WEBVTT

00:00:00.240 --> 00:00:04.480
okay okay i know what you're thinking

00:00:03.360 --> 00:00:12.719
water plus a high amperage

00:00:08.480 --> 00:00:14.400
transformer that can literally blow up

00:00:12.719 --> 00:00:18.320
and take the rest of your computer hardware with it

00:00:16.480 --> 00:00:24.480
what could go wrong good idea or bad idea

00:00:21.119 --> 00:00:26.960
it's here and our goal today is to find

00:00:24.480 --> 00:00:29.119
out if the usual benefits of liquid cooling

00:00:27.920 --> 00:00:35.760
quieter cooler operation can also apply

00:00:32.399 --> 00:00:37.760
to a power supply

00:00:35.760 --> 00:00:42.079
and we wouldn't be able to do it without tunnelbear tunnelbear makes it easy to

00:00:40.000 --> 00:00:46.399
privately and securely browse a more open internet try tunnelbear for free at

00:00:44.719 --> 00:00:56.239
tunnelbear.com LTT

00:00:56.239 --> 00:01:01.600
so the first thing that caught our attention about this product

00:00:59.920 --> 00:01:08.479
aside from it being a water cooled power supply was the standard g1 quarter openings

00:01:05.439 --> 00:01:10.640
complete with bit power logo under the

00:01:08.479 --> 00:01:15.680
modular power connectors if they were trying to do a quality job

00:01:12.880 --> 00:01:21.280
they definitely found a solid partner so let's keep going then

00:01:17.439 --> 00:01:24.240
the internals do still have a fan making

00:01:21.280 --> 00:01:27.680
this a hybrid design similar to ASUS poseidon series of graphics cards

00:01:26.799 --> 00:01:34.960
but at the bottom of the main pcb there is

00:01:31.119 --> 00:01:36.479
also what appears to be a full length

00:01:34.960 --> 00:01:42.159
water block it doesn't seem to have any fins but it

00:01:39.600 --> 00:01:44.640
does have a flow meter interesting

00:01:43.680 --> 00:01:49.840
now the nda that we signed prevents us from

00:01:47.600 --> 00:01:53.360
tearing down our unit to show you exactly what that plate makes contact

00:01:52.479 --> 00:01:58.640
with but everything we can see from peering

00:01:55.920 --> 00:02:02.880
inside suggests that the internal heat sink array is pretty typical of a normal

00:02:00.960 --> 00:02:08.080
air cooled power supply and nothing on the fan side suggested

00:02:05.600 --> 00:02:13.840
that special design considerations had been made to accommodate transferring

00:02:10.879 --> 00:02:18.160
the heat to the water block so

00:02:14.720 --> 00:02:20.080
maybe the top mounted heat sinks run all

00:02:18.160 --> 00:02:24.400
the way through the pcb holes to the base of the block

00:02:21.680 --> 00:02:28.640
or maybe there's just a big thick thermal pad on the bottom of a

00:02:26.400 --> 00:02:32.959
completely normal power supply we actually don't know

00:02:30.879 --> 00:02:36.959
but none of this prevents us from answering our original question anyway

00:02:35.120 --> 00:02:43.280
so let's move on to find out if cooler and quieter can be

00:02:40.560 --> 00:02:47.120
achieved through fsp's method whatever it is

00:02:44.239 --> 00:02:51.680
we tested our unit in both air and liquid cooled mode and also compared it

00:02:49.920 --> 00:02:58.319
to one of the best air-cooled power supplies on the market the Corsair

00:02:54.040 --> 00:03:01.760
ax-1200i we ran each of them at idle

00:02:58.319 --> 00:03:04.640
two-thirds load and max capacity taking

00:03:01.760 --> 00:03:07.440
both thermal and acoustic measurements now

00:03:05.440 --> 00:03:12.720
one does not simply make a system consume

00:03:09.480 --> 00:03:15.200
1200 watts of power without some

00:03:12.720 --> 00:03:22.800
seriously overkill hardware so to achieve that we rigged up an x 299

