case.
In my scenario how the air flows through and exits the case is the main variable that affects the cooling.
How the air flows through and exits the case also is a great factor in your scenarios, as well as what other heat producing sources are in the case.

I've seen "open air" testing for coolers proposed also, but to me, that's a much worse solution than testing in a case, as it has way too many uncontrolled variables, as to airflow restrictions and deviations from what would be a "norm" in a case. (i.e. cables, other cards, other heat sources, other case fans, etc.)

Then you also have to factor in how the cooler is made/designed and mounted compared to a OEM cooler (what direction the air flows through compared to the OEM cooler i.e. up, down, left, right, etc.).
Also as to how it deflects air to the MB and VRMs for cooling them (as almost all OEM coolers are designed to do).
That seems to never be discussed or compensated for with "non-OEM" cooler testing.

There are just so many variables that factor in beyond inlet temp, that there is no really good way to test coolers.

So maybe just being aware of all the variables that enter into the cooling scenario is the most important factor of all, rather than just out of context temp and noise figures.

What do you think?

Not quite sure where to start.

How about with what you said about my post "I understand why that makes sense from your point of view and although I agree to a point, I would say that may be a little too narrow."

You start out saying your system's coolers (radiators) are room temperature, but using that as an example is an even narrower interpretation of what I am talking about. And unless every component in your case is cooled by your loop they will be warmer than they would be on room temperature air.

I do agree, case airflow is critical. I have posted much on the subject and how to go about getting the best, and I will add here that the key is supplying component with cool intake air without their heated exhaust contaminating and warming it up.

Testing with system in a case or on an open bench really makes little or no difference. The key is monitoring the temperature of air going into cooler, not the room .. to remove all temperature variables from the equation. Other variables are still involved (barometric pressure, humidity, etc) but they are always involved, be it inside a case or on a bench. What is a "norm" for a case? If we take your case, mine, and 999,990 others I would be surprised if 5 of them are configured identically .. meaning there is no "norm" the same as there is no "real world environment" like the one tester is using.

OEM cooler are not really relevant .. except maybe for it's basic cooling ability to be used as a reference. Same goes it's airflow requirements.

From an analytical comparison of cooler performance we need to start with only the coolers themselves using a fixed heat source and intake air temperature .. nothing else.

Once we have their basic performance data, we can then factor in other variables and conditions they might be used with. But there are an infinite number these, and unless we remove as many as humanly possible to start with there is no credible base on which we can compare the coolers on their own merit. Hope that makes sense.

 

[MrTeal]

I started hi-jacking another thread and Micrornd replied, so instead of continuing I started this thread.

My original post
Micrornd's reply

Not quite sure where to start.

How about with what you said about my post "I understand why that makes sense from your point of view and although I agree to a point, I would say that may be a little too narrow."

You start out saying your system's coolers (radiators) are room temperature, but using that as an example is an even narrower interpretation of what I am talking about. And unless every component in your case is cooled by your loop they will be warmer than they would be on room temperature air.

I do agree, case airflow is critical. I have posted much on the subject and how to go about getting the best, and I will add here that the key is supplying component with cool intake air without their heated exhaust contaminating and warming it up.

Testing with system in a case or on an open bench really makes little or no difference. The key is monitoring the temperature of air going into cooler, not the room .. to remove all temperature variables from the equation. Other variables are still involved (barometric pressure, humidity, etc) but they are always involved, be it inside a case or on a bench. What is a "norm" for a case? If we take your case, mine, and 999,990 others I would be surprised if 5 of them are configured identically .. meaning there is no "norm" the same as there is no "real world environment" like the one tester is using.

OEM cooler are not really relevant .. except maybe for it's basic cooling ability to be used as a reference. Same goes it's airflow requirements.

From an analytical comparison of cooler performance we need to start with only the coolers themselves using a fixed heat source and intake air temperature .. nothing else.

Once we have their basic performance data, we can then factor in other variables and conditions they might be used with. But there are an infinite number these, and unless we remove as many as humanly possible to start with there is no credible base on which we can compare the coolers on their own merit. Hope that makes sense.

For the average review site, as long as they use a consistent test bed across different review and that case has decent airflow, using ambient is probably not a bad choice as it gives the reader a better feel for how a particular cooler will perform in a real world system. People know their CPU temperature and their room temps, few probably know their internal case temperature.

