Arctic Freezer 33 eSports ONE CPU Air Cooler Review @ [H]

FrgMstr

Just Plain Mean
Staff member
Joined
May 18, 1997
Messages
55,532
Arctic Freezer 33 eSports ONE CPU Air Cooler Review

The Arctic Freezer 33 eSports ONE is a CPU air cooler that packs some very healthy stats into what is a diminutive package if space is of a concern. Its price is also diminutive at only $30. The Freezer 33 touts a thermal coating that creates "micro-turbulences" in order to give it optimised heat-dissipation and increased cooling performance.

If you like our content, please support HardOCP on Patreon.
 
Last edited:
Huge disappointment from Arctic Freezer. They were my go to brand for basic CPU cooling when I was building pcs. I basically always replaced the stock cooler with their Arctic Freezer Pro lineup since it was so cheap and a huge upgrade if a client didn't want to spend a lot on a cooler.
 
Maybe the eSport to which it refers is trying to avoid thermal throttling while playing a game.
 
I got this on my current i7-2600 (soon to be i5-8400) and it's preforming admirably. Prime95 small FFT for 10 minutes and it barely got to 60C. Maybe they'd do better by lowering the max supported TDP.

EDIT: Or like others are saying, your unit was possibly defective.
 
Last edited:
Really surprised that Arctic dropped the ball on this one -- was expecting above average to good results. Maybe the heatpipes were defective (as doubletake stated above)?
 
That's weird... I got an arctic freezer 3 "non-esports" working pretty good on a ryzen [email protected] or 3.9

curious if the black coating is an issue. well that and a 1400 is only half a 1700 lol.

anyway, makes me not feel so bad about not getting the 'prettier' one. thanks for the review!
 
Wow, those are terrible temperatures! Even performs worse than the AMD cooler on low fan speed! I feel stupid.. I've been trumpeting about how good of a cooler it is. Any idle temps on the big black brick?
 
Really surprised that Arctic dropped the ball on this one -- was expecting above average to good results. Maybe the heatpipes were defective (as doubletake stated above)?

you shouldn't be surprised.. lately i doubt a lot of [H] Cooler reviews, a lot of coolers which are very well known of offer wonderful results here are getting smashed and perform like trash or coolers with tend to perform on the average mid range results perform like top dogs.. only site that tend to happen. And it may be something related to the platform they are using without figure it out, i know of some motherboard issues with certain coolers due the thick/thinness of the motherboard PCB layers so it affect mounting pressure anyway, results here shouldn't be taken in consideration when purchasing a new cooler as they even haven't took the proper time to figure what's going on with their test.
 
thanks Kyle o7

yep, that is epicly pathetic performance, that base plate looks nifty as hell, but seriously, even CM that over rates their TDP ability still at least manages to have acceptable cooling and it actually WORKS..that is just shoddy ass performance Arctic
either they royally fucked up, or it was only designed for very specific cpu to be in the TDP rating they used, claiming to be able to deal with up to 200w or something like that and not even coming close to doing so is terrible as hell.

their "quality" has taken a massive noise dive that much is for damn sure, such a shame because it looks "nice enough"
expect terribad performance from a cooler that costs $10 or most boxed cooler, not from one that claims something is cannot deliver and yet pricing similar to say hyper 212 or deepcool gammaxx 300/400 (all priced similar)
 
I'm unable to recall in my 30 plus years of system building a HSF failing so miserably. I've installed a dumptruck load of custom coolers from the likes of Thermalright, Thermaltake, Artic Cooling, Corsair to name a few.

From the results in the article to me it seems there is an issue with the heatpipes not functioning properly or possibly an issue with poor to no contact between the heatpipes and fins. Whatever the reason, the results certainly don't live up to marketing hype. I'd be curious to see up close shots between the fins to see if one of my hunches was on target.

Thanks for the review [H]. Greatly appreciated.
 
you shouldn't be surprised.. lately i doubt a lot of [H] Cooler reviews, a lot of coolers which are very well known of offer wonderful results here are getting smashed and perform like trash or coolers with tend to perform on the average mid range results perform like top dogs.. only site that tend to happen. And it may be something related to the platform they are using without figure it out, i know of some motherboard issues with certain coolers due the thick/thinness of the motherboard PCB layers so it affect mounting pressure anyway, results here shouldn't be taken in consideration when purchasing a new cooler as they even haven't took the proper time to figure what's going on with their test.
Please let me know what exactly you think is wrong with our testing....specifically? We love feedback and criticism, but you are painting with such a broad brush, it does little to help us pinpoint where you think our issues are.
 
