Two Loops, One radiator

[troelsm]

I'm planning on cooling CPU, NB and Gfx and seeing as many watercooling blocks, especially some of the best CPU blocks, are quite restrictive, wouldn't it be better to run 2 loops rather than one.

I'm gonna be using the Typhon III reservoir and from there run two loops like this:

Res -> Radiator -> Res
Res -> CPU -> NB -> Gfx -> Res

Depending on how restrictive the full Gfx block is that might go onto the radiator loop too.

My reason for doing this is that then there wouldn't be any flow restriction on the radiator loop, like there would be if for example the CPU was in the same loop, and this should result in maximum heat transfer.

Would there be any problems doing this? The only potentiel problem I can see is if the water doesn't get mixed well enough in the reservoir and alot of the warm water from the non-radiator loop would be send round and round. Considering that the Typhoon III is designed by a real enthuisiast and not some company clown i don't expect this to be the case though.

Since i don't think i've ever seen this before I'm kind of thinking I'm overlooking something here, enlighten me

[hellcamino]

Did you forget your pump? Keep it simple, res-pump- then whatever routing gives you the shortest runs of tubing.

[bluehaze]

Did you forget your pump? Keep it simple, res-pump- then whatever routing gives you the shortest runs of tubing.

Pumps are mounted on the res with Typhoon III.

I guess this would work, it seems a less than optimal solution though as far as flow rates are concerned. You will have a really high flow rate through radiator and a really low flowrate through all the blocks which is exactly opposite of what you want. To an extent higher flow through blocks = lower temps, and lower flow through rad = lower temps.

It will be interesting to see though if you hookup a controller to each of the pumps the tuning options could make a significant difference finding optimal flow rate through the rad and optimal flow rate through the blocks seperately as long as you could get a decent flow through the blocks with just the one pump.

Edit: Giving it a second thought I'm not sure how well it would work with that particular res though because the inlet of the res is so close to the inlet of the pumps that the temps of water flowing through rad and water flowing through blocks on 2 seperate loops might never equalize or at least take an extremely long time to equalize which with the rad on a seperate loop would be really bad.

[AndrewZorn]

the first thing i thought when i read the topic is you have it completely backwards... how about one loop, two radiators!

but really, what youre doing is restricting the water's access to the radiator, which is the main point of having one. the water being restricted by the radiator happens because of the water being exposed to the maximum inner surface area of the radiator as possible, youre trying to create as much of an interface between the warm water and the cool walls of the radiator as possible.

if you only run some of the water through the radiator, youre submitting less of it to be dissipated into the outside air, while simultaneously submitting more of the water to the hot cpu block...

by using two radiators in one loop youre doing the opposite, all of the water sees more radiator time, instead of less water seeing less radiator time.

EDIT whether or not it's two 'loops' it is still one body of water. what youre setting up is a complicated version of a loop with two pumps, one radiator, and a crapload of water. which wouldnt be so bad, but attempting to separate the ever-warming cpu water and the near-ambient radiator water, only relying on the reservoir to mix these, is making it worse.

[hellcamino]

Pumps are mounted on the res with Typhoon III.

I guess this would work, it seems a less than optimal solution though as far as flow rates are concerned. You will have a really high flow rate through radiator and a really low flowrate through all the blocks which is exactly opposite of what you want. To an extent higher flow through blocks = lower temps, and lower flow through rad = lower temps.

I forgot the typhoon III is a pump top, my bad.


I would just slpit the loops and use two radiators myself.

[troelsm]

I didn't know that lower flow rates through the radiator will yield better temperatures. That just seems plain wrong to me. Can anyone explain why this is? I would think that the more water goes through the radiator the more cooling I get.

If i had it all in one loop i would have really low flow rate through the entire loop along with the radiator.

AndrewZorn: This way I wouldn't restrict the flowrate through the radiator, that's the whole point. This way with two loops more water would flow through the radiator, since the flow won't be restricted by the cooling blocks that are way more restrictive than the radiator.

Bluehaze: I'm only going to use one D5 pump. And i wouldn't have a "really low flowrate through all the blocks". I would have the exact same flowrate as i would if the radiator was in the same loop and i only ran one loop total.

Guess the two main questions are:
Will the water be mixed well enough in the reservoir?
Do radiators really benefit from lower flowrates compared to higher?

[troelsm]

I forgot the typhoon III is a pump top, my bad.


I would just slpit the loops and use two radiators myself.