An additional benefit is that it can show secondary effects you might miss if you just test the cooler input and output temperature for a given heat load. GPUs see that with a blower vs an open air cooler, where while a blower usually performs worse than an open air cooler it does mitigate the need for really good case cooling if you're dumping 500W out the back vs into the case. Similarly, with a CPU cooler and a standard (non-ducted) installation you're dumping the heat from your CPU into the case which then needs to be exhausted by your case ventilation. Using a radiator (or ducting) provides some of that case airflow.

The big problem with using core and ambient temperatures is that it makes comparing reviews across sites impossible; the difference between coolers is going to be less in many cases than the difference in installations. A better solution might be to provide core, case and external ambient temperatures, but that makes it a little more challenging to put into a graph where you can just rank things by a single number.

Ultimately I'd love to see more reviews with a controlled input temperature and a known heat load. "It's a 4790k at 4.5GHz running Prime95" isn't really a proper or repeatable variable. I'd also love to see them tested at different heat loads, different fans speed both with the stock fans and a standard fan used for all reviews, etc. I just don't know if there's enough other people who'd want pages of data for each heatsink to keep a site like that in business.

 

[doyll49]

The problems with using ambient is as stated, it is not consistent The air temperature going into the cooler radiator often changes with individual cooler and fans depending on what their airflow is. Different airflow speed and volume change the path the air flows and therefor the temperature. As we know if the air temp into cooler changes so does the CPU them. But while the cooler intake air temp changes the room remains the same.

There is no "real world" system that is more than remotely similar to most of our systems. Saying it's a "real world" system is saying your car is "real world" and therefore the same as a race car. They both use the same fuel and both have 4 tires.and have other things in common, but the race car uses more fuel, generates way more heat, has better suspension, etc .. and will navigate the race course in a fraction of the time your car will.

For example, a test system has case fans running at a constant speed delivering a constant airflowo of let us say 150cfm total (4x 140mm fans running 800-1000rpm behind grills and filters) .. the same for every cooler tested. Please keep in mind this 150cfm is not just supplying the CPU cooler but everything else in the case except the PSU which is most likely drawing air through another case vent. When the cooler is using a single 1800rpm 120mm fan delivering 64cfm to the cooler all is fine. But when the cooler is using a pair of 2500rpm 140mm fans moving 130cfm of air it becomes another matter entirely. The case's rear exhaust fan is only removing 55cfm at best means another 85cfm of heated air that is coming out of the cooler is swirling around in the case contaminating and heating up the cool air going to cooler.

Another example is down-flow coolers. They usually run quite warm .. often because they are ingesting their on heated exhaust. Ariflow is in, down through cooler, turning out on mobo, then up at RAM, GPU, etc. past cooler and back into fan. Turn the fan over and CPU temps usually drop 7-10c.

Using a CPU running same program does a decent job of generating consistent and repeatable amounts of heat. Not perfect, but not too bad .. at least for all testing done on that platform.

Bur not keeping the cooler intake air temperature consistent can be problematic even on an open test station .. and is definitely an even bigger problem inside of a case.

GPUs using water cooling do more to improve the overall cooling of a system then watercooling on the CPU.

I don't see a need to be able to compare all the review sites as if they were using the same test systems. As long as each of their test systems give similar results with the same group of coolers as another the results can be credible. The problem is most sights will show the D15, R1, PH-TC14PE, Silver Arrow etc. have different CPU temp that are not even close to the same ratio of cooling ability to each other or to other coolers. Running a 117w, 242w and 345w heat source under the same series of coolers will give results that mirror each other. While the 117w load will have much less of a spread than the 345w load, the graphs will be similar. Most of the test sites results do not have this consistency .. mostly because of cooler intake air temp changes.


You mean something like this?

 

[MrTeal]

The problems with using ambient is as stated, it is not consistent The air temperature going into the cooler radiator often changes with individual cooler and fans depending on what their airflow is. Different airflow speed and volume change the path the air flows and therefor the temperature. As we know if the air temp into cooler changes so does the CPU them. But while the cooler intake air temp changes the room remains the same.

There is no "real world" system that is more than remotely similar to most of our systems. Saying it's a "real world" system is saying your car is "real world" and therefore the same as a race car. They both use the same fuel and both have 4 tires.and have other things in common, but the race car uses more fuel, generates way more heat, has better suspension, etc .. and will navigate the race course in a fraction of the time your car will.

True, but as long as the site uses the same test bench for all their testing, the temperature rise between ambient and case should be similar, which just leads to a common offset in all the coolers they test. Again, it's not optimal but it's not a bad choice, and the delta over ambient of a review conducted in a case is likely closer to what an average user can expect in their case given vs on an open air bench.

