Overclock.net banner

181 - 200 of 278 Posts

·
Fix it till its broke
Joined
·
947 Posts
Ahh no need to apologize 👍🏻

I’m not sure why I mentioned RGB either 🤔

Either way my ram has it and I like it.. it’s so bright and blingy, it just sucks you need software to control it, and nothing in the bios other than on/off for the onboard stuff. Although.. with my new cooler the fan covers my ram so that’s kind of a bummer. But it’s ok..
 

·
Registered
Joined
·
217 Posts
Discussion Starter · #182 ·
I’m not sure why I mentioned RGB either 🤔

Either way my ram has it and I like it.. it’s so bright and blingy, it just sucks you need software to control it, and nothing in the bios other than on/off for the onboard stuff. Although.. with my new cooler the fan covers my ram so that’s kind of a bummer. But it’s ok..
I personally prefer basic LED fans due to reasons like this and only last year I got motherboard with proper PWM control instead of few preset modes. When something ends up on motherboard and relies on software to work well, it's always crapshoot. For such reasons I would always prefer to have dedicated hardware solution. For fan speed fan controller (I never had one, because they were really expensive and never truly available) is better (or just getting fans that spin at right speed), for wifi I would always prefer to use dedicated card, for RGB there doesn't seem to be any dedicated front panel controller. That's a shame, because it's always nice to adjust things with knobs instead of hopefully not a dumpster fire software.
 

·
Fix it till its broke
Joined
·
947 Posts
I am running 120x38 case fans mostly, I have 4 of them in my Meshify C. At the top front I have a TY143 and top rear I have a TY147B.. because it doesn't make noise upside down like the A does.. I am looking at replacing all of my 120x38s with TL-C12Pro's. Not the strongest, or the quietest, or the cheapest.. or the biggest :D

I don't plan on adding any more RGB than it already has.. I don't even want to give up my Panaflos, but I can hear two have a rumble going on in the bearing when they are at 12v. They have been running every day since 2006 or so, they don't owe me anything.. 20 bucks each for 110cfm.. cant go wrong with that :)

Except at 12v because they are loud AF and I am getting to old for that business.
 

·
Registered
Joined
·
217 Posts
Discussion Starter · #184 ·
I am running 120x38 case fans mostly, I have 4 of them in my Meshify C. At the top front I have a TY143 and top rear I have a TY147B.. because it doesn't make noise upside down like the A does.. I am looking at replacing all of my 120x38s with TL-C12Pro's. Not the strongest, or the quietest, or the cheapest.. or the biggest :D
And all I have is Silencio S400 (finally upgraded from K280) with single Scythe Kaze 120mm fan as top exhaust and it runs at 520 rpm. That's it. And it turns out that my 14 year old Scythe fan can start at quite low voltage, so it can spin at as little as 400 rpm, but for practical reasons and due to it sometimes failing to start during Windows boot, I set it at 600 rpm minimum. It only spins up to like 800 rpms during BOINC load (all CPU cores + GPU).

I don't plan on adding any more RGB than it already has.. I don't even want to give up my Panaflos, but I can hear two have a rumble going on in the bearing when they are at 12v. They have been running every day since 2006 or so, they don't owe me anything.. 20 bucks each for 110cfm.. cant go wrong with that :)
Ignoring their loudness, there's something fascinating with using really old fans and seeing how long they will last. I never saw 120mm fan fail. The only thing I ever saw close to failure is when I got my small AIO and PWM didn't work. Once PWM pin was bypassed, fan worked fine, but it was unusable due to those 2k rpms.


Except at 12v because they are loud AF and I am getting to old for that business.
Oh, I bet. I once tried 2.5k rpm Yate Loon in PC and that thing was super loud. At least it pushed a lot of air.
 

·
Fix it till its broke
Joined
·
947 Posts
At the machine shop I worked at before the world changed, the CNC lathes and mills we used had 120x38s and 140s as well. Both plastic and metal.. Coolant comes off the tooling as a mist, or in a lathe the spinning chuck turns it into a, aerosol mist. But over the course of a year that mist settles on those blades and turns to a thick oily sludge.. those industrials are tough, and have been running for 18 years. We didn't replace machines, we repaired them until we couldn't. In a nice environment I don't see why you couldn't get 20-25 years out of one.

I use those fans because I like to overclock in the winter and summer. I strongly dislike dialing back an overclock just because the ambient rose 10-15c :D
 

·
Premium Member
Joined
·
27,527 Posts
That's very unrealistic scenario as temperatures will be lower than in any case and therefore some potentially overheating hardware doesn't look as bad as on open air test bench. Intel stock coolers are sometimes tested in test bench and seem to be able to cope with stock CPU loads, but it fails in many cases just due to how close it is to overheating.