00:03:18.239 --> 00:03:26.000
platform with the 16 core core i9 7960x

00:03:22.800 --> 00:03:27.599
overclocked to 4.2 gigahertz then we

00:03:26.000 --> 00:03:35.519
threw four overclocked titan x pascals onto the rig

00:03:31.200 --> 00:03:36.799
for some gp GPU rendering and ah

00:03:35.519 --> 00:03:43.519
beautiful just over 1400 watts from the wall so with some

00:03:41.200 --> 00:03:49.120
CPU voltage adjustments we got our system dialed in to about 1200 watts

00:03:46.799 --> 00:03:54.000
we then used cablemod extenders to be able to have the power supply in a

00:03:51.360 --> 00:03:58.000
separate room in order to control sound leakage

00:03:55.120 --> 00:04:01.360
first up then our air-cooled testing at idle

00:03:58.879 --> 00:04:05.120
both units remained silent neither of their fans turned on

00:04:03.200 --> 00:04:11.680
nothing to see here at two-thirds load the fsp fan spun up

00:04:08.480 --> 00:04:14.319
to about 800 RPM and our thermal imagery

00:04:11.680 --> 00:04:20.000
showed hot spots inside the unit at over 77 degrees celsius our extech sound

00:04:17.440 --> 00:04:24.320
meter read 44 decibels and i'd say subjectively it was about as loud as a

00:04:21.840 --> 00:04:30.080
d5 pump at 12 volts by comparison our Corsair axi reached

00:04:27.040 --> 00:04:31.280
only 700 RPM and was ever so slightly

00:04:30.080 --> 00:04:37.120
quieter at full 1200 watt load the fan spun up

00:04:34.639 --> 00:04:43.040
all the way to about 1200 RPM reaching 50 decibels and it was definitely

00:04:39.600 --> 00:04:45.759
audible over our 12 volt d5 pump making

00:04:43.040 --> 00:04:50.800
it the worst offender in a hypothetical system being pushed to the limits

00:04:48.400 --> 00:04:55.120
we saw two significant hot spots with our thermal camera and we even

00:04:52.800 --> 00:05:00.320
experienced some strange clicking and crackling noises

00:04:56.880 --> 00:05:03.040
for comparison our Corsair hit only 950

00:05:00.320 --> 00:05:07.360
RPM at 46 decibels making it significantly less noticeable over a d5

00:05:06.240 --> 00:05:11.360
not to mention that there was no crackling but it was noticeable over a

00:05:10.160 --> 00:05:16.960
d5 so maybe there is some improvement to be

00:05:13.280 --> 00:05:19.759
had from water cooling and besides this

00:05:16.960 --> 00:05:24.560
the hydro ptm plus 1200 watt was never meant to be air-cooled so

00:05:21.919 --> 00:05:28.960
let's throw this fish back in its natural habitat

00:05:26.160 --> 00:05:34.400
with a 240 millimeter liquid cooler hooked up we started with this power

00:05:31.280 --> 00:05:36.000
supply at two-thirds load and oh well

00:05:34.400 --> 00:05:41.199
that's kind of weird the fan turned on

00:05:37.680 --> 00:05:43.759
it did spin at a slightly lower RPM than

00:05:41.199 --> 00:05:48.000
when it was air-cooled but while our noise measurements did go down we

00:05:46.320 --> 00:05:50.880
actually didn't find that it made a noticeable difference compared to all

00:05:49.919 --> 00:05:57.280
air that midpoint heat spot was gone though

00:05:54.240 --> 00:05:58.400
at full bore the fan spun up to 900 ish

00:05:57.280 --> 00:06:05.120
RPM quite a bit lower than when it was air-cooled

00:06:01.360 --> 00:06:07.680
but it was still audible over a d5 with

00:06:05.120 --> 00:06:12.160
pretty much the same heat spots on the fan side

00:06:08.960 --> 00:06:14.240
so what happened here well without being

00:06:12.160 --> 00:06:17.440
able to take the unit apart to confirm our

00:06:15.440 --> 00:06:22.800
the block is only cooling the back side of the pcb theory