For example, a test system has case fans running at a constant speed delivering a constant airflowo of let us say 150cfm total (4x 140mm fans running 800-1000rpm behind grills and filters) .. the same for every cooler tested. Please keep in mind this 150cfm is not just supplying the CPU cooler but everything else in the case except the PSU which is most likely drawing air through another case vent. When the cooler is using a single 1800rpm 120mm fan delivering 64cfm to the cooler all is fine. But when the cooler is using a pair of 2500rpm 140mm fans moving 130cfm of air it becomes another matter entirely. The case's rear exhaust fan is only removing 55cfm at best means another 85cfm of heated air that is coming out of the cooler is swirling around in the case contaminating and heating up the cool air going to cooler.

That's a bit of a simplification. Barring ducting, the amount of airflow over the CPU cooler doesn't have to match the case airflow at all. Increasing the velocity of the air through the CPU cooler just serves to lower the delta between the cooler and the intake air, and all that air will be swirling around inside the case regardless of your level of case airflow. Increasing (or decreasing) your case airflow will just change the delta between the room ambient and your case temperature. Even if you have much more airflow through the case than through the cooler you're still going to have an offset between case and ambient.

Using a CPU running same program does a decent job of generating consistent and repeatable amounts of heat. Not perfect, but not too bad .. at least for all testing done on that platform.

Agree with your caveat, which is what I meant by repeatable. You might be able to get the same heat out of your CPU on multiple runs, but my same model CPU with identical settings won't. I can't repeat your results. OTOH, a watt is the same for everyone.
Even then, the heat load does change. Even the same CPU will produce a markedly different amount of heat at 35C underwater than at 80C under a small air cooler.

GPUs using water cooling do more to improve the overall cooling of a system then watercooling on the CPU.

Definitely, though you could probably get as good a results on your GPU using an air cooler as the CPU if you made it a 7 slot card as big as an NH-D15.

I don't see a need to be able to compare all the review sites as if they were using the same test systems. As long as each of their test systems give similar results with the same group of coolers as another the results can be credible. The problem is most sights will show the D15, R1, PH-TC14PE, Silver Arrow etc. have different CPU temp that are not even close to the same ratio of cooling ability to each other or to other coolers. Running a 117w, 242w and 345w heat source under the same series of coolers will give results that mirror each other. While the 117w load will have much less of a spread than the 345w load, the graphs will be similar. Most of the test sites results do not have this consistency .. mostly because of cooler intake air temp changes.

Being able to compare between review sites would be nice as no one site can test everything. If more sites did more controlled testing it would make comparisons easier.

The reason for running with multiple different power levels is that heatpiped coolers are actually pretty complicated systems and can display non-linear response at the tops of their ability. You might expect that a cooler on a 200W load would have about twice the delta-T of that cooler on a 100W load, but that's not the case if you're saturating the heatpipes.




Regardless, I'm not saying that trying to isolate just the HSF isn't a better way of testing, just that it's not necessary and in some cases might mask other benefits of the cooler like a CLC exhausting all the hot CPU air out the back of the case.

 

[doyll49]

True, but as long as the site uses the same test bench for all their testing, the temperature rise between ambient and case should be similar, which just leads to a common offset in all the coolers they test. Again, it's not optimal but it's not a bad choice, and the delta over ambient of a review conducted in a case is likely closer to what an average user can expect in their case given vs on an open air bench.

That's is where you are wrong, the cooler intake temp when using a case varies dramatically from room ambient. Way more than when testing on an open bench. I make these statements after running ran many tests with many different cases, with different case fan setups and many different coolers with both stock and different fans.

That's a bit of a simplification. Barring ducting, the amount of airflow over the CPU cooler doesn't have to match the case airflow at all. Increasing the velocity of the air through the CPU cooler just serves to lower the delta between the cooler and the intake air, and all that air will be swirling around inside the case regardless of your level of case airflow. Increasing (or decreasing) your case airflow will just change the delta between the room ambient and your case temperature. Even if you have much more airflow through the case than through the cooler you're still going to have an offset between case and ambient.

The case must flow at least as much air as the combined total of all components .. and that is assuming that there is no contamination of cool intake air going to component by their heated exhaust .. which is almost impossible.

Increasing velocity may or may not change the cooling much, but any change in the temperature of air going into cooler definitely has an almost exact 1:1 change in CPU temp.

You are talking in circles. The whole point of doing open air testing and monitoring the cooler intake air temperature is to supply the coolers with a consistant air temperature.

Changing the amount of case airflow does not mean the the cooler is getting a different temperature air or that the cooler intake is getting cooler air. Again, it is about both where the air flows as well as how much air is flowing.

