Discussion in 'Reviews' started by crazyeyesreaper, Jul 9, 2013.
To read this review go to: http://www.techpowerup.com/reviews/NZXT/Respire_T40/
IMO, if the package came with extra fanclips to attach an additional fan - I would have included it in the tests, It would have given a broader picture of what this cooler is capable of.
The fan it comes with is a DF1202512SELN, Cooler master use the same fan for some of their cases. A similar Yate Loon could be used to test temps and those fans are real dirt cheap.
It might not be so great with one fan, but with two it might really be a performer. Im just sayin
Interesting to note, the H100 was pretty low down the list for idle temp, but near the top for load.
Not a bad cooler i suppose, the base looks a bit cheapo though.
Here it's about €10 more than the Hyper 212 Evo, so totally not worth it imo.
I would like to see how this cooler does with an addition of a second fan.
I dont test second fans unless they are provided because the fan i select would be different from the fans other users may use and as such the conclusion and award is based on its performance out of box and what it offers out of box the option to add second fan is a feature, if another is provided I test that feature. This is because the fan I choose would influence the Noise level, performance, price and more all of which I take into account. Most manufacturers when offering the option for more fans usually have a fan in mind
2x Akasa Venom/Apache -- go go go
I don't think a second fan would contribute very much to the cooler's performance. It's heat capacity is too low. Perhaps if one was to slap two Delta 38mm/200CFM screamers... Otherwise, it just needs more mass.
The mass isnt bad actually but it all comes down to price to the performance it offers
If the heatpipes were soldered to the fins that would be a big plus along with a better fan aka more focused flow etc the fan they included most air goes out the sides of the heatsink instead of across the entire depth of the aluminum fins both of these hurt performance. However both = increase cost. The fact is when its on sale like it was when I wrote the review for $24.99 its a solid option over a stock cooler especially with Ivy Bridge / Haswell systems not everyone overclocks but at the same time not everyone wants to listen to the loud obnoxious stock coolers either thats where the NZXT Respire T40 is rather useful its a huge improvement over stock heatsinks while being easy on the wallet if purchased at the right time. Altho the $39.99 price tag is a bit high if price at 29.99 its a far more attractive option.
Reminds me of my gelid tranquillo for some reason. Probably just the white fan.
Maybe you could do it with some widely available fan like the corsair SP120?
If my memory is right most basic coolers use 6mm heatpipes like 212evo .
I wonder if by using the larger 8mm is the cause of lacking mid load temps .
It would be interesting if coolers mixed size , like 2 6mm and 28mm or with more pipes .
My guess is the 6mm are faster to respond but lack the capacity to handle E CPU's
Your articles were always a little bit strange to me. Why lack of second fan you name as disadvantage? Sometimes it is even advantage. For example if we will talk about Noctua U12S and U14S were is also noted that it lacks second fan.
I will give you an example. I have also tested U12S and U14S. If you want I can post exacts results. Push/ pull gives 1-2 deg. lower temp.
The question is: Do you really want to pay extra 20 eur for such increase of performance?
If we will talk about expensive cooler (except such companies as Noctua which are making expensive fans) I agree they should have extra fan. For example when you will test Kraken X40 you should note that it lacks second fan. But one fan for 40 USD cooler is not disadvantage. It is just normal
The second thing testing methodology. Why you are using Prime 95 for intensive load tests? It is good only for typical load tests. I recommend to use you Linx.
I hope it will help to increase quality of your reviews
1: Prime 95 is usually recommended by the manufacturer
2: CPU designs are changing and as such testing methods will change as well this is true in part due to AVX instruction set and the rise of heavily stressing a CPUs FPUs for max heat aka LinX / Prime95 both do this and in reality this means neither of those tests are truly indicative of actual performance. Prime95 in no way shape or form is even remotely close to a typical load.
3: Many people are still stuck in the past in terms of testing PC hardware, in an ever changing market things must change to keep up the old mentality eventually grows stagnant and becomes an issue.