Sorry, I read the article, and I see that there's good contact between cpu and heatsink, but there's no mention about mounting pressure. I guess there's no problem regarding that.
But it's very strange that even the stock cooler beats this, which have larger surface and fan.
Without good mounting pressure, you would not get a good mate as pictured. Mounting pressure and heat are what give you the TIM flow that is needed for the footprints shown. We use that PK-3 TIM as well. It is incredibly viscous and will show you mating issues very quickly.
 
curious if the black coating is an issue.

Interesting thought. Hey Kyle, can you scratch off any of the coating? Is it just paint? Can you anodize copper? (google says yes, but I've never heard of anyone doing it)
 
Interesting thought. Hey Kyle, can you scratch off any of the coating? Is it just paint? Can you anodize copper? (google says yes, but I've never heard of anyone doing it)
rgMekanic has the cooler in his possession. He is more than free to check it out if he wants and report back.
 
That's weird... I got an arctic freezer 3 "non-esports" working pretty good on a ryzen [email protected] or 3.9

curious if the black coating is an issue. well that and a 1400 is only half a 1700 lol.

anyway, makes me not feel so bad about not getting the 'prettier' one. thanks for the review!

I have the Arctic Freezer 33 eSports Edition. I think the review has the product title incorrect. It's 33, not 3. The difference between the eSports ONE and Edition models is the Edition version comes with 2 Bionix fans for a push pull. I really just got the dual fan version hoping to get extra flimsy fan clips and a matching fan to use as a case exhaust fan, which is what I did. So in effect I have an Arctic Freezer 33 eSports ONE installed as well. I was very surprised at the FAIL and overall results of the review, but here's my thoughts...

I run mine on a stock Ryzen 2 2600 on an MSI B450 Gaming Pro Carbon AC mainboard. My goal was aesthetics, decent cooling, and most important silence. I definitely achieved my goal. I'm running about 40c idle and 63c under Prime 95 small FFT. 63c on CPU load remains just that even while kicking the GTX 960's ass with MSI Kombuster software. I run this in a Fractal Design Meshify C White case. I added 2 14cm Bionix fans to the front for intake, and the extra Bionix fan from my CPU cooler package (12cm) as an exhaust. My total fans topping out at 6; 2x140mm (front intake) 2x120mm (one on cooler one on exhaust), leaving my video cards GPU cooler's lone fan and the Season PSU fan.

The machine is QUIET, fast, and pleasing to me to look at. This cooler gets a thumbs up for me for my application. Now it'd be quick for me to be skeptical of HardOCP's results being they are so far away from my experience with this cooler, but here are some facts. I am using the AMD Ryzen Balanced Power Plan for Windows 10 that currently comes packaged with the latest B450 chipset drivers you can get from AMD's website. With this power plan for some odd reason my 2600 doesn't meet it's maximum boost clock (3900mhz) with small FFTs in Prime95. It hits the max boost speed just fine in games though. I'm not sure if it's some AMD Cool & Quiet algorithm, something pertaining to the Ryzen Power Plan or what. You guys in the [H] forum please let me know your thought on using the AMD Ryzen plan as opposed to the built in Windows 10 High Performance plan. I've already ruled out the Balanced Windows plan as the core speeds fluctuate too much and games stutter for an overall horrible experience. I run my memory in XMP mode @3200mhz btw, before I forget to mention.

Now I am using no where near the 1.475v they are using in the [H] test. I'm below 1.2v even during Prime95 max power consumption test. I mean one of the selling points of the 2600 for me was the low TDP rating and while not exactly, that correlates to lower temperatures usually. That helps me meet my goal of silence. While this cooler has been great for me, looks like HardOCP puts a heat load on this thing it simply cannot handle. I cannot discount their findings because I am using this cooler in a totally different eco system. The overall surface area of the Freezer 33 eSports looks to be less than other large CPU coolers. Also the base does not completely cover the Ryzen IHS. Yes it covers the general die area underneath the IHS but I've sure there's plenty of heat on the perimeter of the IHS that would love to have somewhere to go quickly.

