WEBVTT

00:00:01.880 --> 00:00:07.080
woo look at her go here in my hands is

00:00:05.000 --> 00:00:12.240
the power supply that's responsible for that fireworks display and guess what

00:00:09.360 --> 00:00:18.320
guys we didn't try to kill it and it isn't from some random AliExpress brand

00:00:15.480 --> 00:00:24.000
this unit has five stars on PC part picker and thousands of sales on Amazon

00:00:20.920 --> 00:00:26.279
and yet it failed spectacularly during

00:00:24.000 --> 00:00:30.519
routine testing for power supply circuit our Channel that's been testing Legions

00:00:28.279 --> 00:00:36.040
of power supplies and along the way discovering that many power supply

00:00:32.719 --> 00:00:38.239
manufacturers aren't so honest some are

00:00:36.040 --> 00:00:43.160
over representing their products and wait some are underselling their

00:00:40.239 --> 00:00:48.440
performance that's not what I expected well either way it's finally time to

00:00:45.680 --> 00:00:52.320
take a look at our power supply Gauntlet and when I say Gauntlet I mean it

00:00:50.199 --> 00:00:59.199
because our power supply testing process has resulted in a staggering 23% failure

00:00:56.320 --> 00:01:03.320
rate really separating the good from the uh

00:01:00.640 --> 00:01:09.200
well bright at least speaking of failure I just failed to do a segue to our

00:01:05.400 --> 00:01:12.439
sponsor H MSI their core liquid a15 aiio

00:01:09.200 --> 00:01:14.280
cooler is quiet stylish and best of all

00:01:12.439 --> 00:01:17.840
easy to install cool down your high performance build by picking one up

00:01:15.600 --> 00:01:22.880
today using our Link in the description 9 months ago now we launched our first

00:01:20.320 --> 00:01:28.400
Labs team Channel PSU circuit and since then we've tested 39 power supplies nine

00:01:25.960 --> 00:01:32.119
of which died at some point during our test Suite but we're getting ahead of

00:01:30.439 --> 00:01:38.159
ourselves a little bit first it's time to take a look at our fully armed and

00:01:34.759 --> 00:01:39.880
operational power supply station because

00:01:38.159 --> 00:01:43.600
the keite among you might have noticed that it looks a little different from

00:01:41.759 --> 00:01:48.119
how it looked 2 and 1/2 years ago when we first unboxed this bad boy starting

00:01:45.960 --> 00:01:54.560
at the bottom we've got our AC source which is able to deliver 3 kilovolt amps

00:01:50.880 --> 00:01:57.840
which is basically 3,000 Watts but not

00:01:54.560 --> 00:02:00.079
quite because reasons this is enough for

00:01:57.840 --> 00:02:05.759
us to test up to, 1600 W power power supplies because we need to be able to

00:02:02.520 --> 00:02:07.680
briefly deliver 200% of the rated power

00:02:05.759 --> 00:02:15.080
for the power supply for Excursion testing as for why it isn't exactly

00:02:10.720 --> 00:02:17.040
double the reasons anyway above that we

00:02:15.080 --> 00:02:22.239
have what chroma calls their onoff controller which is just a big relay

00:02:19.680 --> 00:02:25.920
which is just a fancy name for switch then above that we have our short

00:02:24.040 --> 00:02:29.519
circuit and over voltage protection tester moving up we have our DC power

00:02:28.120 --> 00:02:35.640
source which we don't really use at this time and then above that we have our timing and noise analyzer which is

00:02:33.040 --> 00:02:39.599
basically a headless oscilloscope with a host of relays that can be used to

00:02:37.640 --> 00:02:44.560
automate functions like turning the power supply on or off or changing the

00:02:42.280 --> 00:02:48.680
characteristics of our load say for example to add

00:02:45.920 --> 00:02:53.159
capacitance then there's a shelf which does shelf things above the shelf we've

00:02:51.040 --> 00:02:56.800
got the connection panel which allows easy connections so that poor Lucas

00:02:55.000 --> 00:03:01.319
doesn't have to go splunking in the back of a rack every time a wire needs to be

00:02:58.920 --> 00:03:08.280
plugged in of course the real meat and potatoes is up here our 13 loads an

00:03:05.799 --> 00:03:12.319
unlucky number for many a power supply basically these guys use Moss feds to

00:03:10.720 --> 00:03:18.120
dissipate the energy from our power supply as heat we've got 12 400 W loads

