- YouTuber Tech Ingredients made some pretty bold claims recently saying that their
homemade thermal compound can outperform commercial
solutions like Arctic Silver. So naturally, immediately
after that video dropped, we received dozens of emails
asking us to test it out. I mean, Arctic Silver has silver in it. Could anything be more
conductive than silver? Tech Ingredients sure thinks so. So we got our hands on
it, and we are going to be with full attention to all
the warnings trying this out. And the video is brought to you by AORUS. The AORUS 15G features an
Intel 10th Gen processor with strong battery life and performance and tons of features. Check it out today at the
link in the video description. (upbeat music) The craziest part of this is
not only does Tech Ingredients say that this will outperform
a commercial solution, he also says that you can make it at home. As long as you've got some
dendritic copper, epoxy, graphene, diamond and access
to a high powered ultrasound or a cell disruptor. We just went ahead and paid
the 20 bucks or whatever it was to have him send it over. Now, I'm really excited to test this, but there is one major caveat. As is laid out in the instructions,
this bond is permanent, meaning this is more like a thermal epoxy and less like a thermal paste. So when I was putting together
my test bench for today, I had to consider a couple of things. Number one is I needed something
that output enough heat that I was going to be
able to tell the difference between my reference
thermal solutions over here and the Tech Ingredients' thermal epoxy. And number two, I needed a test bench where I wouldn't be
too choked up if my CPU was permanently bonded to
the bottom of my cooler. So I chose an AMD FX-8150. This thing was known for being
an absolute inferno of a chip and not real fast. And this thing. This is the Thermaltake Black Widow SpinQ. I'm not actually sure what it's called, it seems to have two names. Universal CPU cooler. Been a while since I've
dragged the old Crosshair IV out of mothballs. One thing that I went and grabbed is some of these Noctua cleaning wipes. Normally, I'm pretty loosey goosey about the way that I clean CPUs in between thermal paste applications, but because we're
actually trying to measure the difference in performance
between thermal pastes, it's pretty important
that we do a great job of cleaning our CPU here. Let's get some RAM installed. Throw our period appropriate
graphics card on here. We're gonna start with what I'm expecting to be the least performant
thermal interface. This is the Innovation Cooling
IC Graphite Thermal Pad. And to be clear, I'm not saying it's bad, it's just that the sales
pitch for these pads is that they're reusable
and they don't make a mess. Not that their performance
is right up there with the top thermal pastes. So we just plunk that on there and we're ready to mount our cooler. Oh, I actually lied to
you guys a little bit. I think this is going to be the least performant thermal interface. Let's try this one too. This is the included thermal paste with this kind of cheapo cooler. One thing that's nice about
using an AMD CPU for this that uses the clip-based hold down is we're gonna have
identical mounting pressure on all of our thermal interface materials. There's no real room for user error because you can't over or undertighten it. To make our lives easier, we're not gonna bother
with idle temperatures. That's not really where the
men get separated from the boys when it comes to thermal solutions anyway. So let's go ahead and fire
up a classic stress test for a classic CPU, smallest FFT, eight threads. And then we're gonna wait until
the CPU temperature reaches about a stable point,
and we're gonna go ahead and we're gonna click this clock down here to reset our average values, let it run for two or three minutes. 50 degrees under load qualified
as a inferno of a CPU, now we've got the Intel 10th Gen. I'm a little rusty on my FX overclocking, but about 4.5 gigahertz
with about 1.42 volts or so. (beep) After some trial and
error where we got reacquainted with the FX-8150's 70 degree
Celsius maximum temperature, we settled on 1.475
volts with turbo disabled and four gigahertz clock
speed giving us about 68 degrees under load
with our thermal pad. So now what we're gonna
do is reset our timer and take an average
over about two minutes. After a little over two minutes, we've got no thermal throttling, an average temperature
of 67.7 degrees Celsius and a maximum of 69.1. Nice! Now for our next thermal compound. This is nice, we don't have to clean it because all we had on there
was one of these thermal pads, but we're definitely
gonna have to clean it after our next attempt. What do you think Andy, quality thermal compound or
the gross included garbage? - [Andy] Let's do the included garbage. - Let's do the included garbage baby. Now at this point of the
game, I could wait around for the temperature to reach equilibrium or I could take nature into my own hands. By running the CPU at effectively what is a greater ambient temperature we can cause it to heat up to approximately where we
think it's gonna end up and then it should normalize
a little bit sooner. It's a free tech tip. Hey, look how fast it's heating up. The system works. Meanwhile, Steve over at Gamers Nexus with like liquid nitrogen pots. Wow, I knew that these thermal pads were optimized for convenience
rather than performance but dang, even the generic stuff
included with our heat sink is at, what's my average here? 60.1. Now, I'm really not sure what to expect from a quality thermal paste. Are we gonna drop like
another seven degrees? I kinda doubt that. Oh, I always forget about
this, even though it's actually the reason I chose an AMD
chip because obviously, once the CPU and the heat
sink are glued together, I'm not gonna be able to
get at the release mechanism so I wanted something
that worst case scenario I could just yank out of the socket. You gotta twist, ladies and gentlemen. There that's better. Let's get this bad boy cleaned up. These Noctua cleaning
cloths are next level. That is a clean FX CPU,
ladies and gentlemen. No harm no foul. Do we really need to use two
high quality thermal pastes? I think just one is fine. Do you want Noctua or MX-4? - Noctua.