I personally wouldn't see any problem with using some nice and beefy ATX case like Fractal Design Define 7 with two intake fans and single exhaust fan. For GPU use something hot and something that doesn't have boost and many power saving features. So, it could be GTX 480 (not blower version). It should work fine if you change nothing, but coolers and only do thermal tests. For acoustics test you will need open bench and sound dampened room (and for that matter, a passively cooled PSU). I don't see why that wouldn't work well, especially in this era, when "gamer" GPU dumps a lot of heat into case and some of it will inevitably end up in cooler. That's quite a big deal in mid tower ATX cases and anything smaller than that. Besides something over the top like CM HAF series, I haven't seen a case that somehow could effectively deal with exhausted heat from GPU. (Although it wouldn't be impossible to fix. You can use blower card or some case manufacturer could design an adjustable air duct for side panel and mount 140/120mm exhaust fan on side, as well as air intake on the bottom of side panel with adjustable duct for GPU cooling. The reason why I don't like most side fan holes is because they are awkwardly placed. There's always a side panel mounting in place, where GPU both intakes and exhausts air, so it's impossible to know if it's better put fan as intake or exhaust. Even HAF X had this problem, but at least that fan seemingly helped GPUs to stay cool.).



Oh, not anymore. It seems that Thermalright is still doing some fine engineering. However, I think that Fuma 2's concept of using opposing fans is really cool and seems to work quite well. Now I would like to see how Frost Commander 140's heatsink would work with FUMA 2's fan setup. It could be really quiet and exceptionally well performing cooler. The only problem is that there are no fans sold that have blades spinning in opposing direction. And it would be even better, if blades would be from Noiseblocker eLoop fans. Knowing how legal system works, it would be easiest for Noiseblocker to enter cooler manufacturing. That or for Thermalright to collab with Noiseblocker and use Scythe's fan idea for their own product (it's not even Scythe's idea, it seems rather generic). It would be really cool and they could name that cooler "Thermalright True Airflow 140". If they made it right size and perhaps added third fan, this monstrosity could finally dethrone D15 for being "money no object" air cooling solution. Hell, it may even give some big AIOs a run for their money (for that it would also need to have shroud, so all airflow would go through fins, not be lost in case somewhere). And on top of it all, it wouldn't require any expensive materials or much additional manufacturing tools, so it could be sold for lower or comparable price to D15.
What is unrealistic is thinking you can test a cooler inside of a complete system. All that testing gives you is cooler's performance in that specific system. Only way to test coolers in built system is by monitoring and recording air temp entering cooler at same time as CPU temp is recorded. The resulting delta temp is cooler's cooling ability. Monitoring and recording cooler intake air temp is critical part of any cooler testing be it on open bench test station or in case built system. Problem with doing it in case built system is different coolers flow air differently so air temp entering cooler can vary dramatically. Flat downflow coolers are worst for increasing their intake air temp. Air temp entering cooler on open bench test station is more consistent, but flat downflow coolers still pre-heat their intake air (this can usually be resolved by fan flowing air away from mobo instead of toward). The more consistent the air temp entering cooler is the more accurate the results.

You are making assumptions about opposing fans helping. I've seen no testing that supports opposing fans help or hurt.

You talk about adding fan, yet testing has proven more fans make little to no difference. Noctua tested NH-D15S and NH-15 and found 1-3c difference with 1 vs 2 fans. I've done similar testing an found the only time there is even 1-3c difference is at full speed with multi-fan setup also being a few db louder. At same noise level single fan vs multi-fan coolers cool the same.

NH-D15 has never been king of the hill. Same applies to NH-D14. When NH-D14 came out Thermalright Silver Arrow was a few degrees cooler. NH-D14 cools same if not slightly better than NH-D15 with same fans. Other coolers cooling as well or better than D15 are PH-TC14PE, Silver Arrow. SBE, IBE, etc., R1, Fuma 2, Le Grand Macho RT, TRUE Spirit 140 Power, and the list goes on and on.
 

·
Registered
Joined
·
217 Posts
Discussion Starter · #187 ·
What is unrealistic is thinking you can test a cooler inside of a complete system. All that testing gives you is cooler's performance in that specific system. Only way to test coolers in built system is by monitoring and recording air temp entering cooler at same time as CPU temp is recorded. The resulting delta temp is cooler's cooling ability. Monitoring and recording cooler intake air temp is critical part of any cooler testing be it on open bench test station or in case built system. Problem with doing it in case built system is different coolers flow air differently so air temp entering cooler can vary dramatically. Flat downflow coolers are worst for increasing their intake air temp. Air temp entering cooler on open bench test station is more consistent, but flat downflow coolers still pre-heat their intake air (this can usually be resolved by fan flowing air away from mobo instead of toward). The more consistent the air temp entering cooler is the more accurate the results.
I wouldn't say it's unrealistic. Yes it is system specific, but gives a better idea to end user about what kind of performance one could expect in built system. Open test benches can only give a good about ideal cooler capabilities and is good for comparing coolers with each other. However, in complete system different coolers absorb surrounding heat differently. Perhaps using a built system for testing isn't the best method, but I think that effects of GPU heating up cooler, should be tested somehow. I'm pretty sure that nobody that reads those tests doesn't have any dedicated GPU.