Agree with your caveat, which is what I meant by repeatable. You might be able to get the same heat out of your CPU on multiple runs, but my same model CPU with identical settings won't. I can't repeat your results. OTOH, a watt is the same for everyone.
Even then, the heat load does change. Even the same CPU will produce a markedly different amount of heat at 35C underwater than at 80C under a small air cooler.

This is true. But if we both test the same coolers there will be a correlation between your results and mine if they are both receiving air at a constant temperature. The amount of heat displacement of same model coolers while not giving the same temperature on each of our systems will give the same amount of cooling per watt.

Definitely, though you could probably get as good a results on your GPU using an air cooler as the CPU if you made it a 7 slot card as big as an NH-D15.

I agree, but we are not talking about bigger cooler on GPUs. Lets save this for another discussion.

Being able to compare between review sites would be nice as no one site can test everything. If more sites did more controlled testing it would make comparisons easier.

It is not about how many coolers each site tests, but how they test them.

The reason for running with multiple different power levels is that heatpiped coolers are actually pretty complicated systems and can display non-linear response at the tops of their ability. You might expect that a cooler on a 200W load would have about twice the delta-T of that cooler on a 100W load, but that's not the case if you're saturating the heatpipes.

The chart of data I posted shows a direct correlation between amount of heat and cooling ability of both heatpipe and CLC.

Regardless, I'm not saying that trying to isolate just the HSF isn't a better way of testing, just that it's not necessary and in some cases might mask other benefits of the cooler like a CLC exhausting all the hot CPU air out the back of the case.

HSF testing accuracy and changes in case airflow and/or other differnces like exhausting heated air directly out of case are two different and separate issues of cooling systems.

We could have a thread to discuss case airflow too.

 

[MrTeal]

That's is where you are wrong, the cooler intake temp when using a case varies dramatically from room ambient. Way more than when testing on an open bench. I make these statements after running ran many tests with many different cases, with different case fan setups and many different coolers with both stock and different fans.

I didn't say the cooler intake wouldn't rise, just that for a given case and heat load the delta should be the same. IE, if your room ambient is 20C and you're dissipating 200W, if your case temperature at the intake of the HSF is 30C with cooler A, it will also be 30C with cooler B, C, etc. It will create an offset but that offset should be similar for most coolers.

The case must flow at least as much air as the combined total of all components .. and that is assuming that there is no contamination of cool intake air going to component by their heated exhaust .. which is almost impossible.

This is just incorrect. I could install a heatsink and test load in a sealed aluminum enclosure, and the fan on the heatsink could still move 100CFM through the cooler. The air temperature in that aluminum box would rise until it dissipates 200W to the atmosphere through radiation and convection, but the system would eventually reach steady state.

Increasing velocity may or may not change the cooling much, but any change in the temperature of air going into cooler definitely has an almost exact 1:1 change in CPU temp.

Exactly, and in the (extreme) example above, at that steady state with internal air temperature X, there would still be a delta-T between X and the load temperature. That delta T would be about the same with an internal temperature of 80C in the closed system as it would be with an internal case temperature of 20C if I'm using two 6800 CFM fans for intake and exhaust.

Changing the amount of case airflow does not mean the the cooler is getting a different temperature air or that the cooler intake is getting cooler air. Again, it is about both where the air flows as well as how much air is flowing.

It absolutely does. The more case airflow you have, the closer your case temperature (and by extension the cooler intake temperature) will be to the external ambient temperature.

This is true. But if we both test the same coolers there will be a correlation between your results and mine if they are both receiving air at a constant temperature. The amount of heat displacement of same model coolers while not giving the same temperature on each of our systems will give the same amount of cooling per watt.

Agreed, it just makes cross-referencing results more difficult. If Anandtech tests Coolers A, B, C and D while some other site tests A, C, D and E, you can get a pretty good idea of how B might perform relative to E. If the other site tests E, F, G, and H, you can't draw any conclusions about how the first four perform relative to the last four.

The chart of data I posted shows a direct correlation between amount of heat and cooling ability of both heatpipe and CLC.

In the linear part of their operating range, yes. Take a look at the NH-L12, for instance. Granted most of my experience with this is with higher heat loads that you'd ever see on a common CPU (600W+) with smaller air cooler, but you can easily reach the point where power vs delta-T becomes non-linear even with stupid (dual 250CFM TFC1212DE fans). You're maxing out the ability of the heat pipe to be a heat pipe, and start to rely on conduction through the copper itself which is much less efficient.

