Something went wrong with the small watercooled PC...
Something went wrong with the small watercooled PC...
2018-12-03
yesterday we finished building this
piece of crap but today we had to fix
this piece of crap
this holiday season spice up your build
with high-performance options from MSI
like the Emmy Jeezy 390 god-like which
supports up to 40 600 megahertz memory
or their wide range of graphics cards
from AMD and NVIDIA EMA site also has
full product lines including bare-bones
workstations gaming peripherals monitors
and more to see all that MSI has to
offer click the link in the description
below just because you can do some
things doesn't mean you necessarily
should I was a little concerned when I
started building this system with how
densely packed it was becoming that
there may be some cooling issues in fact
a lot of you within my last video were
like hey where are the cooling numbers
where's the temperatures where's the
benchmarks why is I setting into that
video I said load all my operating
system and stuff on there and that's
what we've done today and what we have
discovered is that there are definitely
some cooling inefficiencies in this so
if you take a look right here you can
see just how densely packed it is
there's not a lot of room on the bottom
the feet only stand about four
millimeters off the floor so I've got it
propped up right now because I can put
bigger feet on the bottom to get more
airflow in though you know through the
bottom of the case this is an updraft
case heat rises therefore we are trying
to kind of have the natural you'll rise
of heat give us some convection to where
we're blowing the heat through the case
and up through the top the problem is
with the motherboard right there and the
i/o shield we lose some volume there
this pump uses up quite a bit of volume
plus it puts off heat into the case
because of this heat sink our RAM takes
up some volume our power supply takes up
a huge amount of volume which means that
that much airflow going up through the
bottom is already blocked we have our
ribbon cables along there we have our
tubing which is pretty warm we'll talk
about that in a second and then of
course we've got the top radiator right
here which as you can see once again is
blocked significantly by the ribbon
cable it's also blocked significantly by
the power supply and we are also taking
the heat of the power supply and blowing
it right into the radiator so here's
what this has given us we have a GPU
right now a Titan X it stocks beads on
water that's sitting at 55 C naught
terrible but approximately tendency
higher than I would see them normally in
a larger case with larger radiators and
better air flow so we already know that
that is causing us a problem in terms of
just heat and saturation now I have 4 ek
Vardar 2200 rpm fans on this and even
when I run them at full speed we still
sit about 52 C on the the GPU now the
CPU the 9900 K this was a little bit of
a concern of mine is a hot CPU we know
that it has the s Tim or the solder
throw thermal interface material we've
also seen some reports that it's not
good better than the standard thermal
material but they've taken something
made it slightly better but crammed
twice as many cores in the same small
package so basically you've got 8 cores
and 16 threads worth of hyper-threading
inside of a you know 1150 weighs 1155
socket thick run whatever it is and it's
a it's a lot of focused heat and even
though I have a decent block in decent
pump on here the CPU factory voltages is
getting extremely hot so I'm gonna have
to do a couple of things today I'm have
to take this apart again which I said in
my last video I hoped I wouldn't have to
yeah I feel welcome crossing my fingers
wet I will show you guys right now in
Cinebench I will actually hit a hundred
C headstock voltages I've even down
clocked it I've got it running at 4.3
gig gigahertz on all cores
I'm not even letting any of the cores
get to 5 gigahertz like the factory
turbo will because the factory turbo
will go to 5 gigahertz on one core I
think it is and like 4.8 on 2 and then
it steps down from there right that's
the turbo that Intel Turbo I've turned
off all of the Asus features on this
because what I found was that I was
hitting 105 C and throttling when I
opened up Google Chrome now I've cleared
the min/max on here so you can see but
the package is idling at 42 to 48 C it
kind of moves around we've got as high
as 54 we're not under any load I also
have the fans running a hundred percent
speed right now as well as the pump you
can see that our voltages here on the B
core is only one point one that's
nothing one point one degree or one
point one voltage on a
like at 8700 K would have my
temperatures under load be somewhere in
the 50s so if I go ahead and run the
test right here watch our course granted
this doesn't have all what does have all
the eight cores here they're right there
so check cannot keep an eye on the the
core temperatures right there 70s 80s
9100 101 102 104 and now we're
throttling at 105 as you can see if we
look down here I'm throught only for 4
gigs 3.9 3.8 3.9 3.8 on some so we are
throttling and our score was only a 1780
one because the CPU was throttling I
don't know I think I feel like the
retention system on the Swift tech is
terrible so now we have to tear it apart
and see what we find
so part of my issue with this pump block
combo is the retention plate here is it
sits off the off the back of the