You are making assumptions about opposing fans helping. I've seen no testing that supports opposing fans help or hurt.

You talk about adding fan, yet testing has proven more fans make little to no difference. Noctua tested NH-D15S and NH-15 and found 1-3c difference with 1 vs 2 fans. I've done similar testing an found the only time there is even 1-3c difference is at full speed with multi-fan setup also being a few db louder. At same noise level single fan vs multi-fan coolers cool the same.
They certainly don't hurt cooling:
Results are at 9:16. Changes sure are minor, but there are some changes. Single fan fares worse than 2 fans in any config. Opposing fans didn't do anything for thermals, except that for some reason hotspot was hotter, but core average was the same. 3 fans didn't change results much, although miniscule improvement was seen. I think it was due to cooler being rather small, if it had longer fins and second tower with longer fins, third fan would be effective. and if I remember correctly, adding more fans if airflow is already achieved, then it only increases pressure. So, results will depend on case. Opposing airflow does nothing for thermals (in open bench), but may do something for acoustics. And Fuma 2 comes with slow fans and we all know that for slow fans improvements to blades or motor, barely do anything, so opposite airflow concept may be better tested with faster fans (It's possible to remove blades from Kazes and then put them on different sleeve bearing fans). At least we know that opposing flow doesn't hurt anything, so that's good.

BTW fan stacking has been tested and it doesn't improve airflow (but that can improve pressure):

Adding more fans may help cooler more, when you test it in case with restrictive airflow. And since pretty much every case restricts airflow to some degree, more fans are usually better. On open test bench, effect of 3 fans will not be seen properly.


NH-D15 has never been king of the hill. Same applies to NH-D14. When NH-D14 came out Thermalright Silver Arrow was a few degrees cooler. NH-D14 cools same if not slightly better than NH-D15 with same fans.
You should really know better than that:

They are close, but Silver Arrow lost by a tiny speck.


Other coolers cooling as well or better than D15 are PH-TC14PE, Silver Arrow. SBE, IBE, etc., R1, Fuma 2, Le Grand Macho RT, TRUE Spirit 140 Power, and the list goes on and on.
Oh man, I will sure enjoy dethroning D15. Too bad that your list isn't good. Don't get me wrong, those heatsinks are well performing, but they lack a tiny bit to really match D15. Anyway, here's a list of all those coolers compared to D15:
Phanteks PH-TC14PE CPU Cooler Review - Results (Ascending order) (stock Phanteks won against stock D14, but lost in all out comparison and D14 is slightly worse cooler than D15)
Testing: Setup & Results - Thermalright Silver Arrow Review - Page 4 (some variation of silver arrow also lost to Prolimatech Super Mega)
CRYORIG R1 Ultimate Review (Cryorig R1 Ultimate didn't beat D15 either, it mostly matched it or exchanged places with D15)
Scythe Fuma 2 Review (Fuma 2 actually often beat Noctua by a bit, I will give you that. It's actually odd, because in many reviews it doesn't perform better than D15)
A CPU Cooler You NEED To Know About! Scythe Fuma 2 Review (And here we can see that Fuma 2 is a whole tier below D15, I have no idea why results vary so much. Seriously, in one review Fuma 2 was beating D15 and another it's losing by 15C. Somebody should investigate what is going on.)
Scythe FUMA 2 CPU Cooler Review (Fuma 2 performed slightly worse than D15, but once with equal fans it managed to won by miniscule amount and Fuma 2 was generally quieter than D15. So my verdict is that it matches D15.)
Thermalright Le Grand Macho RT CPU Cooler Review (Le Grand Macho RT performs like Fuma 2, loses to D15 with stock fan curve and wins once fans are normalized. I guess it matches D15, it's hard to say that it wins.)
Thermalright True Spirit 140 Power CPU Cooler Review (True Spirit 140 Power loses by quite a lot to D15)
Cooling - Thermalright's TRUE Spirit 140 "Power" - New Air Cooling King? Mini Review (however in another review True 140 Power had an edge)

And the conclusion is that, despite some review gremlins, most coolers performed like D15. There was no clear, definite winner against D15. D15 has not been meaningfully dethroned, but I wish it was. The main problem being is that all D15 competitors don't have particularly unique design or innovative features that would help them gain an edge. Therefore there isn't any truly meaningful change between them. Competitors will need to start using shrouds or triple tower configs, pretty much anything not already seen to go past current air cooling limits. I personally think that putting a shroud on air cooler would be the easiest and most logical step to improve their performance, as it wouldn't let any generated airflow to not flow outside of heatsink. That's far better idea than opposing fans as it solves the same problem, but far more effectively.
 