 

[doyll49]

I didn't say the cooler intake wouldn't rise, just that for a given case and heat load the delta should be the same. IE, if your room ambient is 20C and you're dissipating 200W, if your case temperature at the intake of the HSF is 30C with cooler A, it will also be 30C with cooler B, C, etc. It will create an offset but that offset should be similar for most coolers.

It is obvious you have not monitored cooler intake and room temperatures in different systems. The fact is there is no consistant correlation between them. There are too many variables involved .. of which case airflow is only a small part. Cooler fan specifications, cooler size and orientation, etc. all play a significant role in what the cooler intake air temp is. And as all coolers and their fans are not the same, their cooler intake air temps are also not the same.

Before we can continue this discussion farther we have to come to agreement on the fact that cooler intake air temperature changes depending on cooler fan speed and what cooler is being used. This temperature change is not consistent nor linear with room ambient .. and therefore the deltas are not the same. This is the reason I preach the use of cooler intake instead of room ambient.

This difference also changes dependent on cooler fan speed.

This has been proven by everyone who has monitored both temperature locations. Not all cases are the same, but the only testing I've seen was done by George Cella in his Enthoo Luxe case with 6.8-8.8c difference between room and cooler (2c variance from linear)

As an analogy, is the temperature in your kitchen the same or change in a linear relationship with your lounge or bedroom? Mine does not, especially the kitchen and lounge. Their temperature changes dependent on if cooking is being done or if fireplace is being used .. even with central heating/cooling and all doors open.

 

[MongGrel]

Why are you asking for advice along these lines if you seem dead set on not listening to what anyone says ?

 

[doyll49]

Why are you asking for advice along these lines if you seem dead set on not listening to what anyone says ?

The thread is to discuss the problems with current cooler testing criteria not asking for advice.

My testing and evaluations of coolers .. and other who use cooler intake air temp have all come to the same conclusion. Cooler / radiator intake air temp is the critical baseline temperature, not the room ambient. While there is a correlation between them, oom ambient is only one of many variables affecting and determining the cooler intake air temp.

Even on a bench test station the temperature of airflow into cooler may be different than room. The PSU, lighting on bench, even tester standing / setting at the station are all emitting heat and effecting the air temperature. How much / how little can be debated, but Anyone who has monitored both temperatures knows this to be true. Therefore there is no reason to discuss which is the best way to monitor .. intake air temperature the most accurate source of air temperature for determining cooler delta temperatures. Using any other source allows for other possible variables to affect/change the air temperature before going into cooler.

 

[MongGrel]

Um ok.

 

[doyll49]

Trolling doesn't change the facts.:biggrin:
Test it out for yourself to prove what is true. But saying things just because of a believe based on no evidence has no place in an analytical discussion. I have posted examples and data verifying what I'm saying. I wish I had kept data of cooler intake and room when setting up cases airflow to optimize cooling. I just assume the rational and logic of the differences between cooler intake air temp and room ambient to be so obvious as to not need much to prove it.

 

[MrTeal]

It is obvious you have not monitored cooler intake and room temperatures in different systems. The fact is there is no consistant correlation between them. There are too many variables involved .. of which case airflow is only a small part. Cooler fan specifications, cooler size and orientation, etc. all play a significant role in what the cooler intake air temp is. And as all coolers and their fans are not the same, their cooler intake air temps are also not the same.

Before we can continue this discussion farther we have to come to agreement on the fact that cooler intake air temperature changes depending on cooler fan speed and what cooler is being used. This temperature change is not consistent nor linear with room ambient .. and therefore the deltas are not the same. This is the reason I preach the use of cooler intake instead of room ambient.

This difference also changes dependent on cooler fan speed.

This has been proven by everyone who has monitored both temperature locations. Not all cases are the same, but the only testing I've seen was done by George Cella in his Enthoo Luxe case with 6.8-8.8c difference between room and cooler (2c variance from linear)

As an analogy, is the temperature in your kitchen the same or change in a linear relationship with your lounge or bedroom? Mine does not, especially the kitchen and lounge. Their temperature changes dependent on if cooking is being done or if fireplace is being used .. even with central heating/cooling and all doors open.

Ah, but he's looking at closed loop water coolers. Unless he has the radiator dangling inside the case, the CPU cooler is actually changing the case airflow. This differs from a tower CPU cooler.

That's why I originally said that using a cooler in the case isn't a bad idea, as some like CLCs have additional benefits that you might not see on an open air test bench monitoring intake temperatures.