00:03:16.120 --> 00:03:24.879
that can be strung together to fully load any rail on any power supply up to

00:03:21.480 --> 00:03:27.519
1600 watts and then our 13th unit has

00:03:24.879 --> 00:03:32.840
two 100 WT loads which we typically use for 5vt standby and- 12 volt since we

00:03:30.720 --> 00:03:38.680
don't really need a big individual load for those it uh can get pretty warm

00:03:36.159 --> 00:03:43.319
Behind These Guys and beside that is a digital multimeter which we use for the

00:03:40.959 --> 00:03:47.319
well the seven thermocouples that go into our test chamber to measure ambient

00:03:45.200 --> 00:03:52.760
air intake and exhaust air for the power supply and measures the connectors to

00:03:49.640 --> 00:03:54.480
make sure that they don't you know melt

00:03:52.760 --> 00:03:59.760
we don't publish our temperature results yet but it is something that we're working toward finally there's the crown

00:03:57.840 --> 00:04:05.200
jewel of our tester our road and schwarts mxo 58 oscilloscope she's got

00:04:02.920 --> 00:04:09.519
eight analog channels 16 digital channels which is nice but the real

00:04:07.519 --> 00:04:14.239
benefit of this scope is the ease of programming and the nice waveforms you

00:04:11.760 --> 00:04:18.919
get to look at truthfully we know we're kind of using the Ferrari of

00:04:15.799 --> 00:04:20.519
oscilloscopes to go get groceries but

00:04:18.919 --> 00:04:25.960
that can be worth it when you acquire this equipment and sometimes it takes

00:04:23.000 --> 00:04:32.360
years to finally start using it which yeah yeah what's the deal with that okay

00:04:29.880 --> 00:04:36.840
I do know the deal with that the chroma is not exactly user friendly or

00:04:34.680 --> 00:04:41.360
documented making matters worse the programming is very similar to assembly

00:04:39.199 --> 00:04:46.960
code which turned the automation process for our testing from a two-month project

00:04:43.680 --> 00:04:49.120
to a 10mon project for poor Lucas

00:04:46.960 --> 00:04:54.479
thankfully I didn't have to do it but now that it's finally done we can figure

00:04:51.199 --> 00:04:56.280
out what power supplies are good and

00:04:54.479 --> 00:05:01.639
which ones are not so good which brings us to an important question what makes a

00:04:58.880 --> 00:05:07.680
power supply good well at its most basic level a good power supply must adhere to

00:05:04.240 --> 00:05:10.240
the ATX specification and deliver these

00:05:07.680 --> 00:05:14.680
voltages to your compy regardless of the kind of load that it's under or the

00:05:12.160 --> 00:05:20.199
input power conditions now we don't follow exactly the ATX prescribed test

00:05:17.440 --> 00:05:25.479
plan but we try wherever it's practical and our tests are heavily inspired by

00:05:22.960 --> 00:05:30.560
ATX the good news is pretty much everything we've tested even the ones

00:05:27.720 --> 00:05:36.319
that failed has pretty pry much operated within the ATX specification until it

00:05:33.520 --> 00:05:41.360
didn't but the bad news is we have seen some big differences in efficiency in

00:05:39.360 --> 00:05:44.960
simple terms the efficiency of a power supply is the percentage of the power

00:05:43.319 --> 00:05:49.520
that you draw from the wall that gets turned into useful energy that your

00:05:46.720 --> 00:05:53.400
computer can use now many manufacturers Market their power supplies with a

00:05:51.319 --> 00:05:58.160
little 80 plus certification that you'll find on the box but in our testing there

00:05:56.039 --> 00:06:05.919
were two power supplies whose measured efficiency didn't match their ratings

00:06:01.080 --> 00:06:09.479
The Corsair SF 850l and the fsp Vita

00:06:05.919 --> 00:06:12.440
850gm both of them actually exceeded

00:06:09.479 --> 00:06:19.319
their rated deficiency a result that is validated actually by clear results own

00:06:15.240 --> 00:06:22.160
80+ testing why would anybody Market

00:06:19.319 --> 00:06:26.880
their unit as 80 plus gold when they could have labeled it 80 plus Platinum

00:06:24.319 --> 00:06:31.520
instead when we talked to Corsair and fsp about this we got similar answers

00:06:29.160 --> 00:06:36.560
from both of them these power supplies it turns out were designed to be gold

00:06:34.199 --> 00:06:41.160
efficiency power supplies but then their Engineers went and did a bit too good of