·
Fix it till its broke
Joined
·
947 Posts

·
Registered
Joined
·
381 Posts
Most AIO suck due to the weak pumps inside of them.
I have a custom loop that replaced my older CM 212 evo+.

Block: EK supremacy MX acetal AMD
Radiator: 240mm XSPC (35mm copper)
I put 2 old 120mm fans i had (you know...those black, dusty ones YOU KNOW YOU HAVE THEM) on the radiator and ran them at 7V(~800rpm)
First pump i got was a Phobya dc12-240, which was silent even at fullspeed - but a very weak pump, most AIOs have similar pumps.
Good enough for beginners + a 5.25" drive bay reservoir that can hold 0.5L of water.
Tubing and barbs/fittings bought at local plumbing shop for less than 7€.

These could keep an FX6300 [email protected] below 52°C(peak) after 2hours of y-cruncher stresstesting, the aircooler usually reached 58-60C.
I was kinda satisfied but i expected a bigger drop in temps after reading all the stories on watercooling.

I accidentally cracked the pump inlet on the phobya at some point and instead of getting another one, i did a lot of research on xtremesystems and i went for the swiftech mcp655(ddc laing)
This pump, at full speed(~4450rpm) on the same loop, could keep the CPU temp below 42°C
I removed one fan from the radiator and temps went up to 44°C, dropped pump speed to 1600rpm and temps went up to 48°C.(Turning off fans, running pump at 100% pushed temps to 52°C)

Was the air-cooler more convenient? Yes and cheaper.
Are AIOs worth it? Get an air cooler.
Is a custom loop worth it? Not on value or convenience, but it's the only thing you can build with your hands for your PC.
Do AIOs and aircoolers perform the same? LOL, a proper loop is not comparable.

Remember, this is a custom loop on a budget(170-200€), there are better blocks, stronger pumps, thicker radiators, static pressure fans.
This is entry-level performance for custom water.

Should i go custom?
Only if someone else would trust you with assembling their loop. ;)

The biggest difference on custom water vs Air/AIO is the gradual increase of temperature no matter the load - there are no spikes higher than 10°C.
With a bigger reservoir it can take an hour or more to reach peak temps, but it only takes a few minutes to cool back down, you can't achieve that with air, you can't hold the excess heat in one place.
 

·
Registered
Joined
·
217 Posts
Discussion Starter · #190 ·
Reviews should be taken with a grain of salt. I have gotten better results on my own then some reviewers.

Thermalright Le Grand Macho RT Review - Test System & Temperature Results | TechPowerUp

It shows as equal. But in my own little comparo my LGMRT did better than my old D14 did on the same hardware. And even better when using a TY-143.
Not sure exactly why for you it performs better, but there are 3 things that might have:
Macho has more static pressure
Your case might have less than ideal airflow
Noctua was better at reabsorbing GPU heat.

Anyway, I doubt that difference was big between those coolers. And BTW D15 is an improved version of D14. There's a significant difference between them both. D15 performs 3-5C better than D14.
 

·
Premium Member
Joined
·
27,527 Posts
I wouldn't say it's unrealistic. Yes it is system specific, but gives a better idea to end user about what kind of performance one could expect in built system. Open test benches can only give a good about ideal cooler capabilities and is good for comparing coolers with each other. However, in complete system different coolers absorb surrounding heat differently. Perhaps using a built system for testing isn't the best method, but I think that effects of GPU heating up cooler, should be tested somehow. I'm pretty sure that nobody that reads those tests doesn't have any dedicated GPU.

They certainly don't hurt cooling:
Results are at 9:16. Changes sure are minor, but there are some changes. Single fan fares worse than 2 fans in any config. Opposing fans didn't do anything for thermals, except that for some reason hotspot was hotter, but core average was the same. 3 fans didn't change results much, although miniscule improvement was seen. I think it was due to cooler being rather small, if it had longer fins and second tower with longer fins, third fan would be effective. and if I remember correctly, adding more fans if airflow is already achieved, then it only increases pressure. So, results will depend on case. Opposing airflow does nothing for thermals (in open bench), but may do something for acoustics. And Fuma 2 comes with slow fans and we all know that for slow fans improvements to blades or motor, barely do anything, so opposite airflow concept may be better tested with faster fans (It's possible to remove blades from Kazes and then put them on different sleeve bearing fans). At least we know that opposing flow doesn't hurt anything, so that's good.