 

[doyll49]

Ah, but he's looking at closed loop water coolers. Unless he has the radiator dangling inside the case, the CPU cooles actually changing the case airflow. This differs from a tower CPU cooler.

That's why I originally said that using a cooler in the case isn't a bad idea, as some like CLCs have additional benefits that you might not see on an open air test bench monitoring intake temperatures.

Actually what George found was the opposite of what usually happens when using CLCs.. CLCs (AIOs too) usually give lower airflow temps inside of case than tower coolers do.

Enthoo Luxe is a very well ventilated case. full top and bottom venting as well as 2x 140mm front and back is almost all vented with a 12.5x16.5cm and . 4.5x16cm grills plus all PCIe covers are vented.

Have you actually monitored the testing of a bunch of coolers in a case to see the effect different coolers have on case airflow and temperatures? I don't think you have. If you had you wouldn't be saying room ambient and cooler intake air temps have a linear comparison. Not unless the case airflow is tuned to the cooler .. or coolers of same style and airflow.

 

[MrTeal]

Actually what George found was the opposite of what usually happens when using CLCs.. CLCs (AIOs too) usually give lower airflow temps inside of case than tower coolers do.

Enthoo Luxe is a very well ventilated case. full top and bottom venting as well as 2x 140mm front and back is almost all vented with a 12.5x16.5cm and . 4.5x16cm grills plus all PCIe covers are vented.

Have you actually monitored the testing of a bunch of coolers in a case to see the effect different coolers have on case airflow and temperatures? I don't think you have. If you had you wouldn't be saying room ambient and cooler intake air temps have a linear comparison. Not unless the case airflow is tuned to the cooler .. or coolers of same style and airflow.

Less in ATX tower cases, most of my testing has been in 4U cases and open air.

Those results that George saw are a little counter-intuative. I'd be interested to see a link to his results, and whether he removed the top 140mm fan when testing the CLCs.

 

[doyll49]

Can I assume these are standard 432x384x132mm 4U cases? If this is true only a few consumer coolers that will fit .. and none of the top tier coolers. I have done no testing with racks so can only guess how they perform, but if they are the 4x front intakes with straight through airflow I would assume the cool quite well. A bit on the noisy side, but then they are usually in their own areas so it's not much if any issue.

I think the testing was published on as video review. He give me the data separately.
Here is link to it
http://www.hitechlegion.com/reviews...10i-gt-and-h100i-gtx-clc-liquid-cooler-review

 

[MongGrel]

 

[doyll49]

@MongGrel
Don't see many of those Ultra Copper around. Great cooler, but than so is NH-D14.

Is the D14 flowing up or down? I assume this orientation is for RAM clearance .. or is it GPU clearance?

 

[MongGrel]

Up.

The Antec 1200 has a 200 above it in the top, is the only fan hasn't been modded out.

It will clear RAM horizontal, I went vertical on purpose.

Just took kicking the fans around a bit.

It's not really for clearance, was for airflow, it sucks heat off the backside of the GPU.

 

[doyll49]

I have done the same when case vent alignment / size was better up then back. I also used a shroud/duct to flow the exhaust out without it contaminating cool air.

 

[MrTeal]

Can I assume these are standard 432x384x132mm 4U cases? If this is true only a few consumer coolers that will fit .. and none of the top tier coolers. I have done no testing with racks so can only guess how they perform, but if they are the 4x front intakes with straight through airflow I would assume the cool quite well. A bit on the noisy side, but then they are usually in their own areas so it's not much if any issue.

I think the testing was published on as video review. He give me the data separately.
Here is link to it
http://www.hitechlegion.com/reviews...10i-gt-and-h100i-gtx-clc-liquid-cooler-review

Standard cheap 4U server cases, and basically empty 4U boxes. Originally larger rack cases were looked at to fit large tower coolers, but the lack of performance in the application didn't justify the extra cost. Even with 250CFM Deltas on it, an NH-D14 couldn't dissipate 500W with acceptable temperatures. The CoolIt 120mm CLC (IE the H80i) worked well enough, though for my own personal application I ended up going with Swiftech H220s. Even in a 35C ambient I was able to keep die temps at an acceptable 100C when dissipating 800W at the chip, with a direct die mounting using CL Liquid Ultra.

As you mentioned though, reliability on the CLCs was a concern, especially with the H220s prior to their new stepping. Of the 16 I originally bought, I had to return half of them due to pump failures. Granted running 24/7 at extreme temperatures probably didn't help, but it's still an obvious area for them to work on.