My idle and load temps in this build of 40c and 63c respectively were recorded during a hot day (90 Fahrenheit) windows open at my house. I like to test in worst case scenarios hehe. The temperatures were maintained with the following fan curve settings I created in my BIOS with the main focus on silent as possible for as long as possible:

Front intake 140mm x2
40c 25%RPM
55c 30%RPM
70c 50%RPM
85c 100%RPM

I matched exhaust and CPU curves 120mm x2
40c 25%RPM
55c 40%RPM
70c 50%RPM
85c 100%RPM

My 960GTX cards fan doesn't even spin up unless the GPU gets around 60-65c so it's virtually silent and the Seasonic PSU makes no noise.

So that is my report on this cooler as a new owner of it. I have no reason to doubt HardOCP as I find this site to be very straightforward and honest. I think the results for me are to be expected after my diligent research and the results for HardOCP are to be expected if you are going to make this stack of aluminum work [H]ard. Hey, they want to say their product supports up to 200watts, someones gonna call your bluff. Looks like HardOCP did that. Just not enough mass on the cooler perhaps?

I'd love to hear other peoples opinions regarding the different Power plans for Windows 10. Also I'd like to hear some opinions regarding overclocking of the 2600, because I've elected not to. It's my understanding if you OC these modern CPUs your core clocks will stay at those high speeds and I like the clocks coming down when the load is off. I prefer the cool and quiet, but still having some oomph these days, I'm older and overclocking the hell out of everything is in my past now heh. One more tidbit about the Arctic Freezer 33 eSports is I am scared I may have screwed it down too hard. My AM4 adapter brackets slightly bend. I figured the provided screws were of the shoulder-bolt type for that perfect fit and tightened with that in mind, still looks like it may be too much, I don't know... what do you guys think? Everything is running fine (except for Winblows 10, oh Linux Mint please support games).

I'll try to provide a pic of my system here now. It is an awful pic from my phone, looks so much better in person:

MSI.jpg

Thank you HardOCP for reviewing products I am interested in, even after I just bought them haha!
 
Last edited:
Wow, what a turd. My goto brand for CPU cooling made this?! The ArticFreeezer i30 was such a great product. 212 for future Intel builds it is.
 
I have the Arctic Freezer 33 eSports Edition. I think the review has the product title incorrect. It's 33, not 3. The difference between the eSports ONE and Edition models is the Edition version comes with 2 Bionix fans for a push pull. I really just got the dual fan version hoping to get extra flimsy fan clips and a matching fan to use as a case exhaust fan, which is what I did. So in effect I have an Arctic Freezer 33 eSports ONE installed as well. I was very surprised at the FAIL and overall results of the review, but here's my thoughts...

I run mine on a stock Ryzen 2 2600 on an MSI B450 Gaming Pro Carbon AC mainboard. My goal was aesthetics, decent cooling, and most important silence. I definitely achieved my goal. I'm running about 40c idle and 63c under Prime 95 small FFT. 63c on CPU load remains just that even while kicking the GTX 960's ass with MSI Kombuster software. I run this in a Fractal Design Meshify C White case. I added 2 14cm Bionix fans to the front for intake, and the extra Bionix fan from my CPU cooler package (12cm) as an exhaust. My total fans topping out at 6; 2x140mm (front intake) 2x120mm (one on cooler one on exhaust), leaving my video cards GPU cooler's lone fan and the Season PSU fan.

The machine is QUIET, fast, and pleasing to me to look at. This cooler gets a thumbs up for me for my application. Now it'd be quick for me to be skeptical of HardOCP's results being they are so far away from my experience with this cooler, but here are some facts. I am using the AMD Ryzen Balanced Power Plan for Windows 10 that currently comes packaged with the latest B450 chipset drivers you can get from AMD's website. With this power plan for some odd reason my 2600 doesn't meet it's maximum boost clock (3900mhz) with small FFTs in Prime95. It hits the max boost speed just fine in games though. I'm not sure if it's some AMD Cool & Quiet algorithm, something pertaining to the Ryzen Power Plan or what. You guys in the [H] forum please let me know your thought on using the AMD Ryzen plan as opposed to the built in Windows 10 High Performance plan. I've already ruled out the Balanced Windows plan as the core speeds fluctuate too much and games stutter for an overall horrible experience. I run my memory in XMP mode @3200mhz btw, before I forget to mention.