BTW fan stacking has been tested and it doesn't improve airflow (but that can improve pressure):

Adding more fans may help cooler more, when you test it in case with restrictive airflow. And since pretty much every case restricts airflow to some degree, more fans are usually better. On open test bench, effect of 3 fans will not be seen properly.



You should really know better than that:

They are close, but Silver Arrow lost by a tiny speck.



Oh man, I will sure enjoy dethroning D15. Too bad that your list isn't good. Don't get me wrong, those heatsinks are well performing, but they lack a tiny bit to really match D15. Anyway, here's a list of all those coolers compared to D15:
Phanteks PH-TC14PE CPU Cooler Review - Results (Ascending order) (stock Phanteks won against stock D14, but lost in all out comparison and D14 is slightly worse cooler than D15)
Testing: Setup & Results - Thermalright Silver Arrow Review - Page 4 (some variation of silver arrow also lost to Prolimatech Super Mega)
CRYORIG R1 Ultimate Review (Cryorig R1 Ultimate didn't beat D15 either, it mostly matched it or exchanged places with D15)
Scythe Fuma 2 Review (Fuma 2 actually often beat Noctua by a bit, I will give you that. It's actually odd, because in many reviews it doesn't perform better than D15)
A CPU Cooler You NEED To Know About! Scythe Fuma 2 Review (And here we can see that Fuma 2 is a whole tier below D15, I have no idea why results vary so much. Seriously, in one review Fuma 2 was beating D15 and another it's losing by 15C. Somebody should investigate what is going on.)
Scythe FUMA 2 CPU Cooler Review (Fuma 2 performed slightly worse than D15, but once with equal fans it managed to won by miniscule amount and Fuma 2 was generally quieter than D15. So my verdict is that it matches D15.)
Thermalright Le Grand Macho RT CPU Cooler Review (Le Grand Macho RT performs like Fuma 2, loses to D15 with stock fan curve and wins once fans are normalized. I guess it matches D15, it's hard to say that it wins.)
Thermalright True Spirit 140 Power CPU Cooler Review (True Spirit 140 Power loses by quite a lot to D15)
Cooling - Thermalright's TRUE Spirit 140 "Power" - New Air Cooling King? Mini Review (however in another review True 140 Power had an edge)

And the conclusion is that, despite some review gremlins, most coolers performed like D15. There was no clear, definite winner against D15. D15 has not been meaningfully dethroned, but I wish it was. The main problem being is that all D15 competitors don't have particularly unique design or innovative features that would help them gain an edge. Therefore there isn't any truly meaningful change between them. Competitors will need to start using shrouds or triple tower configs, pretty much anything not already seen to go past current air cooling limits. I personally think that putting a shroud on air cooler would be the easiest and most logical step to improve their performance, as it wouldn't let any generated airflow to not flow outside of heatsink. That's far better idea than opposing fans as it solves the same problem, but far more effectively.
I don't believe it does give us a better feel of how cooler performs. Problem is when cooler is tested system specific we are not getting just the cooler's performance, but a combination of everything in system. I learned a lot when I was testing coolers for reviews. Many review sites don't don't want accurate test results. They only what cooler to fall into step with what advertising and other reviews sites have posted. They depend on being kind to all their sample supply sources by giving them all thumbs up reviews.

I've also built hundreds of systems over the years and I'm telling that if coolers give low temps on open bench they give low temps in cases when case fans are setup properly. End of story!

As for users having dedicated GPUs, sure most do. But most demanding GPU applications are gamers with a few graphic's people added on. They are the ones using high wattage GPUs. But even then the key is having case airflow optimized for components being used. Any of the top coolers will do similar job of cooling when case airflow is optimized as it should be.


Going off on stacking fans is a joke. Stacking increase pressure rating but only if airflow is straightened between fans. It has little to do with topic at hand. Difference between 1x, 2x and 3x fans on single and twin towers is insignificant .. At most 3c. Reason some reviews don't agree is because most reviewer are using flawed test procedures. Few monitor and record air temp entering cooler at same time as they record CPU temps in their tests .. and every degree warmer air is entering cool is degrees hotter the CPU temps is (at same load and fan speed).

If review isn't recording air temp into cooler test is flawed and basically irrelevant!! Especially if testing is done in a case built system!!

I'm not going to look at your links because it's a waste of time. I'm sure most if not all are not monitoring and recording air temp entering cooler when the record CPU temp.