motherboard and it has these little like
foam pads the problem is those crush and
then as you can see it's sort of like
bows and I'm not sure we're getting a
really good contact here on the block
but the only we're gonna be able to
check that now is actually take it off
and I found it easier to remove the
whole motherboard than to remove just
the just the block because this brackets
gonna probably fall off in the back and
as you can see right here although
technically is a cutout right here
taking out the GPU and all that stuff
does not make it necessarily easier to
get to the back of this thing so I'd
rather just work on it outside of the
case it's just more screws ok so right
off the bat though I'm not too happy
with that thermal paste application if
you look at it I mean it doesn't go all
the way to the edge it's very thin I
suspect this one might need a little bit
of lapping as well where we have the
lose at this point
and it still spins a little bit so I
means the block itself could be high
centered okay so I lapped the CPU and
left the block
it's definitely flat now and cleaned out
the micro channels underneath too so
we'll be changing out the fluid
obviously so I'm gonna put this guy back
together I'm not expecting this bill to
obviously be amazing on temperatures I
just want something reasonable and I
dropped one of these screws so I have to
find it I've always got a screw loose
if you know what I'm saying okay here we
go
lupus together my own special formula
mix just drinking water and another
concentrates but I digress let's go
ahead and you know why I was worried I
dropped coolant down inside of the power
supply ah
okay so right 24c inside of Windows
right now all the same settings we had
before so much better I'm gonna go ahead
and go back to my fan expert now and put
the fan curve back to where I had it
let's just go standard fan curve all
right moment of truth now with Cinebench
remember last time we ran this I hit 105
on the package 105 on some core is it
throttled way down to like 3.3 or
something like that I remembered it was
353 57 58 59 60 61 60 62 and it's still
sitting at 4.3 so this should be
significantly faster right now you hear
how the fans aren't ramping up I think
this is a combination of a couple things
I think it was a combination of lapping
both the 1873 I think we might be able
to actually put this back up to like the
4.7 and then allow a couple of course
ago five gigs like it's supposed to but
I think a couple of things happened here
one it was clear that the temperature
was getting too hot for the for the true
fluid from primo chill as after one day
it was starting to clog up and and like
solidify like jellify is the best way to
describe it
inside the micro channel inside of the
apogee drive - it's not a lot of space
there it's very thin so that could
really have been causing us a terrible
inefficiency of the block we also so we
change
shut the fluid we also left the CPU IHS
as well as the block I also left the IHS
on the CPU as well as the base plate or
the cooling plate for the block itself
and then we applied more thermal paste I
did the X pattern rather than just the P
method like I did in the first one and
our temperatures are significantly
better so I'm gonna try I'm gonna try I
want to try five two gigahertz 1.2 but I
don't think it's way enough voltage see
if it crashes though 77 ATC 81 82 83
fans kicked up we are still 25 °c below
t.j.maxx
and we know Senate bench pushes harder
than Mike Adobe does sorry there 21:36
on our score and our max temp on any
core was 85 and 85 on the package so and
so 5 5 gigahertz all core super doable
it fills laughing cuz he gets to play
with this thing this is his little
travel rig so of course he's happy
alright guys we're gonna go thanks for
watching this video sorry we had to do a
third part on this but these are the
type of videos that are important
because seeing us put it all together
and stuff is one part of these builds
but when you have a problem finding out
the problem is it's just as important
and the other thing I changed I forgot
to mention earlier was I came up with
that better retention system on the back
because I told you how it was kind of
bowing down
so I basically used plastic washers
against the motherboard and then metal
washers against the bracket so there's
no short - to make a shim basically so
that it couldn't smash down because what
I feel is happening is the I feel like
there was a little bit of a bowing
happening with the bracket on the top
which was causing us to not get a good
even seat of the block versus the CPU so
we lapped it didn't really polish it it
was pretty shiny lapped the CPU block
and came up with a new retention system
changed out the thermal paste changed
out the coolant and this is what we got
I think most of the improvement we saw
today though was specifically the
retention system and the lapping of the
CPU so there we go I feel so much better
now is worried does it tear the system
apart and then be
guys I'm sorry but yeah we're done now
thanks for watching guys and as always
we'll see you in the next one
now lobby guys really like this small
form-factor build I did but if that's
not your thing in size matters then
you're definitely gonna want to check
out this one
We are a participant in the Amazon Services LLC Associates Program, an affiliate advertising program designed to provide a means for us to earn fees by linking to Amazon.com and affiliated sites.