Looking for suggestions for heatsinks to review
- Thread starter b1lk1
- Start date
- Status
Prolimatech's Armageddon would be a good place to start since their Megahalem aka Megatron heatsink was the first new product to come close to and let alone beat a TRUE. Thermalright's Cogage Arrow and maybe Noctua's NH-D14 would be a good addition as well although the Noctua product has been out for a while now I believe. Oh, I almost forgot. Coolink's Corator too.
Now those are all $60+ heatsinks so what about something that most people can actually afford? There aren't too many mainstream heatsinks out there with really good performance let alone new offerings to the market segment.
You could always go with the good ol' Scythe Mugen 2. It has been out for a long time but gives great cooling at $36 depending on the online retailer you go through. Xigmatek released a refresh of their Dark Knight cooler with 1156/1366 compatibility right out of the box so that could be an option as is their Balder heatsink.
It's strange that companies like Zerotherm have been pretty much silent for the last year or so after release great coolers like their Nirvana.
Now those are all $60+ heatsinks so what about something that most people can actually afford? There aren't too many mainstream heatsinks out there with really good performance let alone new offerings to the market segment.
You could always go with the good ol' Scythe Mugen 2. It has been out for a long time but gives great cooling at $36 depending on the online retailer you go through. Xigmatek released a refresh of their Dark Knight cooler with 1156/1366 compatibility right out of the box so that could be an option as is their Balder heatsink.
It's strange that companies like Zerotherm have been pretty much silent for the last year or so after release great coolers like their Nirvana.
Last edited:
Psi*
Tech Monkey
Maybe a unique-ness without too much complication could be including 2 different TIMs ... a common "garden variety" that many would use versus a metal alloy?
Add lapping, at the end of course with the "garden variety" TIM. And do this with 2 or more i7 920s for comparison to the metal alloy TIM. My expectation is that they would be similar ... although that is just a guess.
Just reading what is going on other forums, there is still a lot of noise about lapping. I wonder if a metal alloy might accomplish the same thing. And I have never seen any review including lapping ... certainly a PITA, I understand. I mention the i7 920 as it is current and can be had for ~$200 at the right places.
Simpler, would be just 2 or 3 i7s with the identical TIM & heatsink ... I don't assume that all i7s are made equal, but it seems that is a common assumption.
Add lapping, at the end of course with the "garden variety" TIM. And do this with 2 or more i7 920s for comparison to the metal alloy TIM. My expectation is that they would be similar ... although that is just a guess.
Just reading what is going on other forums, there is still a lot of noise about lapping. I wonder if a metal alloy might accomplish the same thing. And I have never seen any review including lapping ... certainly a PITA, I understand. I mention the i7 920 as it is current and can be had for ~$200 at the right places.
Simpler, would be just 2 or 3 i7s with the identical TIM & heatsink ... I don't assume that all i7s are made equal, but it seems that is a common assumption.
b1lk1
Tech Monkey
Multiple TIMs would make things difficult. Same with lapping. I use one set type of TIM and lapping is just not done by the majority of users. I prefer to use the actual heatsink as they are pulled from the package.
As for methodology and CPU, we do realize that all are not equal, but they are all well within a small percentage of each other and the average user is likely to get close to our results with similar hardware.
We will be using the i5 platform as it is the more mainstream and most common to be used by the majority of readers. You are not going to see a large difference between a 4GHz i5 and 4GHz i7. I do not mean to state that they are the same, but we are talking less than 10-15% in total heat and I guarantee I will run my CPU hotter and longer than 99.99% of the world feels comfortable doing.
As for methodology and CPU, we do realize that all are not equal, but they are all well within a small percentage of each other and the average user is likely to get close to our results with similar hardware.
We will be using the i5 platform as it is the more mainstream and most common to be used by the majority of readers. You are not going to see a large difference between a 4GHz i5 and 4GHz i7. I do not mean to state that they are the same, but we are talking less than 10-15% in total heat and I guarantee I will run my CPU hotter and longer than 99.99% of the world feels comfortable doing.