- All right. So I want a very similar amount. But I'm also gonna make
sure that I spread it out. Man, it's weird seeing
the old green AMD logo. Here on the desktop, the
shortcut is the old green logo. I totally forgot it used to be green, and then they started
competing in graphics, and it's like, oh! Well, we can have team
green and team green. Bang on Andy. 59.1, you nailed it, one degree better. Before using this product,
you might want to review some principles of using
thermal interface materials demonstrated in our
Tech Ingredients video. Wear gloves, protect your work surface with some disposable plastic film. Contact with epoxy can irritate the skin. Got it. After preparing and cleaning the surfaces remove the caps on the syringes and inject the quantity you plan to use. The mixing ratio is three
parts large orange syringe and one part small black syringe. Watch out, the orange
syringe is more viscous. So I'll need to be a little bit careful to get the right quantities here. - You don't want too much.
- Ah, too much! Dang it, they even warned me. I knew this was gonna happen. Does that look about twice
the size of my orange ones? What do you think?
- Yeah. - Okay, we'll do three
more orange ones, shwamp. Good. I love science, the instructions
written for scientists. Stir the components
thoroughly until the color is a completely homogenous dark brown and then mix for another 30 seconds. I'm gonna use this
plastic McDonald's knife. Homogenous and dark brown. Wow, it's kinda grainy,
which I guess makes sense given that it's made of like
little grains of diamond, graphite, and-- - Yeah, that's grainy.
- Dendritic copper. Okay, we're gonna mix it for
another 30 seconds though. I'm very specific about that. I am going to follow
instructions for once in my life 'cause this stuff seems like
serious science chizz here. All right, wow, it really starts
to thicken up a bit there. Interesting. Okay, that seems like about
30 seconds more of mixing. Wow, all right, cool. Apparently we've got about 45 minutes before we can't work with it anymore. Now, we need to put a
thin layer of this stuff on each of our surfaces. This is not the best
spready spreading tool. You know what, I'm just
gonna use my glove. So those are on there. Now get this, there's an
element to the instructions that I actually didn't read until now. Once they're on there,
you're gonna wanna grab a fingernail sized piece
of medium sandpaper, and abrade the epoxy
mixture for a few seconds into each surface before
placing them into contact. Oh yeah, here's some 400 grit sandpaper, which I guess is mediumish. This hurts me so much. We're going heavier guys. I made the mistake of following
the written instructions without looking carefully
at the video instructions. And it seems like he put it on much, much heavier than I expected. So I'm just gonna go ahead, I'm gonna mix up a whole bunch more. And we're gonna try and get
this on there before it cures. So we're going heavier
here, we're going heavier. Here we go. I guess this is the part of the video where I acknowledge that we
shot this over a couple of days because of that whole issue with the 69 degree thermal limit. So I'm putting this on as
far as I can tell correctly from Tech Ingredients' video. Obviously, he didn't
apply it to a CPU but, hey, that's what you guys need me for. It's definitely on there. Andy, what's gonna happen?
- Kaboom. - No, shh, no, what's
actually gonna happen? Do you think it's gonna
outperform the paste? - I think so.
- It's glue. - Yeah.
- It's thermal glue. Dendritic copper.