Now I am using no where near the 1.475v they are using in the [H] test. I'm below 1.2v even during Prime95 max power consumption test. I mean one of the selling points of the 2600 for me was the low TDP rating and while not exactly, that correlates to lower temperatures usually. That helps me meet my goal of silence. While this cooler has been great for me, looks like HardOCP puts a heat load on this thing it simply cannot handle. I cannot discount their findings because I am using this cooler in a totally different eco system. The overall surface area of the Freezer 33 eSports looks to be less than other large CPU coolers. Also the base does not completely cover the Ryzen IHS. Yes it covers the general die area underneath the IHS but I've sure there's plenty of heat on the perimeter of the IHS that would love to have somewhere to go quickly.

My idle and load temps in this build of 40c and 63c respectively were recorded during a hot day (90 Fahrenheit) windows open at my house. I like to test in worst case scenarios hehe. The temperatures were maintained with the following fan curve settings I created in my BIOS with the main focus on silent as possible for as long as possible:

Front intake 140mm x2
40c 25%RPM
55c 30%RPM
70c 50%RPM
85c 100%RPM

I matched exhaust and CPU curves 120mm x2
40c 25%RPM
55c 40%RPM
70c 50%RPM
85c 100%RPM

My 960GTX cards fan doesn't even spin up unless the GPU gets around 60-65c so it's virtually silent and the Seasonic PSU makes no noise.

So that is my report on this cooler as a new owner of it. I have no reason to doubt HardOCP as I find this site to be very straightforward and honest. I think the results for me are to be expected after my diligent research and the results for HardOCP are to be expected if you are going to make this stack of aluminum work [H]ard. Hey, they want to say their product supports up to 200watts, someones gonna call your bluff. Looks like HardOCP did that. Just not enough mass on the cooler perhaps?

I'd love to hear other peoples opinions regarding the different Power plans for Windows 10. Also I'd like to hear some opinions regarding overclocking of the 2600, because I've elected not to. It's my understanding if you OC these modern CPUs your core clocks will stay at those high speeds and I like the clocks coming down when the load is off. I prefer the cool and quiet, but still having some oomph these days, I'm older and overclocking the hell out of everything is in my past now heh. One more tidbit about the Arctic Freezer 33 eSports is I am scared I may have screwed it down too hard. My AM4 adapter brackets slightly bend. I figured the provided screws were of the shoulder-bolt type for that perfect fit and tightened with that in mind, still looks like it may be too much, I don't know... what do you guys think? Everything is running fine (except for Winblows 10, oh Linux Mint please support games).

I'll try to provide a pic of my system here now. It is an awful pic from my phone, looks so much better in person:

View attachment 103280

Thank you HardOCP for reviewing products I am interested in, even after I just bought them haha!
Great post, thanks for sharing. Yeah, we are worlds apart when it comes to Ryzen loads and voltages.

Some people seem to be getting their feelings hurt that this cooler did not pass our testing. Quite frankly, I did not expect it to make it through the process. Not the first time we have seen a HSF fail either. It has still amazed me that the True Spirit and Reeven Justice made it through the tests.
 
I don't even understand how those results would be possible unless the heatpipes were defective. I can only imagine that there is little to no fluid in them for them for the whole cooler to be performing so terribly. That'd be a huge QC oversight.

There's just no way that this is a normal result for that product. Must be defective.
 
Looking over the list, I'm impressed with the CM Hyper 212 evo for its price. I think I'm using the slightly older CM 212 plus, instead of the evo, and though my overclock is mild to moderate (4.2ghz on 3770k), it has never let me down. I'll have to look through the list again and see if anything beats it for the same price when it's time for a new build.
 
There's just no way that this is a normal result for that product. Must be defective.

I wouldn't be so sure. Look at the Dark Rock 4, it's much larger but has slightly slower fans and it too failed.

It's just the way CPU's are tested on the site. You won't see many people running their Ryzen CPU's at 1.475v, but I believe it's done to really stress the coolers and provide data for people who really want to push their CPU's to the absolute limit, even if that limit may slightly reduce the lifespan of their CPU.

On a different note, I've tried the 140mm version of these arctic fans as case fans and like the review, they report well over the RPM of the speed they're actually going. Arctic seems to have buggered up something along the line with the pwm reporting. In terms of moving air, they're fairly quiet but in my PC's case (Phantek Ethoo Evolv ATX) they vibrated and that resonated throughout the case so much that I had to get rid of them.
 
I sorry but that can't be correct.

There's just no way that this is a normal result for that product. Must be defective.
Well, I guess that negates all our very careful testing and retesting of the product and lengths we go to have a very consistent environment to test in so we get apples to apples results. :confused: All that said, show me some other reviews and please point out their testing parameters. I think some fanboys are getting their feelings hurt.
 