I have a Replacement CPU Cooling Fan Heatsink for AMD 775. This CPU Fan Cooler works excellent in heat dissipation and extends the life and functionality of computer. Also It can keep my processor cool with optimal temperature and improve air flow for CPU.
eunoia
Partition Master
In the "top air" category, definitely the new Prolimatech. For us more "frugal" types, how about finding one of the new Intel stock ones that are shipping with the i7 980X? There's bound to be deals on those, I'm guessing a good percentage of those buyers will be giving these away. Does it compete with a Cooler Master Hyper 212?
Last edited:
b1lk1
Tech Monkey
We don't test by classes, we just put a bunch together and make them go head to head. I have noted what has come our way so far, keep it coming!
PS: As for the fans, we like to use them as they come out of the box. If time permits, I wil be open to using one good fan for static testing, but that will exclude some models that come with un-replaceable fans.
PS: As for the fans, we like to use them as they come out of the box. If time permits, I wil be open to using one good fan for static testing, but that will exclude some models that come with un-replaceable fans.
b1lk1
Tech Monkey
Some sure are, but Noctua uses some great fans and there are other companies that also use very good fans.
Anything that uses a standard sleeve bearing gets put on my blacklist, they're great for a couple weeks and then for no reason they start to whine, wobble, rattle and just generally fail in the most unpleasant manor possible. I don't think i've had a sleeve fan last more than 6 months.... and i've had a lot. 10k hours at 25C.... 25 degrees, what a joke.
Dual ball bearing is a minimum for me, but my preference is either FD's (Fluid Dynamic) or Ceramic, yet to have any of those fail.
One thing i wouldn't mind trying is building a custom cooler. I've got a mini-fridge with a TEC (peltier) in it that can cause massive amounts of ice to build up on the inside, i'm just wondering if it can sustain that chill when bolted to a 200 watt heating element. I've seen TEC's in some very expensive coolers and all they do is make a lot of noise since the manufacturers are so conservative with the power (i know, TEC's can go below freezing, condensation, water, frying electricals, etc). TEC's are cheap, but if they can get a CPU down to about 12C, it'll probably be worth it.... just wonder how to go about it, but thats for another time (The whole time delay to reach the cool temps is a concern).
Dual ball bearing is a minimum for me, but my preference is either FD's (Fluid Dynamic) or Ceramic, yet to have any of those fail.
One thing i wouldn't mind trying is building a custom cooler. I've got a mini-fridge with a TEC (peltier) in it that can cause massive amounts of ice to build up on the inside, i'm just wondering if it can sustain that chill when bolted to a 200 watt heating element. I've seen TEC's in some very expensive coolers and all they do is make a lot of noise since the manufacturers are so conservative with the power (i know, TEC's can go below freezing, condensation, water, frying electricals, etc). TEC's are cheap, but if they can get a CPU down to about 12C, it'll probably be worth it.... just wonder how to go about it, but thats for another time (The whole time delay to reach the cool temps is a concern).
Last edited:
I'll give another vote for the Noctua NH-D14. It's the best performing cooler I have seen to date, and it comes close enough to my watercooling loop that once my CPU situation is resolved, I plan to do plenty of load testing... The Noctua is amazing in both performance and installation.
Tharic-Nar, If you plan to build your own TEC then just make sure you can cool the reverse side of the peltier... it must be capable of cooling the pelt's own heat in addition to the heat the pelt absorbs from the CPU. If they get too hot they not only lose efficiency but also performance.
I almost built one myself as there are some very interesting DIY modders out there that have guides, but the problem is they are extremely power inefficient, take a great deal of work, and for the cost of 24/7 use it would be cheaper to buy a cascade or sub-zero system that delivers much better performance anyway. I elected to stick with non-chilled watercooling instead. A peltier is not a cooler, all it does is require an even more powerful cooler be attached to the hot side that can handle the pelts own heat along with the CPU's heat load. A real 200W pelt + 200W CPU would mean 400W of heat you'd need to cool... most pelts greatly overestimate the wattage rating though from what I've seen. Honestly I'd recommend something like this instead: http://www.frozencpu.com/products/9725/ex-vap-19/OCZ_Cryo-Z_Phase_Change_Single_Evaporator_CPU_Cooling_Unit_HOT_ITEM_OCZTCRYO_Sockets_478_754_775_1156_1366_939_940_AM2_AM2_AM3.html
Tharic-Nar, If you plan to build your own TEC then just make sure you can cool the reverse side of the peltier... it must be capable of cooling the pelt's own heat in addition to the heat the pelt absorbs from the CPU. If they get too hot they not only lose efficiency but also performance.