No, Silver Arrow (1st gen), Silver Arrow SB-E (2nd gen), Silver Arrow IB-E (3d gen) all out perform D15 with same fans. Same goes for original NH-D14 being as good or slightly better than newer NH-D15. NH-D15 uses 1500rpm fans while all Silver Arrow variants except Extremes have 1300rpm fans with Extreme models having 2500rpm fans. Even with 200rpm slower fans they are within 1c, so you can imagine what they would do with same fans as D15 has. ;) Silver Arrow IB-E and XB-E Extreme wiht 2500rpm fans drops temps 5-10c lower than D15 temps. NH-D14 we tested against TRUE Spirit 140 Power was 9c cooler at 2500rpm than at 1200rpm on i7 920 @ 4.2GHz


Edit:
Le Grand Macho RT cools as well or better than D15 even with it's single 1300rpm fan. Review sites that do proper job all had it as good or better than D15
 

·
Registered
Joined
·
217 Posts
Discussion Starter · #192 ·
Most AIO suck due to the weak pumps inside of them.
I have a custom loop that replaced my older CM 212 evo+.
Well, I have read that most AIOs suck in comparison to proper loop due to amount of liquid they have. But obviously radiator area, pump and etc play their role.


These could keep an FX6300 [email protected] below 52°C(peak) after 2hours of y-cruncher stresstesting, the aircooler usually reached 58-60C.
I was kinda satisfied but i expected a bigger drop in temps after reading all the stories on watercooling.
I would have too. I tested Mugen 4 PCGH with one fan removed and it reached similar temps (middle 50s). When I was doing suicide runs, I clocked FX 6300 to 5.288GHz with as much voltage as board could provide (It was 1.72V, because board was Asrock 970 R2 Pro 3). For that I used CM Hyper 103 and despite crazy amount of electricity pulled from the wall (over 300 watts with only one module left enabled during Windows boot), it kept it below 60s. CM Hyper 103 is also capable cooler with moderate overclocks. The main reason why my suicide runs topped out at 5.288GHz is because VRMs were really hot, like over 110C hot (on previous board Gigabyte 78LMP-S2P rev.5, I have reached nearly 150C on VRMs and that left brown stain on it, motherboard was working fine for 2 years). It was for validation only and Windows was very unstable at that point, but anyway, for some reason FX chips are cooled rather well with just about anything.

I accidentally cracked the pump inlet on the phobya at some point and instead of getting another one, i did a lot of research on xtremesystems and i went for the swiftech mcp655(ddc laing)
This pump, at full speed(~4450rpm) on the same loop, could keep the CPU temp below 42°C
I removed one fan from the radiator and temps went up to 44°C, dropped pump speed to 1600rpm and temps went up to 48°C.(Turning off fans, running pump at 100% pushed temps to 52°C)
Wasn't that better pump very loud at such rpms?


Was the air-cooler more convenient? Yes and cheaper.
Are AIOs worth it? Get an air cooler.
Is a custom loop worth it? Not on value or convenience, but it's the only thing you can build with your hands for your PC.
Do AIOs and aircoolers perform the same? LOL, a proper loop is not comparable.
That's the spirit. With water you either go big or go home, nothing else makes sense.


The biggest difference on custom water vs Air/AIO is the gradual increase of temperature no matter the load - there are no spikes higher than 10°C.
With a bigger reservoir it can take an hour or more to reach peak temps, but it only takes a few minutes to cool back down, you can't achieve that with air, you can't hold the excess heat in one place.
Custom loop can be very fun, I think. If I ever wanted that, I would definitely would ghetto it first. I would certainly use car or moto radiator, fish tank pump, cheapo 2L or 5L drinking water bottle, PVC tubing and some random cheap block, perhaps something form Ali. From what I have read, this is basically how custom loops initially were. Very ghetto, but very capable. What is lovely about custom water cooling are opportunities to use things not intended for computer cooling. Hell, if you want you can use house radiator if you think that EK stuff is too small and use big 50L jug of water as reservoir if you think that EK is dumb. Basically unlimited choice of radiators also lets you to experiment with non computer fans. If you want you can also try different liquids like cooking oil instead of antifreeze. There's so much fun to be had with crazy ideas, too bad that all of them make PC useless as daily PC and you will never be putting side panel on, but it's worth it. However if one only wants to put custom loop in case and be sane, they may as well just get a 360mm or 280mm AIO.
 

·
Premium Member
Joined
·
27,527 Posts
Well, I have read that most AIOs suck in comparison to proper loop due to amount of liquid they have. But obviously radiator area, pump and etc play their role.



I would have too. I tested Mugen 4 PCGH with one fan removed and it reached similar temps (middle 50s). When I was doing suicide runs, I clocked FX 6300 to 5.288GHz with as much voltage as board could provide (It was 1.72V, because board was Asrock 970 R2 Pro 3). For that I used CM Hyper 103 and despite crazy amount of electricity pulled from the wall (over 300 watts with only one module left enabled during Windows boot), it kept it below 60s. CM Hyper 103 is also capable cooler with moderate overclocks. The main reason why my suicide runs topped out at 5.288GHz is because VRMs were really hot, like over 110C hot (on previous board Gigabyte 78LMP-S2P rev.5, I have reached nearly 150C on VRMs and that left brown stain on it, motherboard was working fine for 2 years). It was for validation only and Windows was very unstable at that point, but anyway, for some reason FX chips are cooled rather well with just about anything.