Here is our testing procedure and method laid out if you are not aware of it or have skipped over it in every cooler review we publish.

CPU
Our AMD Ryzen 1700 CPU will be running at 3.9GHz across all cores and will be being stressed by Prime95 using Small FFTs in both our CPU only test, as well as our combined testing. Our CPU Package power measures ~170 watts under full load.



GPU
In combined testing our GTX 670 will be running Furmark in addition to Prime95. This puts an additional 250 watts of heat into the system that the cooler will have to contend with.



Case
Testing is being performed in our Obsidian Series 750D Airflow Edition Full Tower ATX Case. Chassis fans include two Corsair 120mm fans pulling intake duty in the front, and a Corsair 140mm fan exhausting in the rear.

Fans are set to a locked 60% speed during all testing, as we found that is the best balance between performance and noise.


Thermal Paste

Thermal compound being used is Promilatech PK-3 Nano Aluminum. This is a very viscous compound rated at 11.2 W/m and requires no burn in or setup time.



Temperatures
Ambient temperature will be kept at as consistent temperature as possible for the duration of the tests. Temperatures are being measured in 4 places during both tests using our Sperry DT-506 Quad Input Thermometer, case intake, cooler/radiator intake, cooler/radiator exhaust, and case exhaust.



Idle
Average idle temperatures will be recorded after a thirty minute period of system inactivity.



Load
Load temperatures are measured every 5 minutes from all four points with the Sperry DT-506 thermometer, as well as the CPU Temperature (Tdie) and Package Wattage as reported in HwiNFO 64. Both the CPU only as well as the combined tests are 1 hour long, at which point the average temperatures will be used as our data point.



Sound
Sound levels will be measured with a BAFX Products BAFX3370 Digital Sound Level Meter from a distance of four feet away from the side of the case. With everything turned off and the room completely "silent," the meter registered a sound level of 39dB(A).
 
As an Amazon Associate, HardForum may earn from qualifying purchases.
kyle is not about feelings, all the testing could be all the consistent and careful you want, first those are not your "careful testing" I know you are trying to defend the reviewer because of course it's your personal, the testing procedure I think it have a nice approach as it should be, however still something must be flawed with the system a bad system airflow as the difference in GPU + CPU temps point to (just my opinion), as there's no way in the hell a cooler master Hyper 212 EVO will perform at Noctua NH-U12 and better than Reeven Justice and or Thermalright TRUE 140 Direct, when in fact reeven justice tend to perform at or better than NH-D14 levels... it was pointed several times by other users in other reviews that is not normal specially for people like me who have used, who build machines and who test machines for weeks,

what's worse by your own words:



suddenly you are surprised that the Thermalright TRUE 140 Direct passed richard's test when they perform 10C worse than the NH-U12?

Be quiet Dark rock 4 it's another case of WTF, 87C in CPU only test, but then it fail CPU + GPU combined? we are speaking of a variation of 13C by just adding a GTX 670? which is not a 250W GPU as pointed in the review, that's a 170W GPU which also underclock running furmark also the model is not described in the setup (it does matter).. so a lot of things point to a crap Case airflow, not enough cool intake air and/or not enough Exhaust (just a single 140mm exhaust fan at 60%?.. it have to be a joke and an insult to everyone who care for a proper case airflow, also what are the fans used?) we know YOU want to test things in a "realworld" way, but there are many things wrong with that system and it's an issue with the reviewer not with you, you are loading the system to unrealistic levels at unrealistic voltage levels to put unrealistic heat load inside a case that just can't handle that heat recirculating inside and then happily say the cooler fails?.. sorry but that's not real world and that calls for every review with that system and procedure simply *Wrong*.

you wanted to talk about feedback, well my first feedback start with a terribly prepared case airflow.

Roger that. You think we do it wrong. I think we are doing it exactly correct as reviewers should. The system is set up to show weaknesses in products, and I think it does that very well. Funny that we have about 20 cooler reviews on this particular system, and this is the first time that anyone has made such a fuss about it being flawed. And to that point, we have not had a single manufacture suggest that either....you know, the guys products being reviewed? What our test point out is that once you heat-soak a case, you had best have a heavy duty HSF on it....or go the smart route and use an AIO that exhausts the case. It really is that simple. And our testing bores that out. Sure, we could spend $100 on extra fans to show off a $30 cooler in a good light, but that is not really what it is all about is it? Better yet, test it in a 70F room and then all of a sudden that $30 cooler looks REALLY GOOD.