00:06:38.560 --> 00:06:45.639
a job and they exceeded their target but here's the thing even though their

00:06:43.440 --> 00:06:49.800
tested units exceeded the target they're not confident that every single unit off

00:06:48.280 --> 00:06:54.319
the line is going to hit Platinum efficiency so they decided they'd rather

00:06:51.599 --> 00:06:58.800
be conservative with their ratings also in the case of Corsair the design team

00:06:56.520 --> 00:07:02.759
had just Plum finished the box and label designs with that gold rating on it so

00:07:01.479 --> 00:07:08.280
they ran with that we verify more than just 80 plus

00:07:05.919 --> 00:07:12.319
though which honestly is kind of falling out of fashion and part of the reason

00:07:10.039 --> 00:07:18.840
for that is 80 plus only factors in efficiency from 20% to 100% load so

00:07:16.080 --> 00:07:24.080
instead we measure all the way from 2% load to 110% so that you know how your

00:07:22.599 --> 00:07:30.240
power supply is going to perform when your computer's asleep all the way up to

00:07:26.240 --> 00:07:31.599
when it is oops slightly overloaded now

00:07:30.240 --> 00:07:36.879
I can't say that it'll make a big difference on your power bill if you

00:07:33.240 --> 00:07:39.520
pick up a c850 or c1000 from Enz XT both

00:07:36.879 --> 00:07:44.199
of which have outstanding 2% load power efficiency LTT hat Pro is off to them

00:07:42.400 --> 00:07:50.360
what I will say is that this expanded testing scope helps us to separate the

00:07:46.840 --> 00:07:53.080
overbuilt units from the underbuilt ones

00:07:50.360 --> 00:07:59.159
while testing at 110% we have registered four power supply deaths total including

00:07:55.720 --> 00:08:02.120
uh this one which is not a good result

00:07:59.159 --> 00:08:06.720
in fairness to these power supplies 110% of rated power is out of spec and we

00:08:05.120 --> 00:08:10.599
perform some of these tests with the input voltage and frequency being

00:08:08.440 --> 00:08:15.919
slightly off to simulate non-ideal conditions but thing is with all of the

00:08:14.159 --> 00:08:20.479
protections that a power supply is supposed to have an overpower or an

00:08:18.440 --> 00:08:25.520
overcurrent event really shouldn't result in outrate Failure but sometimes

00:08:23.199 --> 00:08:28.680
it does the game max power supply that we featured in the thumbnail the one

00:08:27.080 --> 00:08:33.560
that Lucas is looking at here spectacular failed during this very test

00:08:31.680 --> 00:08:37.839
the AC Source on the tester went into protection mode and when Lucas popped

00:08:35.399 --> 00:08:44.159
the power supply out of the test chamber he wisely fired up his phone camera

00:08:40.800 --> 00:08:46.360
plugged the unit into the wall and yeah

00:08:44.159 --> 00:08:54.880
you don't want that going on inside your PC but I mean lonus and Lucas that was

00:08:51.120 --> 00:08:57.160
probably just one bad unit right no see

00:08:54.880 --> 00:09:01.640
here's the thing anytime a power supply dies during our testing we obtain

00:08:59.000 --> 00:09:08.480
another one and almost every time the second unit has also

00:09:05.240 --> 00:09:10.279
died across our overcurrent overpower

00:09:08.480 --> 00:09:15.320
and ShortCircuit protection testing we've found six power supplies that have

00:09:12.240 --> 00:09:19.000
gone to sleep permanently securing our

00:09:15.320 --> 00:09:21.560
coveted we cannot recommend award see

00:09:19.000 --> 00:09:27.040
while catastrophic failer protection is technically a recommended part of the

00:09:23.959 --> 00:09:28.800
ATX specification we feel that any power

00:09:27.040 --> 00:09:34.200
supply that you spend good money on should do three things one it should

00:09:31.560 --> 00:09:39.279
protect you we can't have your house and there for you Lighting on fire number

00:09:36.680 --> 00:09:43.360
two it should protect your computer you can't have your motherboard and

00:09:40.800 --> 00:09:48.040
therefore your computer Lighting on fire and three it should protect itself we

00:09:46.279 --> 00:09:52.760
can't have your power supply and therefore your power supply Lighting on

00:09:50.399 --> 00:09:57.560
fire producing molten material or making startling

00:09:54.200 --> 00:10:00.120
noises another surprisingly effective

00:09:57.560 --> 00:10:04.240
killer has been our Brown out test which has claimed the lives of these three