Wasn't that better pump very loud at such rpms?



That's the spirit. With water you either go big or go home, nothing else makes sense.



Custom loop can be very fun, I think. If I ever wanted that, I would definitely would ghetto it first. I would certainly use car or moto radiator, fish tank pump, cheapo 2L or 5L drinking water bottle, PVC tubing and some random cheap block, perhaps something form Ali. From what I have read, this is basically how custom loops initially were. Very ghetto, but very capable. What is lovely about custom water cooling are opportunities to use things not intended for computer cooling. Hell, if you want you can use house radiator if you think that EK stuff is too small and use big 50L jug of water as reservoir if you think that EK is dumb. Basically unlimited choice of radiators also lets you to experiment with non computer fans. If you want you can also try different liquids like cooking oil instead of antifreeze. There's so much fun to be had with crazy ideas, too bad that all of them make PC useless as daily PC and you will never be putting side panel on, but it's worth it. However if one only wants to put custom loop in case and be sane, they may as well just get a 360mm or 280mm AIO.
You need to start doing instead of parroting what you read. Then you will learn that most of what you read is untruth. There is only a scattering of truth out there.

That said, CLC pump have flowrates about equal to what a healthy young adult urinates .. about 60 liter per hour is peak flowrate. AIO's that are not CLC have 100-600L/h flowrates. Most commonly used pump in custom loops is EK-D5 rated 1500L/h.

Compound the above with facts like CLCs all have aluminum radiators .. while AIOs and custom all use copper radiators.
CLCs all have factory sealed fittings with no provision to be taken apart .. while all AIOs and custom loop have threaded fittings and fill ports usually with a reservoir as well.

CLCs were orignally patented as LCLC (Low Cost Liquid Cooling). Name was later changed to CLC. Marketing hype used "water cooling" as hype working to draw in peeps who couldn't build their own custom loops .. either because of lack of skills or lack of money.

CLCs are made with cheapest possible components to kinda work .. AIOs are much better but not even close to quality of custom loop components.
 

·
Rig Advisor
Joined
·
3,492 Posts
This was a good read, but I think we can all agree, we are here now.

2489190
 

·
Premium Member
Joined
·
15,117 Posts
Yes, repeat the same thing that GN, Linus and all the others have said repeatedly... except getting it wrong on the most important parts. xD (I kid! Yay dry humor!)

AIOs typically do worse, yes, but handle heat "spikes" far better due to the nature of the loop. Water can soak up more heat so it acts like a heatbank rather than just a heatsink, letting you accumulate more while it slowly builds up to maximum. At that point it's more a matter of how thick the radiator is and how powerful the pump is.

I'll take a 120mm rad AIO over a tower cooler any day of the week.
Even a water tower cooler? 🌨
 

·
Registered
Joined
·
217 Posts
Discussion Starter · #196 ·
I don't believe it does give us a better feel of how cooler performs. Problem is when cooler is tested system specific we are not getting just the cooler's performance, but a combination of everything in system. I learned a lot when I was testing coolers for reviews. Many review sites don't don't want accurate test results. They only what cooler to fall into step with what advertising and other reviews sites have posted. They depend on being kind to all their sample supply sources by giving them all thumbs up reviews.
Well, that's a bit of conspiracy. Sure somebody wants money, but the problem is that there are some independent reviewers that only care about proper reviews. Bribing some just wouldn't matter if they kept talking nonsense long enough. And on YouTube or internet in general, you can't just bribe everyone. There will always be someone independent.

I've also built hundreds of systems over the years and I'm telling that if coolers give low temps on open bench they give low temps in cases when case fans are setup properly. End of story!
Still, not as good temps as in open air bench.


As for users having dedicated GPUs, sure most do. But most demanding GPU applications are gamers with a few graphic's people added on. They are the ones using high wattage GPUs. But even then the key is having case airflow optimized for components being used. Any of the top coolers will do similar job of cooling when case airflow is optimized as it should be.
No matter how you set it up, GPU air will inevitably go through air cooler. Can't be helped, unless your case has ducting that separates them and I haven't seen anything like that.

Going off on stacking fans is a joke. Stacking increase pressure rating but only if airflow is straightened between fans. It has little to do with topic at hand. Difference between 1x, 2x and 3x fans on single and twin towers is insignificant .. At most 3c.
Sure it's a joke, but it shows what happens in air coolers with more fans and why help only help so much.


Reason some reviews don't agree is because most reviewer are using flawed test procedures. Few monitor and record air temp entering cooler at same time as they record CPU temps in their tests .. and every degree warmer air is entering cool is degrees hotter the CPU temps is (at same load and fan speed).
I see many reviewers monitoring ambient temperature too and some comparing deltas.