00:06:19.520 --> 00:06:24.960
it remains just that a theory but one

00:06:22.800 --> 00:06:30.240
that is somewhat corroborated by these renders that we found on fsp's website

00:06:27.600 --> 00:06:34.080
to be clear we did observe an improvement

00:06:31.759 --> 00:06:39.120
and it is possible that a liquid-cooled power supply could achieve better

00:06:36.639 --> 00:06:44.720
cooling and noise characteristics than an equivalent air-cooled one

00:06:41.680 --> 00:06:47.360
but the implementation on this one did

00:06:44.720 --> 00:06:52.560
leave a bit to be desired we had really hoped that the fan wouldn't be needed at

00:06:50.000 --> 00:06:56.400
all as long as the water cooling didn't fail

00:06:54.000 --> 00:07:00.479
so i would like to see a second attempt at this concept but maybe one that is

00:06:59.199 --> 00:07:05.759
more tightly integrated with the water block

00:07:03.199 --> 00:07:10.720
actually directly making contact with the heat producing parts of the power

00:07:07.759 --> 00:07:14.240
supply maybe then fsp would be onto something

00:07:11.840 --> 00:07:16.720
as it stands at the time of filming we actually don't know the price of this

00:07:15.759 --> 00:07:21.680
unit but we don't think it really matters anyway

00:07:19.599 --> 00:07:26.560
however much it is we really don't see the point of paying

00:07:23.919 --> 00:07:31.520
extra for it over another high quality high wattage power supply unless you

00:07:29.039 --> 00:07:34.639
just must have everything in your system liquid cooled

00:07:33.520 --> 00:07:40.400
for like ultimo bragging rights

00:07:37.919 --> 00:07:45.120
speaking of ultimo bragging rights you can brag about your squarespace

00:07:42.479 --> 00:07:48.319
website when you create it in just a matter of like i mean anything from

00:07:47.039 --> 00:07:52.880
minutes to actually you can spend a long time on a squarespace site but the point is no

00:07:51.199 --> 00:07:57.199
matter how much time you spend on it it's gonna look great they've got 24 7

00:07:55.520 --> 00:08:00.639
support via live chat and email if you're having trouble it starts at only

00:07:59.199 --> 00:08:06.400
12 bucks a month and you get a free domain for the year if you buy squarespace

00:08:03.360 --> 00:08:08.000
and this is cool cover pages on the

00:08:06.400 --> 00:08:13.120
subject of having something up in just minutes allows you to set up a beautiful

00:08:10.879 --> 00:08:17.039
one-page online presence in minutes and they've got tons of great features

00:08:14.879 --> 00:08:20.800
they've got responsive design so your website will look great on any device

00:08:18.960 --> 00:08:24.000
they've got their logo creator they allow you to publish an apple news

00:08:22.560 --> 00:08:28.479
format so millions of potential readers can see

00:08:25.840 --> 00:08:32.240
your squarespace blog so start a trial with no credit card required and start

00:08:30.240 --> 00:08:35.519
building your website today then when you sign up for good head over to

00:08:33.839 --> 00:08:40.240
squarespace.com LTT and use offer code LTT to get 10 off

00:08:38.800 --> 00:08:43.279
your first purchase we'll have that linked below

00:08:42.000 --> 00:08:48.720
so thanks for watching guys if this video sucked you know what to do but if it's awesome hit the like button get

00:08:46.640 --> 00:08:52.160
subscribed maybe consider checking out where to buy the stuff we featured at

00:08:50.480 --> 00:08:55.600
the link in the video description as well as our merch tour and our community

00:08:54.160 --> 00:08:59.200
forum we have many links in our description we have the best links