Well i appreciate your input but that's not really what i asked about

Edit: My two last replies sounded a bit pissy. It wasn't meant like that but english isn't my first language and sometime i come across a bit angry.

[bluehaze]

I didn't know that lower flow rates through the radiator will yield better temperatures. That just seems plain wrong to me. Can anyone explain why this is? I would think that the more water goes through the radiator the more cooling I get.

If i had it all in one loop i would have really low flow rate through the entire loop along with the radiator.

Bluehaze: I'm only going to use one D5 pump. And i wouldn't have a "really low flowrate through all the blocks". I would have the exact same flowrate as i would if the radiator was in the same loop and i only ran one loop total.

The water gets cooled while moving through the radiator, you need to find an optimal speed in which the water has time to cool yet doesn't sit in the rad after it's already cooled as much as it is realistically going to cool while moving. If the water just blasts through the rad the air doesnt have as much time to absorb the heat, if the water sits in the rad too long the air will absorb all the heat but the rest of the water will get hotter meaning the rad now has to remove more heat.

My point was with two pumps and two loops you could individually tune the speed of water through rad and speed through blocks. I think this would only work though with a bigger res and a fair distance from the inlet of res and inlet of pumps in order to have any chance at being effective as otherwise you will just end up recycling the same water through rad and the same water through blocks...I don't think the water temps would equalize.

And it's all a moot point if your only going to use one pump as it's impossible to run seperate loop through rad and seperate loop through the blocks with 1 pump.

[hellcamino]

Well i appreciate your input but that's not really what i asked about

Edit: My two last replies sounded a bit pissy. It wasn't meant like that but english isn't my first language and sometime i come across a bit angry.

Thanks for the explanation or I wouldn't have responded again :P, a D5 running through only 3 blocks and a radiator will work just fine if the blocks are appropriately chosen for low restriction. Adding a second dummy loop running from the same pump won't help you out any and will actually hurt your temps. Water will take the path of least resistance so it would flow 90% through your rad and 10% through your blocks (a guess).

[millertime359]

Guess the two main questions are:
Will the water be mixed well enough in the reservoir?
Do radiators really benefit from lower flowrates compared to higher?

You do have to remember the heat dissipates through the rad. You do need to allow the water some time in the rad for the heat to dissipate. Also though, if the flow rate is increased, the water would move faster, but would still spend same amount of total amount of time in the rad.

My biggest concern would be getting enough cooling with the fact that your loop is basically using the res as it rad.

If you are worried about flow, I would just get 2 220 rads and run 1 run rad CPU and the other MB GFX rad. I think that is the way the typhoon was designed to work, but I could be wrong. I think BoxGods or DB are the best people to put this question too.

[troelsm]

The water gets cooled while moving through the radiator, you need to find an optimal speed in which the water has time to cool yet doesn't sit in the rad after it's already cooled as much as it is realistically going to cool while moving. If the water just blasts through the rad the air doesnt have as much time to absorb the heat, if the water sits in the rad too long the air will absorb all the heat but the rest of the water will get hotter meaning the rad now has to remove more heat.

My point was with two pumps and two loops you could individually tune the speed of water through rad and speed through blocks. I think this would only work though with a bigger res and a fair distance from the inlet of res and inlet of pumps in order to have any chance at being effective as otherwise you will just end up recycling the same water through rad and the same water through blocks...I don't think the water temps would equalize.

And it's all a moot point if your only going to use one pump as it's impossible to run seperate loop through rad and seperate loop through the blocks with 1 pump.

I would really like to see some sort of scientific data that shows the performance of a rad actually decreasing at some point where the flow gets too high. That really sounds bizare to me.

The thing about equalizing the temperature in the reservoir is what i'm most concerned with. Hopefully the guy (don't remember his nickname) who designed the typhoon III can comment on that.

How come i can't run two loops with only one pump out of the typhoon? Isn't both inlets and both outlets connected to the same "tank"?

[hellcamino]

How come i can't run two loops with only one pump out of the typhoon? Isn't both inlets and both outlets connected to the same "tank"?

The loops must match in terms of restriction or one will get much more flow than the other.

[troelsm]

Thanks for the explanation or I wouldn't have responded again :P, a D5 running through only 3 blocks and a radiator will work just fine if the blocks are appropriately chosen for low restriction. Adding a second dummy loop running from the same pump won't help you out any and will actually hurt your temps. Water will take the path of least resistance so it would flow 90% through your rad and 10% through your blocks (a guess).