I'm not going to look at your links because it's a waste of time. I'm sure most if not all are not monitoring and recording air temp entering cooler when the record CPU temp.
That's your choice then. My reasoning tells me that there isn't significantly better cooler than all of what you listen and D15. And there really isn't any reason why there should be. Those coolers are all very similar, with similar fans and similar tower form factor.

No, Silver Arrow (1st gen), Silver Arrow SB-E (2nd gen), Silver Arrow IB-E (3d gen) all out perform D15 with same fans.
Haven't seen any strong evidence to support that. They perform kinda the same or Silver Arrow wins by fractions of degree. That's a bit too weak.

Same goes for original NH-D14 being as good or slightly better than newer NH-D15.
Disagree again. D15's heatsink is bigger. It is 1 cm wider than D14's. Also D14 comes with one smaller fan. There's no way D14 is better than D15.


NH-D15 uses 1500rpm fans while all Silver Arrow variants except Extremes have 1300rpm fans with Extreme models having 2500rpm fans. Even with 200rpm slower fans they are within 1c, so you can imagine what they would do with same fans as D15 has. ;)
They will basically match Noctua. 200 rpms in fast fans mean almost nothing.

Silver Arrow IB-E and XB-E Extreme with 2500rpm fans drops temps 5-10c lower than D15 temps.
I don't see any evidence to support those claims. Here's a review of Silver Extreme:

Stock it is beating Noctua D14 by 10 degrees. However, D15 is bigger cooler with more capable fans, so let's reduce 2C due to better heatsink and 3C due to fans and we get 5C lead for Silver Arrow.


Passive performance shows how good heatsinks are alone and there's only 2C difference between them. So most gains in Silver Arrow are from those crazy fast fans. Perhaps I should reconsider Silver Arrow Extreme as air cooler king, but 5-10C lower temps just aren't happening. Once fans are normalized, Silver Arrow Extreme is only a bit better than Noctua.

NH-D14 we tested against TRUE Spirit 140 Power was 9c cooler at 2500rpm than at 1200rpm on i7 920 @ 4.2GHz
Post a link. There's no way it beats dual tower cooler with only slim single tower.

Le Grand Macho RT cools as well or better than D15 even with it's single 1300rpm fan. Review sites that do proper job all had it as good or better than D15
You better show from where you are pulling your results from, because they surely don't match majority.
 

·
Fix it till its broke
Joined
·
947 Posts
To be fair TS140P has 6x 8mm pipes, it is slim, but it is tall af. The cold plate will make or break it. Mine is ok, but its not the greatest. Works great on Intel. On LGMRT I could run linpack extreme on my 3770K @ 4500MHz with no fan attached to the cooler, and case fans on low.. 5v.. Its a good cooler. I could also run my 3600XT @ 4400MHz with no fan. Noctua is just overhyped, they aren't bad.
 

·
Premium Member
Joined
·
15,117 Posts
You better show from where you are pulling your results from, because they surely don't match majority.
FrostyTech is good, if still around.

This was one of the best air cooler designs:
The same cooler can be found made by other companies, just remember what it looks like and score 1 if you can find one.

There was this 1 mainly PSU vendor that made a really good cooler, too.
It ruled the top for a long time.
(Something)_ TS-140
 

·
Premium Member
Joined
·
27,527 Posts
Your review of Silver Arrow Extreme is most likely in a case with same fans as used with D15 and other coolers .. which means case is only flowing about half as much airflow as Extreme TY-143 fans are flowing at full speed .. so of course cooler is forced to re-use it's heated exhaust air that is 7-15c hotter than room .. so resulting CPU temps are same 7-15c higher.

I've sspent way too many hours testing case airflow and coolers to believe hype your are reading and believe is truth. A system that has been setup with proper case airflow will have no ,more than 3c warmer than room going into CPU cooler even when GPU is at 100% load. But that is not a case with stock case fans (which are 95% of the time almost worthless and place is wrong places as well) .. and without monitoring air temp 2.5cm in front of CPU & GPU cooler during stress testing user has no idea what coolers are doing .. because they have no base line air temp .. and using room ambient taken from thermometer somewhere else in room doesn't work!
 

·
Laptop Enthusiast
Joined
·
9,399 Posts
This thread is still going on? Good lord. Like... it's long since deviated from the original topic.

Benefit of AIO: Heatsoak/heatbank for shorter load and/or temp spikes. (Thermal density yo) Because no fan curve can react quickly enough to handle some heat spikes from either Intel or AMD.

Benefit of Tower Cooler: Rapid heatsinking. Pushes the heat away faster if done correctly.


There, we're done. Turn a page. xD
 
181 - 200 of 278 Posts
Top