Why Guess ?

[Witchdoctor]

I am new the the forum here.

Thought I would share this with the community

For years when setting up water cooling systems I have simply made my best guess at heat loads based on Max TDP of said components, I have recently put that all behind me with a simple watts calculator. you simply plug it in and then plug your you PC into it and it gives you an acurate load in any increment you would like, in this case watts. This information is helpfull when trying to calculate what type of heat you have to disapate that aids in sizing of Rads and what type of RPM you will need with said surface area's you decide on. It also tells you where you are at with your PSU be it enough or if it is struggeling

Here is a test I ran this morning using my sig rig with Furmark SLI and LinX both running at 100% on the CPU and GPU's with a modest overclock

This is the best 20 bucks I have spent in a long time providing definative information instead of the typical speculation

 

[Bobnova]

Just as fair warning, those units can be pretty bad at reading APFC loads sometimes. OW's talked about them a little bit here in there.
Usually they're within 10% though, which is good enough.
I have one, they're cool.

 

[MattNo5ss]

Kill A Watts are okay for estimating overall system power consumption, but I don't think they are too good for estimating heat output of your water cooled components. You would probably get something quite a bit higher from the at-the-wall monitor. There's nothing wrong with building an overkill loop though, so at at least the Watt reading will be incorrect on the high end instead of being too low.

EDIT: Off topic...Witch Doctor is my favorite D3 class

 

[Mandrake4565]

Wow and I thought my Fx 8350/580 Lightning pulled a lot of watts from the wall when pushed!
View attachment 142761

 

[Bobnova]

If we assume for a moment that they are exactly accurate (they're usually close enough), they tell you to within probably 0.01% how much heat is being generated inside your case.
If we take the psu efficiency out of the equation (multiply the read value by psu eff at that wattage), then take out the wattage eaten by things not in the wc loop (ram, moto, hdds, etc) you have a pretty good idea of how much heat the loop is dealing with.
All the electricity consumed by the computer turns into heat eventually. All.
The only part that does not turn into heat immediately inside the case is the energy actually put into the air to induce movement that stays in the air as kinetic energy. Even that becomes heat as the air slows and stops.
I really do mean all, too. Not most, not some, all.

 

[MattNo5ss]

If we take the psu efficiency out of the equation (multiply the read value by psu eff at that wattage), then take out the wattage eaten by things not in the wc loop (ram, moto, hdds, etc) you have a pretty good idea of how much heat the loop is dealing with.

This is what I had in mind, PSU efficiency and things that aren't in the loop need to be taken into account for a more accurate estimate.

 

[GTXJackBauer]

Yup, I got 2 of those. Nice to see what the system approximately is pulling off the wall. I usually don't use that for my heat load calculations but is another great tool to use as well if you get the other components subtracted as its been said. When I just want to see what my rig is doing at the time being I use my digital PSU's software and this unit. Both are within a few watts of each other which is great.

 

[Witchdoctor]

Thanks for the input gent's


I was thinking more along the lines of Bobnova as all electrical current at some point turns into heat energy, be it the board, memory, spinners and so forth, So if is agreed upon these components generate heat due to said electical input that would ultimantly be pulled through a rad at some point for dissipation, would they not have to be accounted for in a heat load calculation ?


Edit:

What's the skinny on sizing screenshots, can not get the screen shot in the OP readable ?

Thanks

 

[EarthDog]

IMO, not really, no. Depending on the loop's physical configuration, the heat off the not included in the loop parts may not even pass through a rad on it's way out of the case so how could that be included? Regardless, radiant heat of said parts versus the vrm, memory, hdd actually being in the loop are completely different things to measure. So, while there are likely considerations for such things it's not remotely like having those components actually IN a loop capturing all the heat and not having it diluted(cooled off) by the ambient air.

This method will overestimate the heatload... Which is never a bad thing really.

 

[Witchdoctor]

Ok, I was thinking along the lines of the radiant heat off said components would raise case air temps effecting the efficiency of the rads there fore having to be accounted for, Appreciate the input, kinda new to this type of stuff.

So is it a fair statement to say to get an accurate account of the heat load one would only have to account for what is physically in the loop itself and all other associated components can and should be discounted ?

 

[EarthDog]

You are spot on above. I wouldn't say completely discounted but I woukdnt add that wattage to the parts you are physically cooling though. Ok, I guess I am saying.don't.worry about it, lol. Just know.for every degree C your intake air increases so.do your temps.

That load inside the case can also be dissipated with case airflow. Only count what is in your loop to be cooled is my take.

 

[madcratebuilder]

This is what I had in mind, PSU efficiency and things that aren't in the loop need to be taken into account for a more accurate estimate.

I monitor the wattage at the AC outlet with a Kill-A-Watt, I'm able to monitor wattage with Cyberpower gadget, great for showing wattage increase when you load a devise like the gpu, cpu or even when doing file transfers or dvd burning. I also monitor the PSU with the Link2 software for the AX1200i.

This gives my a general idea of how much energy is being consumed and with a little math I calculate the heat energy removed by the Rads. I now use the Aquaero software to display this info.

Aquaero manulk

The aquaero calculates the current power dissipation using the temperature difference between two temperature sensors and the current flow rate. Water has a thermal capacity of 4187 Ws/(kg * K), meaning an energy of 4187 Ws is required to increase the temperate of 1 kg of water by 1 Kelvin. Power = thermal capacity * volume flow * temperature difference
Due to inaccuracies of temperature and flow sensors and temperature inertia
of the cooling system, power measurement is a bit problematic in general.
Especially during system start up, the actual power dissipation will differ significantly from values calculated by the aquaero. Furthermore, any coolant
additive alters the thermal capacity, this is factored in in the aquaero. The
power measurement functionality should therefore just be seen as an interesting additional information.