That's a good point. I didn't really think about that but it's actually quite logical. Guess that pretty much settles it

A little side question: Say i were to run one loop with, CPU 350, Gigabyte ex-58 NB block, a full gfx block (probably the koolance gtx 295 block) and the swiftech mcr320. Would that be allright to run in one loop or is that too restrictive?

[bluehaze]

I would really like to see some sort of scientific data that shows the performance of a rad actually decreasing at some point where the flow gets too high. That really sounds bizare to me.

The thing about equalizing the temperature in the reservoir is what i'm most concerned with. Hopefully the guy (don't remember his nickname) who designed the typhoon III can comment on that.

How come i can't run two loops with only one pump out of the typhoon? Isn't both inlets and both outlets connected to the same "tank"?

It's physically possible to hook it up like that but it won't work.

As far as scientific data your on your own but just given the general nature of how a radiator works water passes through it and as it passes through it air blowing through the radiator absorbs the heat. So the faster the water moves through the radiator the less time their is for the air to dissapate that heat.

When everything is on one loop you have to find a happy medium within which the water moving through the blocks has doesn't have enough time to overheat and the water moving through the radiator has enough time to dissipate the heat.

With seperate loops you dont have this restriction and could individually tune the speed of water through the rad and the speed of water through the blocks.

Only reason I think this might work is because you would effectively have the loop being fed with water cooled via a different loop that has lower temperatures. Would be interested to see the results if anyone has ever tried this...probably doesn't work though or else the method would be all over the place by now

[troelsm]

It's physically possible to hook it up like that but it won't work.

As far as scientific data your on your own but just given the general nature of how a radiator works water passes through it and as it passes through it air blowing through the radiator absorbs the heat. So the faster the water moves through the radiator the less time their is for the air to dissapate that heat.

When everything is on one loop you have to find a happy medium within which the water moving through the blocks has doesn't have enough time to overheat and the water moving through the radiator has enough time to dissipate the heat.

With seperate loops you dont have this restriction and could individually tune the speed of water through the rad and the speed of water through the blocks.

Only reason I think this might work is because you would effectively have the loop being fed with water cooled via a different loop that has lower temperatures. Would be interested to see the results if anyone has ever tried this...probably doesn't work though or else the method would be all over the place by now

I see your point that the water needing to have some time to dissipate the heat, but isn't it kinda like saying that if you put your head out the window of a moving car, at some speed it will start to get less cold because the air hitting you dont have enough time to draw the heat off your face?

You could probably say that if you just look at a specific amount of water going through the radiator the water going very fast wouldn't be cooled very much, but that wouldn't matter since there are so much more going through.

Well the point about the water going through the least restrictive loop is valid so i guess it doesn't really matter

[DarthBeavis]

All I will say is with the T3 you want TWO LOOPS . . . .period. Anyone here who says otherwise . . .even someone like our senior people . . .are wrong if they say one loop with a T3 is better than two. What is on those loops is another matter.

[millertime359]

I don't think you ever got over to that other theard Troelsm, but here you go from:

Will all this talk over running seprate loops off of this, could you have a loop that is rad only and a loop with your CPU and GPU blocks only. I wouldn't think it would work, but the T3 isn't your standard res will it allow you to do that?

I would think that it WOULD work actually. Keep in mind that the water in the T3 is exchanged almost 16 times a minute. A larger rad would do better as its more restricted likely.

It is cool that people are starting to see the potential here. Even I have no idea what the limits are on the T3 because there are all these new variables. Build people cool tools and they will in turn build things you can't even imagine.

[nikhsub1]

...If the water just blasts through the rad the air doesnt have as much time to absorb the heat, if the water sits in the rad too long the air will absorb all the heat but the rest of the water will get hotter meaning the rad now has to remove more heat.

This could not be more incorrect. Please do some reading before blasting 'fantasy physics' on a public forum.

As far as scientific data your on your own but just given the general nature of how a radiator works water passes through it and as it passes through it air blowing through the radiator absorbs the heat. So the faster the water moves through the radiator the less time their is for the air to dissapate that heat.

When everything is on one loop you have to find a happy medium within which the water moving through the blocks has doesn't have enough time to overheat and the water moving through the radiator has enough time to dissipate the heat.

So then you're just making all this up? This has been discussed to death for many years. I wish I knew how this false myth keeps getting perpetuated.