1536567289781kdyc78n_3_2.png

Now the other side of the argument, and this is the way most cooler tests are done, is to do these on an open bench. If you think that is more realistic than putting the cooler in the case, and them comparing them, well more power to you.

Quite frankly, you come across as eating up too much marketing that is served up to you IMO. There are times you get what you pay for, no matter the brand name.

As for your statements about the cards and wattage levels, those are all predicated on actual wattage loads we see when testing. I know you read a spec sheet somewhere, so you think you are right, but we wrote our test review based on facts that we gathered, not what you read on a spec sheet. And the that GTX 670 does NOT undervolt/clock on Furmark, that is one of the reasons it is being used in that setup. I fleshed out and tested that entire system here in my office before shipping it to Rich. I am very familiar with it. The fact is you do not know a lot of what you are talking about here.

But at the end of the day, this Arctic Freezer 33 eSports ONE is one of the worst air coolers we have tested. Your money is better spent elsewhere if you are overclocking IMO. By all means, put it in middle of the road stock systems. It is not an "enthusiast" product however.
 
we have to agree to disagree then. if what in reallity are weak points you consider are stronger points on your test rig who I am to say everything it's wrong?.. good luck then.

PD: btw I never say coolers should be tested in open bench, I was one of the first who supported the idea of this kind of test procedure, however i do not agree in test coolers inside a unrealistic heat soffocated case, that mean that while you show a cooler in a goodway everyone else will have totally opposed experience, it's your site that will remain as untrustworthy as no result here is a sign of real world, I can understand why you want to do this way however is incorrect, no cooler tested in that environment perform that way in the real life, where people care of case airflow to care their hundreds dollar machines.. I would never recommend a Hyper 212 over a Thermalright TRUE 140 just because according your test it perform better when in real life the latter crush the first one, and the same with other coolers, you say go smart and go with 100$+ AIO, I say go smarter and save money by going with a reeven justice and spending extra money in case fans to have an overall nicely cooled machine, Hey and this is from a guy using AIO since the first corsair H50, but again, more power to you if you believe anything of what you say is correct in real life with respect of cooling performance.
Roger that.
 
however i do not agree in test coolers inside a unrealistic heat soffocated case.

We test in a Corsair 750D airflow edition, in stock configuration. Much like someone would buy when building a PC. We chose this case for 2 reasons, 1 it's easy to work in, and 2 is because it is literally (or was) one of the most popular cases on the market. Meaning, it's relatable to what people are actually building a system in.

Check out ATX cases. You will see most all have a similar configuration as to what we are using, two 120mm front intakes and a single 140mm exhaust.

While I will completely agree that our CPU heat load is higher than most (the rig was made to push cooler to their limit, and... this is an overclocking website), this rig was designed specifically to replicate what most people are using. An ATX case with 2 intakes and 1 exhaust, a 250ish watt video card with an aftermarket cooler (chosen specifically, didn't want a blower card). To say our test rig is "unrealistic" is impressively far fetched.
 
While I will completely agree that our CPU heat load is higher than most (the rig was made to push cooler to their limit, and... this is an overclocking website).

This is why I am not alarmed that while this cooler performs stellar for me, it failed the [H] test. There's a reason I was bent on getting a 65TDP part or at the MAXIMUM a 90-100TDP part. I like my quiet, and I'll give up 20 fps for some quiet anytime. Brings me back to when Kyle wrote an article (which he also had a youtube vid for, NEED more [H] on youtube, blue hair or not Kyle) where he was trying to get some Intel CPU to hit 5Ghz. He considered the project a fail because he didn't reach his goal, but his work was a big win for me. I was still on the fence as for Intel or AMD for a new system, and his results of nearly 20c drop in temps by deliding were of great interest to me and I was thankful he did all that work with those expensive parts for me. Now I had the knowledge that with some work I could have a pretty spicy fast Intel system with some decent silence if I chose the effort to delid. 20c lower = some serious RPM drops for the fans.

Another example would be the system that I was replacing. I had been given one of those small form factor Shuttle machines which had an LGA1366 CPU in it. Let me tell you that CPU did NOT belong in that environment. I would easily hit 90-100c and the throttling fun began. I even cut a 120mm blow hole on top and installed a low profile 120mm fan and tested it pushing in and exhausting out to try to help the heat (blowing in was better, usually is). The facts remained that for me to have that system not so hot without sounding like a jet engine, I had to compromise. Basically because some idiot at shuttle put an Intel Core i7 920 rated for 130TDP it was just a heat box. Remember not my build I got this box free!