 

[RnRollie]

unfortunately, the whole point is moot IF you assume the total wattage consumption = heat AND take radiant heat INSIDE the case into account... because... if you would dimension your raddage for the TOTAL comsumption,... they wont fit in your case
Therefore putting radiant heat from non-water cooled components into the equation is moot
But at least the system would be overdimensioned , which is a good thing

 

[EarthDog]

unfortunately, the whole point is moot IF you assume the total wattage consumption = heat AND take radiant heat INSIDE the case into account... because... if you would dimension your raddage for the TOTAL comsumption,... they wont fit in your case
Therefore putting radiant heat from non-water cooled components into the equation is moot
But at least the system would be overdimensioned , which is a good thing

You are spot on above. I wouldn't say completely discounted but I woukdnt add that wattage to the parts you are physically cooling though. Ok, I guess I am saying.don't.worry about it, lol. Just know.for every degree C your intake air increases so.do your temps.

That load inside the case can also be dissipated with case airflow. Only count what is in your loop to be cooled is my take.

 

[Witchdoctor]

Would it not depend on what case you have?
I get that you would only take into account the actual load of the components in the loop and typically radiant heat would be dealt with through intake and outtake fans

I have a set of circumstances where all my intake air comes through a rad and all of my outtake air goes through a rad as well

so the only escape for the air is through a rad. I assume I would still ignore those radiant heat loads from the other components or with this set of circumstance would it need to be calculated in ?

Sorry for the dumb questions, like I said above I am no expert at this, but trying to get my head wrapped around this as I see many here and other places under sizing rad surface area and RPM's on fans servicing said surface areas form what I ascertain is their true heat load. I would hate to have someone move in the wrong direction from something I stated that was not correct.

Thanks to everyone for the input, good stuff

 

[RnRollie]

yes, it depends on case
but... in almost all cases, it is best to have the air intake through the rads, and exhaust elsewhere.

for example if rads on top & front, set the fans as intake, and dump the "hot" air out the rear.
Yes the temps inside the case will be slightly higher, but it wont kill anything... unless you are so underdimensioned that the loop reaches boiling and the air through the rad jumps from ambient to +70°C ... but in that case you have other worries

 

[EarthDog]

Would it not depend on what case you have?
I get that you would only take into account the actual load of the components in the loop and typically radiant heat would be dealt with through intake and outtake fans

I have a set of circumstances where all my intake air comes through a rad and all of my outtake air goes through a rad as well

so the only escape for the air is through a rad. I assume I would still ignore those radiant heat loads from the other components or with this set of circumstance would it need to be calculated in ?

Sorry for the dumb questions, like I said above I am no expert at this, but trying to get my head wrapped around this as I see many here and other places under sizing rad surface area and RPM's on fans servicing said surface areas form what I ascertain is their true heat load. I would hate to have someone move in the wrong direction from something I stated that was not correct.

Thanks to everyone for the input, good stuff

Again, for intake->rad->case->rad->exhaust setup, see what I said above (every degree C your intake temperature raises, so will your overall temps).

People undersize rads because they are generally clueless of all the data out there say from our stickies here or from Martin and Skinee labs who actually test this stuff and show how air flow and ambient temperatures affect the loop. That and perhaps their goals are different with water cooling. So while people like you or I or benchers want the absolute lowest temps possible, others do not mind air-like temperatures and quiet. But those two sites (Martin/Skinee) really discusses a lot of what you are questioning. You may want to read the stickies as well as they go over heat load... not once mentioning radient heat from internals I may add.

GENERALLY, and there are many exceptions, but for a high end gaming rig, a general rule of thumb that will likely not leave one underradded is 2x120 for each part in the loop. Again, plenty of exceptions, but for the vast majority of people that frequent here or OCA, that is plenty for better than air temps and quiet operation.

EDIT: As far as being case dependent... of course it is. That said, I see people suggesting to 'regular' users, for a top mounted radiator, to intake through the top of the case and have it exhaust out the rear. While this will likely drop temps from 1 to a couple of degrees C (remember I have said for every degree C your ambient/intake air rises so will your temps...), One has to consider if that is even worth it. I would rather have better airFLOW (front/sides = intake, top/rear = exhaust) than to gain a degree or two.... even as a bencher... but that is just me.

 

[Witchdoctor]

Thanks guys,

I have spent a lot of time reading Martin and Skinee testing and results and agree, no where can I find radiant heat addressed ???
One would ascertain it does not need to be addressed

It would be interesting to compare performance of identical systems using loop load only and total load to see what type of performance differences we would see and how it effects the dt's

 

[EarthDog]

You would have to use the same system, loop, TIM application, and mount to minimize variables. Two different systems with the same hardware would still yield different results unless you only want to measure Δ and not actual values. Even then though, results can vary. I mean you know that each CPU is different. I could put a line of 4770K @ 4.5Ghz 1.3v CPUs and each would read a different temp for MANY reasons (leakage to name one, mount, TIM, etc).

You would literally need the ability to mount the/a rad in the case, and be able to take it out and not be affected by case temps...or the easy way, as I have said a couple of times in this thread already, temps raise 1:1 for every degree C difference so subtract the difference between the intake air temp (probe in front of intake fans) and the intake temperature of the rad inside the case (probe in front of that rad).... that would be your difference without physically testing it.

 

[Witchdoctor]

Agreed, it would be difficult to have variables remain static. Even ambient temps would effect results


Thanks again everyone for your participation in this thread.