I list it as a con because for the most part the cost to manufacturer these fans is tiny, evidence of this is a Yateloon which can be had for $4 at various etailers thats after manufacturing cost of said fan, shipping packaging worker wages etc. It is far cheaper for the cooler manufacturers so in essence on some coolers the cost to produce could increase a products cost a few $1 while improving performance. If adding a second fan from the get go can improve performance without drastically increasing price on the consumers side while also remaining quiet can be done why not do it? It makes the product look more attractive and improves its price to performance. This is why its a con, you raised Noctua as an example, if that 1-2C can take it from being in the middle of the pack to challenging the likes of a $100 cooler like the PH-TC14PE yet being half the size does that not make for a more attractive cooler? It is smaller, quieter and cools as well as the much larger product and if the fan was included it would cost far less. Granted all this depends on the CPU used and many other variables that people like to overlook.
Regardless I take notice of your recommendation for LinX but I will not be using it, LinX uses a higher voltage or vrm load I cant remember of the top of my head than any other application in use today, what this means is that LinX is rather pointless in regards to my testing. Of course Prime 95 is the same with AVX stressing the FPU primarily. Both if improperly run and setup can eventually cause damage to a chip if stressed hard enough. Since I am testing coolers the chance of hitting thermal throttle point is a real possibility and hitting that point hard enough and for long enough will kill a chip.
That said, thank you for the recommendation all input is welcome but at this stage I wont be switching to LinX. However things will be changing in the near future in which ways I wont say just yet .
Prime95 is a distributed computing app, and gives a workload that is similar to what those guys who do FAH and BOINC will see, running 24/7.
LinX is a test that was built to stress systems to the max, which nearly NO user will EVER do. This is further compounded by Haswell chips that use AVX2, and get higher voltage than they would with "typical" workloads. Very few things actually use AVX2 right now.
I don't understand why cooler reviews are done on a 3960/LGA 2011 to be frank. These are not mainstream processors plus the fact that Ivy and Haswell do not solder and use Tim makes the results difficult for mainstream users.
Because it was what hardware I had to give him at the time? why beat the dead horse that was beat around months ago when he started doing reviews? Why repeat the same comments you made back then?
W1zzard and I have outfitted Crazy with Haswell hardware, he's currently working on getting reviews done with that.
Hardware Crazy uses is MY choice, I sent it all to him, minus the Haswell CPU. That's MY responsibility, not his, get off his case, asshat. You asked these questions before, so now I take this as pure trolling, which is against forum rules, let me remind you.
As a direct answer, SKT 2011 was used to push coolers to the limit. I even sent him a 3960X chip that has excessive power consumption, simply to push the coolers hard.
I don't believe I made the comment previously and what gives you the crown to call someone an asshat? My resonse was just a why and I see you answered it..... TY
You were a reviewer. You should understand the reason for high-heatload testing of a cooler, no?
Nevermind, you obviously did not.
And lighten up dude, just a word.
Dave, I don't. Both Haswell and Ivy run hotter and would love to see he results with mainstream as I mentioned. Sorry your taking it the wrong way. I do builds and have used NZXT product occassionally so mainstream is what I look at. My 3930K does not run close to the temps as a 3770K or 4770K
No, they don't. Maybe in your experience they did, but overall, SKT2011 has higher TDP and higher power consumption under OC. His chip pulls 150W @ stock. In dealing with coolers, overall heatload is what matters, not really temps reported. Like all CPUs are not equal in OC, they are not equal in power consumped or temps. Many factors play a role.
And again, it's about the hardware I had on-hand. I have limited resources, just like anyone else. When I was able to update his rig, I did. Those parts he's using, minus the CPU, came out of my pocket. TPU paid for the CPU.
Crazy uses 3960X, with more cache, comparing the chips is not prudent. Your chip runs cooler than his..it should..it uses far less cache. Mine is similar to his, although not quite as high.
Really? my 3930K does not run close to temps on a 3770K or 4770K. It is like air conditioning. remember, not all 3770K's run overclocked at 1.25 volts. Some need 1.3-1.35 volts for 4.5GHz and you know those temps will be much hotter than a 3930K or 3960K. Besides the point is moot cause xtreme chips are not mainstream. Just sayin.. Heck, I could run my 3930 at 1.45 volts and be cooler than a 3770K at 1.35 volts
This is why I called you an asshat. You know all chips are not equal, and comparing yours to his is not as black and white as you'd like to present. I could have sent crazy AMD chips, and then people would have complained then too.
Well honestly the way i see it does not really matter what platform you use if all the testing is done on that same platform,we are comparing cooler vs cooler here after all,no reason the deltas between them would be different on haswell/ivy.
Well responses can be interpreted differently.
Separate names with a comma.