 

[doyll49]

Standard cheap 4U server cases, and basically empty 4U boxes. Originally larger rack cases were looked at to fit large tower coolers, but the lack of performance in the application didn't justify the extra cost. Even with 250CFM Deltas on it, an NH-D14 couldn't dissipate 500W with acceptable temperatures. The CoolIt 120mm CLC (IE the H80i) worked well enough, though for my own personal application I ended up going with Swiftech H220s. Even in a 35C ambient I was able to keep die temps at an acceptable 100C when dissipating 800W at the chip, with a direct die mounting using CL Liquid Ultra.

As you mentioned though, reliability on the CLCs was a concern, especially with the H220s prior to their new stepping. Of the 16 I originally bought, I had to return half of them due to pump failures. Granted running 24/7 at extreme temperatures probably didn't help, but it's still an obvious area for them to work on.

Indeed, The original H220 did have pump issues. I'm assuming this was resolved and the new X series seem to be very good. Water cooling does have advantages, but the cost is just too high for my budget. CLCs have such low end components I just don't trust them.

I agree, even top air is only good to 350-425w (guessing as I don't have equipment to test). Even then would need 20-25c ambient. And from what I've seen with current tech that is about as good as it's going to get.

 

[Micrornd]

I started hi-jacking another thread and Micrornd replied, so instead of continuing I started this thread.

My original post
Micrornd's reply

Not quite sure where to start.

How about with what you said about my post "I understand why that makes sense from your point of view and although I agree to a point, I would say that may be a little too narrow."

You start out saying your system's coolers (radiators) are room temperature, but using that as an example is an even narrower interpretation of what I am talking about. And unless every component in your case is cooled by your loop they will be warmer than they would be on room temperature air.

No I think you mis-read, I did not say my radiators were at room ambient temp, I said "In my particular case, room ambient temp is the same as cooler intake temps".
This was in reference to the fact you were talking about how cooler intake air temps are not the same as room ambient temps.
I was only pointing out that while it is true that with a normal air cooler, buried in a case, that air intake temps at the cooler will be normally higher than room ambient, that is not always the case with radiators.

My CPU temps do not exceed 55c at full load though (with much less cfm than stock) and a big part of that is because the radiators air intake temp is room ambient temp and not pre-heated to case ambient temp.

So I think we are on the same page on this.

I do agree, case airflow is critical. I have posted much on the subject and how to go about getting the best, and I will add here that the key is supplying component with cool intake air without their heated exhaust contaminating and warming it up.

Testing with system in a case or on an open bench really makes little or no difference. The key is monitoring the temperature of air going into cooler, not the room .. to remove all temperature variables from the equation. Other variables are still involved (barometric pressure, humidity, etc) but they are always involved, be it inside a case or on a bench. What is a "norm" for a case? If we take your case, mine, and 999,990 others I would be surprised if 5 of them are configured identically .. meaning there is no "norm" the same as there is no "real world environment" like the one tester is using.

I believe that is what I said?
"There are just so many variables that factor in beyond inlet temp, that there is no really good way to test coolers."

So maybe just being aware of all the variables that enter into the cooling scenario is the most important factor of all, rather than just out of context temp and noise figures."

OEM cooler are not really relevant .. except maybe for it's basic cooling ability to be used as a reference. Same goes it's airflow requirements.

We definitely don't agree here, and this is one of the areas I was talking about where your view is too narrow.
(This is based on Intel, although I assume AMD is the same)
You are only looking at how well the CPU is being cooled, not how the OEM cooler works.
OEM (Intel) coolers either draw air over, blow air over, or deflect air onto the VRMs and other components around the socket.
That is one of the reasons almost all OEM MBs locate those components in the same place, use the same height heatsinks on VRMs, why some VRMs, but sometimes not all have heatsinks, etc.

Ignoring the fact that a good air cooler, such as the Noctuas, do relatively nothing to cool those components is wrong.
It means that component temps around the socket are higher and more heat is released in the case by those components, than would be by an OEM cooler, raising the air intake temp at the cooler.
Those same higher VRM (and other component temps) can become a critical factor for those that overclock, unless additional dedicated cooling is added.
While cooling the CPU is the major job of the cooler, it is not the only job for an OEM cooler and this must either be done by a replacement cooler or additional work and parts must be added to make up for it.
This is especially true for water cooled systems, VRM and MB components that were normally cooled air from the OEM coolers, must be compensated for by additional airflow by other means.

The point is that if 2 aftermarket coolers cool the CPU the same, but one also provides cooling air for the VRMs, it is the better cooler and better buy and this must also be factored in, not an afterthought.