[Waterlogged]

I wish I knew how this false myth keeps getting perpetuated.

This is the real $250,000,000,000 question.

[bluehaze]

This could not be more incorrect. Please do some reading before blasting 'fantasy physics' on a public forum.

So then you're just making all this up? This has been discussed to death for many years. I wish I knew how this false myth keeps getting perpetuated.

Where were you during all the discussion last night? So how does it work then? Because i'm still confused about why it wouldn't work if you have two seperate loops and are able to adjust flow through the rad seperately from flow through the blocks. And I have looked and haven't seen this discussed anywhere. Would love to read about it though if you happen to know where I might find some info

Reason I am confused is if you raise your fan speed thus increasing air flow through the radiator your temps go down, so wouldnt lowering water flow through the rad have the same effect as increasing air speed as long as the water was flowing through the rad independently of the blocks? But I guess if this is the case then you now have increased volume of water flowing through the blocks as compared to volume of water flowing through the rad regardless if they are seperate loops because they are still sharing the same water source...

I don't know man...it's all pretty confusing.

Cheers

[Darth_Penguin]

Where were you during all the discussion last night? So how does it work then? Because i'm still confused about why it wouldn't work if you have two seperate loops and are able to adjust flow through the rad seperately from flow through the blocks. And I have looked and haven't seen this discussed anywhere. Would love to read about it though if you happen to know where I might find some info

Reason I am confused is if you raise your fan speed thus increasing air flow through the radiator your temps go down, so wouldnt lowering water flow through the rad have the same effect as increasing air speed as long as the water was flowing through the rad independently of the blocks? But I guess if this is the case then you now have increased volume of water flowing through the blocks as compared to volume of water flowing through the rad regardless if they are seperate loops because they are still sharing the same water source...

I don't know man...it's all pretty confusing.

Cheers

Im pretty sure nikhsub1's comment to you was regarding your theory about flow and not the loops of the T3. I posted the equation last night for you in the other thread but you didn't comment on it. Your logic "sounds" plausible in the sense that slowing the water through the rad should pull more heat but this isn't actually the case. It has been my whole point to you that slowing things down does NOT make things cooler. Look at the equation and you will see that amount of heat transferred is directly proportional to flow.

[nikhsub1]

Reason I am confused is if you raise your fan speed thus increasing air flow through the radiator your temps go down, so wouldnt lowering water flow through the rad have the same effect as increasing air speed as long as the water was flowing through the rad independently of the blocks?

So you acknowledge that more airflow = better temps but slower water flow should equal better temps? Does this theory not sound contradictory to you? You say that the more time the water spends in the rad will cool better correct? OK, lets look at it like this then. You have 2 identical loops except one is running 1GPM and the other is running at 2GPM. At 1GPM the water spends twice as long in the rad as compared to 2GPM, however, at 2GPM the water passes through the rad twice as often. So now answer me which loop spends more time in the rad?

Im pretty sure nikhsub1's comment to you was regarding your theory about flow and not the loops of the T3.

[kinghong1970]

dunno nikhsub1, maybe if the pump was so crappy that it comes on and off, then water would sit in a loop and allow for better heat transfer?



excuse my sillyness...

maybe bluehaze is looking at stationary water boiling in a pot...
bluehaze, the cooling system in your loop does not process in batches, it constantly flows and while if you look at how little amount of heat is absorbed in a block in a given time, you have to take into factor the flow rate...

dunno if this makes sense...

if the loop were to remove temp in batches, then i think the temp of your cpu will spike up and down... also, bigger volume capacity on the blocks would be needed.. no?

[bluehaze]

So you acknowledge that more airflow = better temps but slower water flow should equal better temps? Does this theory not sound contradictory to you? You say that the more time the water spends in the rad will cool better correct? OK, lets look at it like this then. You have 2 identical loops except one is running 1GPM and the other is running at 2GPM. At 1GPM the water spends twice as long in the rad as compared to 2GPM, however, at 2GPM the water passes through the rad twice as often. So now answer me which loop spends more time in the rad?

Well of course the 2gpm what confuses me is not that, what confuses me is that lowering the flow of the water by proportion increases the volume of air and vice versa. So if you are only capable of maintaining a certain volume of air wouldn't lowering the volume of water result in decreased temps?

[AndrewZorn]

not to be confused with that fact that many people do turn their pumps down from the max to get the best results, but this comes from the balancing of heat dump from the pump vs flow