Scythe FUMA 2 CPU Air Cooler Review: Thermals, Noise, & Pressure Benchmarks
video description
Date: 2021-01-09
Related videos
Comments and reviews: 10
Hampus
In the testing I've seen around the web and my experience after buying it is that it is inefficient in the last 300rpm and only improve temps by 1 degree or 2 while becoming much louder.
Running it at stock pwm curves (around 50-70% pwm) at full load it is barely audible over idleing chassi fans.
Air coolers like this really have to be run free at more efficient fan speeds otherwise they perform piss poor.
A great example is the u12a in SweClockers.com testing where there is only 3 degree difference between full fan speed and using low noise adapter (somewhere around 7v-9v if I remember right), But a 10db increase. A h115i pro rgb will perform identically to u12a in quiet mode as u12a with low noise adapter but both running at full tilt they will have the same noise with the h115i being a lot cooler.
Air coolers often don't scale in performance over 70% pwm but the size of water coolers make them scale well up to full speed.
I have said it before. This noise normalized testing is something that is not representable for how the coolers will actually perform for users.
The fuma have looked like it is a 2 degrees off D15 and u12a but at the same time holding the same noise and would perform about equal as a h115i I noise and cooling in situations users actually use them in.
reply
In the testing I've seen around the web and my experience after buying it is that it is inefficient in the last 300rpm and only improve temps by 1 degree or 2 while becoming much louder.
Running it at stock pwm curves (around 50-70% pwm) at full load it is barely audible over idleing chassi fans.
Air coolers like this really have to be run free at more efficient fan speeds otherwise they perform piss poor.
A great example is the u12a in SweClockers.com testing where there is only 3 degree difference between full fan speed and using low noise adapter (somewhere around 7v-9v if I remember right), But a 10db increase. A h115i pro rgb will perform identically to u12a in quiet mode as u12a with low noise adapter but both running at full tilt they will have the same noise with the h115i being a lot cooler.
Air coolers often don't scale in performance over 70% pwm but the size of water coolers make them scale well up to full speed.
I have said it before. This noise normalized testing is something that is not representable for how the coolers will actually perform for users.
The fuma have looked like it is a 2 degrees off D15 and u12a but at the same time holding the same noise and would perform about equal as a h115i I noise and cooling in situations users actually use them in.
reply
Oren
other than the really massive ones (the grand kama was bizarre), i've always enjoyed installing Scythe products for my customers. Nice high quality workmanship, always quiet, standard mounts so you can easily replace fans - not that i've seen any fail - they are great quality. Unlike other brands they usually employ designs requiring lower pressure (the fins are not tightly packed) so they can run the fans slow. In fact, the only problems i've had with them is some old motherboards will trigger the low-fan-speed warning in the Katana-3 default pwm curve (it can go as low as 450 RPM when the CPU is idle). As long as you don't overclock too much, they are perfect for the price.
reply
other than the really massive ones (the grand kama was bizarre), i've always enjoyed installing Scythe products for my customers. Nice high quality workmanship, always quiet, standard mounts so you can easily replace fans - not that i've seen any fail - they are great quality. Unlike other brands they usually employ designs requiring lower pressure (the fins are not tightly packed) so they can run the fans slow. In fact, the only problems i've had with them is some old motherboards will trigger the low-fan-speed warning in the Katana-3 default pwm curve (it can go as low as 450 RPM when the CPU is idle). As long as you don't overclock too much, they are perfect for the price.
reply
Khunvyel
Is thermal paste included here? I suppose not. And if, then what quality?
Aside from that, this sounds like it is currently the best quality and compatibility cooler for anyone who just keeps stuff at stock and doesn't overclock manually and wanting rather low noise levels with a very decent price. Noctua has some work cut out for them it seems. Good to know, I've held off trying them until your review came up. One of the next builds scheduled for me that falls in that category will likely see the Fuma 2, but that solely relies on the presence and quality of the paste.
reply
Is thermal paste included here? I suppose not. And if, then what quality?
Aside from that, this sounds like it is currently the best quality and compatibility cooler for anyone who just keeps stuff at stock and doesn't overclock manually and wanting rather low noise levels with a very decent price. Noctua has some work cut out for them it seems. Good to know, I've held off trying them until your review came up. One of the next builds scheduled for me that falls in that category will likely see the Fuma 2, but that solely relies on the presence and quality of the paste.
reply
Innocent_Bystander
I'm actually kind of disappointed that you caved into these asinine tactics. Just for the spam that you know came from them, my only response would be a heavy dose of my driving finger.
Even if it happened to be top of the charts, they need to learn how to conduct business in a respectful way. The comment spam is unprovable. Correlation does not equal causation, and it'll be difficult if not impossible to prove that those comments came from them. Without any other information, I wouldn't hold that against them, but they can kick rocks over the rest.
reply
I'm actually kind of disappointed that you caved into these asinine tactics. Just for the spam that you know came from them, my only response would be a heavy dose of my driving finger.