I almost built one myself as there are some very interesting DIY modders out there that have guides, but the problem is they are extremely power inefficient, take a great deal of work, and for the cost of 24/7 use it would be cheaper to buy a cascade or sub-zero system that delivers much better performance anyway. I elected to stick with non-chilled watercooling instead. A peltier is not a cooler, all it does is require an even more powerful cooler be attached to the hot side that can handle the pelts own heat along with the CPU's heat load. A real 200W pelt + 200W CPU would mean 400W of heat you'd need to cool... most pelts greatly overestimate the wattage rating though from what I've seen. Honestly I'd recommend something like this instead: http://www.frozencpu.com/products/9725/ex-vap-19/OCZ_Cryo-Z_Phase_Change_Single_Evaporator_CPU_Cooling_Unit_HOT_ITEM_OCZTCRYO_Sockets_478_754_775_1156_1366_939_940_AM2_AM2_AM3.html
I wouldn't call Noctua's fans proprietary. They can be swapped out if you want something really foolish like a Panaflo Delta. I was thinking more along the lines of Arctic Cooling's fan on the Freezer Pro or just about anything from Zalman's XX00 series.
I'd vote for Thermalrights. Even when not on top, they're always hovering nearby.
BTW, why don't you guys get a proper heater (or a row of them, for seasoning the grease) instead of introducing randomness with actual CPU?
Things can be done to last or just to sell, technology isn't the key when marketing kicks in. I'd rather go for sleeve from Noctua then hyper-dynamic-fantastic-ceramic from Arctic Cooling.
That's what killed Peltier's in OC, the ones small enough to fit CPU just doesn't have the throughput to pump 100W to the other side. Kougar pointed out the rest
BTW, why don't you guys get a proper heater (or a row of them, for seasoning the grease) instead of introducing randomness with actual CPU?
Well, the difference between sleeve and dynamic bearings is mostly naming, as there are hardly any (in computer fans) that doesn't operate in both regimes.
Things can be done to last or just to sell, technology isn't the key when marketing kicks in. I'd rather go for sleeve from Noctua then hyper-dynamic-fantastic-ceramic from Arctic Cooling.
No, it can't.
That's what killed Peltier's in OC, the ones small enough to fit CPU just doesn't have the throughput to pump 100W to the other side. Kougar pointed out the rest
Psi*
Tech Monkey
THIS!++
A "real" CPU just doesn't make so much sense anymore ... if you are reviewing HSFs. This makes the process more scientific versus wondering how some fav benchmark program is stressing a processor. What I really want to know is how good a HSF is in removing heat & I don't care about how the heat is generated. I think this makes more sense given the multi-mounts some HSF have.
I don't review heatsinks for Techgage, but besides the obvious difficulties how would you mount a heater to the base of the cooler? Especially so it does not actually heat the cooler itself and only transfers heat to the base? Especially with top-down coolers, that traditionally just don't perform as well as tower coolers. The heater would also need a built-in ability for measuring temperatures, measuring the edges of the cooler base would not suffice as these are not designed to actually provide cooling.
Another issue is heatpipe placement, heatpipe type (there are at least 4 types), and the CPU used can make a huge difference. When Intel CPU's first featured two cores, they used two die under the IHS... so at the time coolers that were build for a single die in the middle of the IHS did not cool them as well as they should have.
The last issue is some users prefer temperature readings inside an actual case. For instance if you use a top-down cooler inside a case, the results are going to be lower than what you'd see with an open-air testbed because top-down coolers simply blow hot air all around the inside of the case, while tower coolers usually direct most of it out the back exhaust fan. It's amazingly complicated how heatsink testing can be.
- Status