- Yeah. - Diamond. Man, I hate this when
you've got one dirty finger from what you were doing before and you gotta wash your
whole hand, you know? I'm just gonna take it to 59
'cause then we're gonna see whether it keeps going up or goes down. This is interesting. It's a good thing we left
a little bit of headroom between the thermal pad and the 70 degree
thermal throttling limit, because you can see here, we are actually in between pad performance and that 70 degree limit with the epoxy. Fortunately, we're not throttling, so there should be the same
amount of thermal output and our result should still be valid. I spoke too soon. During our two minute
run, we thermal throttled. You can see all these cores
dropped down to 1.4 gigahertz. They ramped back up. You can see we're running
at full speed now. But it clearly was outperformed by every
other solution we tested. Now I kinda feel bad about
the way I presented the intro because I had intended
to like seem skeptical, but I actually thought that
this was gonna outperform it. I didn't mean for this to just be like, dunking on their thermal compound. It's the next day, it
occurs to me that maybe our thermal epoxy doesn't
perform at peak efficiency until it's actually hardened. So we're gonna fire the machine back up, I let it sit, not running,
'cause I wasn't sure how having it be hot would affect the curing. And we're gonna take one
more crack at it here. Just checking the numbers from yesterday, I think it's gonna perform
quite well compared to them. Wanna see a nice flat line before we start our two minute run though. It starts out high from
our heat gun shortcut, then comes back down and now it's creeping
its way back up again. I'm trying to wrap my
brain around this behavior. How can you have the same
thermal interface material in there that performs differently over time? It seems to be very
clearly doing that though. Maybe Tech Ingredients could
do a follow up explaining this. I'd love to know, like is it expanding? How long should it take
to reach equilibrium? I'm disappointed, I was
ready to come in this morning and say, puh, well, stupid me, I shouldn't have been testing
it before it was cured anyway. But we're actually seeing pretty much the same behavior at this point. We just touched 68 degrees for the first time a
little while back there. And you can see we've got a
pretty consistent trend upward. Right as I was saying
that it thermal throttled, so two of our cores dropped
down to 1.4 gigahertz. That we can't give it a
temperature score at that point, that's just a fail. With thermal performance worse
than even our graphite pad, it's safe to say I won't be recommending Tech Ingredients' thermal epoxy
for your CPU anytime soon. But it's also safe to say that they never recommended it for that either. It's for situations where you cannot use a higher performance thermal grease. And you know what, I think my expectations were probably a bit unreasonable here. But given that the only
reference that I had for its relatively high performance was a competing thermal epoxy that nobody in their right
mind would put on a CPU. I just didn't really
have a frame of reference and neither did you so I can understand what you guys wanted me to test it for. All I can hope now is that
it's performance as an epoxy is disappointing, and at
least I can get my CPU back. Okay, I think it's glued. But as we've seen before,
it should still be possible to pull the CPU out of the socket. Yeah, there we go. So it looks like A, we
definitely had enough to cover the entire IHS
'cause we've got that seepage all the way around the edge. And B, we've obviously got a fair amount of mounting pressure on
here because otherwise why would it seep out from
around the edges there? Remember how I was able to
twist off the old thermal paste? Well, the epoxy component works. (laughs)
Shucks! Oh, my goodness, look how
durable this stuff is. I'm scratching at it with a knife and it's not even entirely coming off. I know sometimes blunt force can be better for cracking epoxy. (loud banging sound) So we have no choice but to keep going. Enter the Powerfist. Oh! Oh! Wow, I think it's kind of a 30% chance this thing still works after
what just happened to it. But what happens now if I just
put like a thermal pad on it? Oh, wow! That's bent. I'm not even gonna try
to mount it properly. And I'm just gonna see if it posts. Well the CPU still works. I mean, we can never use it for certain kinds of testing anymore because it's covered in epoxy, but at least I can give it
back to the inventory guys knowing that my inventory
checkout sheet is clear. Just like I'm cleared to segue
into our sponsor Manscaped. Manscaped is featuring
their new Weed Whacker nose and ear hair trimmer. It features the same
proprietary skin safe technology to help reduce nicks and tugs
with a powerful 9000 RPM motor powering the 360 degree
rotary dual blade system. It's got a rechargeable lithium battery with a 90 minute charge. It's water resistant, and comes with an easily
replaceable blade. So check out their performance package which includes both the Lawnmower
3.0 and the Weed Whacker and get 20% off and free shipping by using offer code LTT at manscaped.com. We're gonna have them linked below. If you guys enjoyed this video and you like thermal compound experiments, maybe check out the one from
actually three years ago where we bought the
cheapest bulk gigantic jar of thermal compound that we'd ever seen and tested its performance.