Even if it happened to be top of the charts, they need to learn how to conduct business in a respectful way. The comment spam is unprovable. Correlation does not equal causation, and it'll be difficult if not impossible to prove that those comments came from them. Without any other information, I wouldn't hold that against them, but they can kick rocks over the rest.
reply
Dionamuh
To be clear: are the actual temperatures shown (delta T over ambient) +21 degrees? So 50 degrees is actually 71 degrees?
I've been using the Fuma (1) for many years. First on a Xeon, now on a Ryzen 9 3900X (after ordering the AM4 mounting hardware from Scythe for a couple of dollars). Very quiet, performs adequately. The Ryzen 9 3900X has a tendency to have very short high temperature spikes because of it's chiplet design and high density, which shows in monitoring software, but temperatures overall are I'd say pretty good.
reply
To be clear: are the actual temperatures shown (delta T over ambient) +21 degrees? So 50 degrees is actually 71 degrees?
I've been using the Fuma (1) for many years. First on a Xeon, now on a Ryzen 9 3900X (after ordering the AM4 mounting hardware from Scythe for a couple of dollars). Very quiet, performs adequately. The Ryzen 9 3900X has a tendency to have very short high temperature spikes because of it's chiplet design and high density, which shows in monitoring software, but temperatures overall are I'd say pretty good.
reply
SmashStomp
Great review, only thing I would have liked to have seen was the performance numbers with aftermarket higher rpm fans.. To see how much of a impact it would have thermals.. This would make a great argument for future upgradability with the cpu cooler in increasing performance if you need it down the road.
On whether it would be better to invest in new fans or get a new cooler all together with a new gen of cpus that has higher tdp.
reply
Great review, only thing I would have liked to have seen was the performance numbers with aftermarket higher rpm fans.. To see how much of a impact it would have thermals.. This would make a great argument for future upgradability with the cpu cooler in increasing performance if you need it down the road.
On whether it would be better to invest in new fans or get a new cooler all together with a new gen of cpus that has higher tdp.
reply
Reinier
Actually, I've been asking for Scythe cooler reviews because they look to be good performers for the price here in Europe. Not sure how their prices look in the US though.
But you're too late! ;-) I bought a Deepcool Gammax GTE V2 Black for 25 a few weeks ago. Looks really nice with a Corsair LL 120 (which was a pain to install as it's wider than the stock fan, so the metal clip thingies could barely reach their attachment point).
reply
Actually, I've been asking for Scythe cooler reviews because they look to be good performers for the price here in Europe. Not sure how their prices look in the US though.
But you're too late! ;-) I bought a Deepcool Gammax GTE V2 Black for 25 a few weeks ago. Looks really nice with a Corsair LL 120 (which was a pain to install as it's wider than the stock fan, so the metal clip thingies could barely reach their attachment point).
reply
James
I have a scythe ninja 5 and it's badass. Although terrible in my last case (fractal node 304) because of size. Currently running it open air while i finish manufacturing my custom case.
Running 30-55 C under mild to moderate load on a ryzen 3900X with conductonaut. I don't tend to be able to sustain load long enough for the monster heatsink to actually really warm up, even compiling code!
reply
I have a scythe ninja 5 and it's badass. Although terrible in my last case (fractal node 304) because of size. Currently running it open air while i finish manufacturing my custom case.
Running 30-55 C under mild to moderate load on a ryzen 3900X with conductonaut. I don't tend to be able to sustain load long enough for the monster heatsink to actually really warm up, even compiling code!
reply
undefined
On the other hand you can still get the same performance or better in same special cases, with an old 8700k also 6cores/12 threads oc-ble to 5Ghz, no need to rush to upgrade to either ryzen latest and greatest or intel 10 series, prices are anyway crazy now, so till they really settle down in a year or 2, :-), one can still manage with an 8700k. Just repeat this mantra Must not buy! :-)
reply
On the other hand you can still get the same performance or better in same special cases, with an old 8700k also 6cores/12 threads oc-ble to 5Ghz, no need to rush to upgrade to either ryzen latest and greatest or intel 10 series, prices are anyway crazy now, so till they really settle down in a year or 2, :-), one can still manage with an 8700k. Just repeat this mantra Must not buy! :-)
reply
Helios1234p
Have you tried testing the coolers with a standardized set of fans to see how effective the heatsinks themselves are? The fans are clearly the limiting factor here. For someone on a budget, they might end up getting the Scythe but knowing that the heatsink itself is efficient, it means they wouldn't have to upgrade the whole cooler in the future but just the fans.
reply
Have you tried testing the coolers with a standardized set of fans to see how effective the heatsinks themselves are? The fans are clearly the limiting factor here. For someone on a budget, they might end up getting the Scythe but knowing that the heatsink itself is efficient, it means they wouldn't have to upgrade the whole cooler in the future but just the fans.
reply
Add a review, comment
Other channel videos