00:10:01.920 --> 00:10:09.120
power supplies this test is pretty simple you know when the lights flicker

00:10:06.240 --> 00:10:12.959
briefly well we're testing to see how long that needs to last before your

00:10:10.880 --> 00:10:16.800
computer will shut down we're not even looking at surges or anything like that

00:10:15.160 --> 00:10:21.959
it just turns out that some power supplies really don't like power getting

00:10:19.399 --> 00:10:26.360
turned off briefly and then back on a surprising non-k killer though is our

00:10:23.920 --> 00:10:33.440
Excursion testing where we hit the power supply with a 150% load for 1

00:10:29.600 --> 00:10:35.519
millisecond and then a 200% load for. 1

00:10:33.440 --> 00:10:41.560
milliseconds to see how it'll handle something looking at you

00:10:38.399 --> 00:10:44.000
gpus briefly drawing way more than it

00:10:41.560 --> 00:10:48.200
should while this hasn't killed anything yet there are differences in performance

00:10:46.279 --> 00:10:51.560
in these tests so if you plan on overclocking you're going to want to

00:10:49.519 --> 00:10:56.800
look for a power supply with all green check marks here now that we know what

00:10:54.040 --> 00:11:00.920
we're testing for let's set up a test why don't you explain step one because

00:10:58.720 --> 00:11:03.920
it's not obvious to me looking at it yeah so the first thing we do we unbox

00:11:02.519 --> 00:11:07.519
it find out any of the manufacturer's claims and write a lot of it down for

00:11:05.720 --> 00:11:10.800
metadata that later gets up the loaded to the website then we go through input

00:11:09.240 --> 00:11:14.480
and output ratings for the power to make sure that we're loading it correctly and

00:11:12.279 --> 00:11:17.800
that the loads are properly portioned to it and configured for that so we're not

00:11:16.000 --> 00:11:20.560
nailing it on one rail and then leaving another one alone and then we go through

00:11:19.040 --> 00:11:23.680
and we decide what tests thrun on it which is exactly the same for all of our

00:11:22.200 --> 00:11:26.519
standardized testing that we go up on PSU circuit and the lab's website and

00:11:25.200 --> 00:11:29.720
then we take kind of an inventory of all the cables that comes with it and quick

00:11:28.240 --> 00:11:34.600
measurements of the length so we have that data as well and then we can just run this cell in Python and some code

00:11:33.440 --> 00:11:40.040
I've written on the back end uh generates a whole mess of data and files

00:11:37.720 --> 00:11:43.000
that we can run off of I said step one you gave me what sounded like at least

00:11:41.519 --> 00:11:46.279
like half a dozen steps that's great that's efficient step one 80 plus

00:11:44.959 --> 00:11:49.920
Platinum this guy it's not even titanium I think

00:11:48.279 --> 00:11:54.519
titanium is better next it's time for Lucas to turn on the tester and the

00:11:56.040 --> 00:12:01.360
scope there a lot of buttons yeah now I

00:11:59.240 --> 00:12:04.920
can even reach most of them next up we shove our power supply into the chamber

00:12:03.240 --> 00:12:08.760
and reposition our thermal couples according to the exact physical

00:12:06.360 --> 00:12:12.639
dimensions of the unit then it's time to plug into this guy connector board

00:12:10.959 --> 00:12:15.839
interface board I don't know what it's called and it doesn't matter but shout

00:12:13.959 --> 00:12:21.240
out cic for sending this thing over to us because it saves a ton of time we can

00:12:19.040 --> 00:12:25.800
plug in as many connections as you could possibly find on a modern power supply

00:12:23.480 --> 00:12:30.680
and the best part is this was some special work from Lucas some tedious

00:12:28.120 --> 00:12:35.079
work from Lucas we take our measurements right at the bottom of where everything

00:12:32.839 --> 00:12:39.480
plugs in which is important because otherwise the length of these wires

00:12:37.600 --> 00:12:43.279
could affect our readings and we don't want that oh there's one more thing we

00:12:41.480 --> 00:12:48.000
have a thermo couple that we plug right into the uh 12vt high power you know the

00:12:46.440 --> 00:12:54.600
one that has a reputation for being kind of melty once everything's hooked up we

00:12:51.519 --> 00:12:58.760
start our tests at 0 degre C then we

00:12:54.600 --> 00:13:01.160
move to 20° C and finally 40° everything

00:12:58.760 --> 00:13:05.360
is a automated except the temperature of this chamber we could automate that but