From an analytical comparison of cooler performance we need to start with only the coolers themselves using a fixed heat source and intake air temperature .. nothing else.

Once we have their basic performance data, we can then factor in other variables and conditions they might be used with. But there are an infinite number these, and unless we remove as many as humanly possible to start with there is no credible base on which we can compare the coolers on their own merit. Hope that makes sense.

Yes, that makes sense, but as I mentioned above, I think there are other primary factors to consider at this base level, not later.
Some sort of standard should be established, but to be of any real use it must in some way involve restrictions, such as found in the real world and not just an idealized test bench.
At the minimum, tested with standardized restrictions to exhaust flow, raised inlet temp, cross-airflow, etc. to attempt to bring it into the real world.

 

[doyll49]

I meant using room temperature air into radiator is alos narrow minded, my bad.

Cooler / radiator intake air temp can usually be kept from 2c to 5c pretty easily. Obviously what case it is, etc. do make a difference.

I think what you said was

I've seen "open air" testing for coolers proposed also, but to me, that's a much worse solution than testing in a case, as it has way too many uncontrolled variables, as to airflow restrictions and deviations from what would be a "norm" in a case. (i.e. cables, other cards, other heat sources, other case fans, etc.)

That is what I was replying to, and to me I understand that to say testing in a case has less variables than on an open bench. Maybe I misunderstood.

I think by monitoring the cooler intake we can get results that others can come close to. Obviously different component (CPU) will generate different amounts of heat, but the result for a bunch of same coolers tested on different systems will have very similar difference in temperatures from one test bed to another .. as long as the cooler intake is base temperature used is similar. Testing with 30c desert air versus 22c humid coastal air makes quite a difference in cooling ability .. the 22c coastal air can absorbs way more heat.


I definitely agree that aftermarket coolers need to cool motherboard component at least as well as OEM coolers. My reference to OEM coolers was only to cooler used being used as a reference for comparing CPU temps to more capable coolers.

But I have found most aftermarket coolers when used with low to medium height RAM and fans set as close as possible to mobo cool mobo just fine .. espeically if their idle speed is kept fast enough to keep a constant airflow. It is not uncommon for a top tier cooler to have higher mobo component temps at idle than at full load.


Agree again.
What do you propose as basic standards? I'm having a hard time with "tested with standardized restrictions to exhaust flow, raised inlet temp, cross-airflow, etc." Could you give a more detailed description of these parameters please?


When testing I do watch mobo temps, but unless they are out of the normal range I don't mention them .. and I should.

Thanks for the reply.

 

[MrTeal]

I watched the video, and the AOI coolers are mounted in the top location. I checked the NH-D15 review, but that was in a different case so I'm not sure on the mounting or existence of the top 140mm fan for the install there.

This does have me intrigued though, so I'm thinking of doing some testing on it. I'd propose the following test based on what I have around.

Case: Coolermaster CM690
Heat Load: Custom resistive heat load
Case ventilation: Front 140mm, rear 140mm (make TBD)
Coolers to Test: H80i, H220, NH-D14, Hyper 212 Evo
Measurement Points: Room Ambient, Cooler intake, Hotplate temperatures

Test would consist of three heat loads: 95W (TDP of mainstream K), 140W (TDP of HEDT parts), 300W (Overclocked HEDT). For each heat load, 9 test points would be sampled with the system and cooler fans set to various levels, PWM %s TBD. IE, at 95W I might test the 9 possible combinations of the system and CPU fans set to 50, 75 and 100% PWM.

Unfortunately I can't seem to find my anemometer right now, but if I find it I'd measure exhaust velocity as well.


It might not happen for awhile, but any comments on the test plan?

 

[doyll49]

@MrTeal
George tests all coolers compared in a review the same day as the review cooler. This is so variables like humidity, barometric pressure, etc. are all the same.

I don't retest all coolers, but I do a baseline test with a previously tested cooler to be sure conditions are the same for each test session. Living in temperate climate on the coast means conditions don't vary much.

When testing in a case I monitor the the air temp near case, in airflow into cooler and in airflow out of cooler. I have not been monitoring exhaust velocity but I like the idea, and as I have a hotwire anemometer I'll start doing it. I use an Aquaero 5 LT to control fans and use rpm rather than %PWM, but do record %PWM. I also record the rpm at each speed. I use Thermalright TY-143 for reference fans. I like them because they run from 500rpm to 2500rpm. I've made a custom adapter shroud from 140 to 120mm for smaller coolers. I also monitor noise level at 30 cm and 1 meter.

I'm hoping to do a new test system soon.