So I ran that thing with 2 of the 4 cores disabled, sometimes even only on a single core when the weather was very hot. Also disabling HT did wonders for heat (SMT I love it so much more than HT, thanks AMD!!!). The cooling solution and chassis for that hardware simply was inadequate and I think that is the case here. I love my Arctic-Freezer-33-eSports-Edition-Long-Named cooler, it's awesome for me.

TLDR
YES! - Arctic Freezer 33 for stock Ryzen 5 2nd Gen 2600
NO! - Overlocked Ryzen 7 1700 @ 1.475v

Makes sense to me. I imagine if HardOCP replicated my system and tested it again it would get a seal of approval for use in my setup.
 
Wow who would have thunk a $30 heatsink review would ruffle so many feathers.

Kyle brings up an interesting point in his rebuttle... a lot of the sites that review heatsinks use an open board on a test bench, not in a case. Not real world scenarios. Also I think that by reading a site named HardOCP, one would expect things to get tortured and abused. That's why I'm here, but I guess not everyone else.

Keep up the torture, Kyle. Personally, I want to know how things perform in a WORST case scenario, not the best scenario.
 
We test in a Corsair 750D airflow edition, in stock configuration. Much like someone would buy when building a PC. We chose this case for 2 reasons, 1 it's easy to work in, and 2 is because it is literally (or was) one of the most popular cases on the market. Meaning, it's relatable to what people are actually building a system in.

Check out ATX cases. You will see most all have a similar configuration as to what we are using, two 120mm front intakes and a single 140mm exhaust.

While I will completely agree that our CPU heat load is higher than most (the rig was made to push cooler to their limit, and... this is an overclocking website), this rig was designed specifically to replicate what most people are using. An ATX case with 2 intakes and 1 exhaust, a 250ish watt video card with an aftermarket cooler (chosen specifically, didn't want a blower card). To say our test rig is "unrealistic" is impressively far fetched.
Here's the only problem I see with this argument.

The site is for enthusiasts. The CPU isn't at stock settings, the motherboard isn't at stock settings, the video card isn't at stock settings... all awesome btw.

But the case is left stock.:confused: I'm kind of confused why you choose a setup that "most people are using" while simultaneously doing an overclock that most people aren't using. I get the idea you want worst possible scenario, but I'd like to see you do this with at least 1 more intake fan (bottom mount) and 1 more exhaust (top mount) or perhaps swap in a couple of 140mm for the 120's. After all, some of those AIO units you're testing against are dual fan units that might be bolted to the top of the case, the least you could do is put in another exhaust fan to make the situation a bit more fair so the naysayers have less to pick apart even when a cooler still fails spectacularly. :D

Now do I think it would have saved this cooler? Fuck no. This thing is a marketing disaster and an over-spec'd piece of shit. It likely would have failed this scenario while in an open-air test bench. It doesn't even fully cover the AM4 heatspreader while claiming to be AM4 compatible. It's got 4, 6mm heatpipes when most decent coolers I've seen out there with similar heatpipe arrangements are barely rated for 95w or 125w TDP CPU's. It might be fine for some guy doing a cheap Ryzen 5 or i5 budget setup and getting a decent OC, but that's it. This is [H], go big or go home, and I'm glad you guys called them out on their marketing bullshit. It's what you do best.

I'm sure you guys will disagree with me. That's cool. Your methodology does have merit even though I disagree slightly with it. Just something to consider for future is all.:cool:
 
Last edited:
TLDR
YES! - Arctic Freezer 33 for stock Ryzen 5 2nd Gen 2600
NO! - Overlocked Ryzen 7 1700 @ 1.475v

Makes sense to me. I imagine if HardOCP replicated my system and tested it again it would get a seal of approval for use in my setup.
100% agreed. We abuse these to show you a realistic worst case scenario.
 
But the case is left stock.:confused: I'm kind of confused why you choose a setup that "most people are using" while simultaneously doing an overclock that most people aren't using.
And that is a very valid point, but where do you go from there. You put 10 fans in? Then folks will argue about what fans, what configurations, what speeds. It is a deep rabbit hole....trust me. Been there, done that.
 
Back
Top