00:13:03.639 --> 00:13:08.040
it's actually nice to have a few check-in times to make sure that

00:13:06.760 --> 00:13:14.839
everything is looking good with the results our hardest tests are only done

00:13:10.639 --> 00:13:16.760
at 20° C 16 hours later we get this

00:13:14.839 --> 00:13:20.800
which as you can tell needs some processing but fortunately that's

00:13:18.399 --> 00:13:25.160
automated by some code that Lucas wrote then from there we get a script template

00:13:22.959 --> 00:13:29.160
for PSU circuit videos that gets filled in with the results that are sent to

00:13:27.240 --> 00:13:34.240
Emily for the voice in the videos we we use 11 Labs an AI text to speech dubbing

00:13:31.639 --> 00:13:37.680
service and specifically we use Brian because we like his voice the most it

00:13:36.040 --> 00:13:41.399
can take a couple of tries to get a good read from Brian and I need to balance

00:13:39.480 --> 00:13:47.480
how natural he sounds with how accurate his reading is some more Niche phrases

00:13:43.600 --> 00:13:50.800
like ATX tend to trip them up 1 atx3

00:13:47.480 --> 00:13:52.440
2024 manufactured by cesic each video

00:13:50.800 --> 00:13:56.880
has a template with parts of the voice over that don't change in blue down here

00:13:55.079 --> 00:14:01.160
and then I insert the generated voice above that in green then the footage and

00:13:59.160 --> 00:14:04.800
Graphics get added in and after 2 to 3 hours of work we have a video now if

00:14:03.079 --> 00:14:08.800
you've ever watched PSU circuit you're probably wondering why don't you guys

00:14:06.720 --> 00:14:14.120
just have a human voice the videos and then have a human fully edit it and I

00:14:11.160 --> 00:14:17.959
get it I would genuinely love to do that but let me put it this way so far on

00:14:16.519 --> 00:14:24.759
this channel we have made an eyew watering $138 in AdSense

00:14:22.279 --> 00:14:29.519
$138 plus a few bucks here and there from private internet access VPN

00:14:26.680 --> 00:14:35.240
Affiliates and that's across all of the so the economics of humans doing that

00:14:32.120 --> 00:14:36.399
stuff they just don't make any sense but

00:14:35.240 --> 00:14:41.959
that doesn't mean that the Channel's going anywhere we are committed to continuing to test power supplies and

00:14:40.320 --> 00:14:46.160
continuing to upload these videos because the data isn't just valuable to

00:14:44.079 --> 00:14:50.920
viewers of the videos it's also incredibly valuable to LTT's ability to

00:14:49.000 --> 00:14:55.759
do build videos that feature power supplies that we can truly stand behind

00:14:53.600 --> 00:14:59.680
also if you don't like the AI voice you're not alone but you can just read

00:14:57.600 --> 00:15:03.680
the articles at LTT Labs.com you'll even find some extra data there

00:15:01.320 --> 00:15:08.279
you might even find this segue to our sponsor Microcenter March is Monitor

00:15:06.240 --> 00:15:12.079
Madness month at Microcenter try saying that three times fast they have great

00:15:09.839 --> 00:15:18.120
deals on monitors of all shapes sizes and specs like this Alienware 34in OLED

00:15:14.639 --> 00:15:19.480
175 HZ refresh rate Beast for $400 off

00:15:18.120 --> 00:15:24.000
whether it's something simple for your work setup or a companion for your Ultra

00:15:22.120 --> 00:15:27.639
powerful gaming rig they have something for everyone plus if you want to save

00:15:26.120 --> 00:15:31.959
even more money you can take advantage of their GPU trade inogram

00:15:29.560 --> 00:15:36.240
if you purchase the GPU at Microcenter simply bring it back and receive an

00:15:33.720 --> 00:15:40.279
offer on the same day they can even help you recycle your old Tech or help donate

00:15:38.800 --> 00:15:43.440
it to someone in need if you need to make space for your new purchase so

00:15:42.120 --> 00:15:47.959
click our Link in the description to check out their monitor Madness event

00:15:46.079 --> 00:15:52.839
we'll also have another link there where you can get a free 128 gig flash drive

00:15:50.440 --> 00:15:56.399
if you visit their new Santa Clara store when it opens this year if you guys

00:15:54.759 --> 00:16:00.880
enjoyed this video why not check out the one we did on a 3,000 W power supply

00:15:58.639 --> 00:16:04.880
where you can see what our testing capabilities used to be