Pressure

[zip22]

so let me get this straight,

-you now have a 3/4" input to your pump? this probably helped the situation

-you added a feed back loop, so water can either go into your cooling loop, or directly back to the pump input? i understand how this would make life easier for your pump, but i fail to see how this would help out your cooling loop at all. of course it will make your pump quieter, since it doesnt have to do anything.

im curious as to your temps with the valve completely closed, my guess would be better than with it open

 

[voigts]

Thanks for the thorough explanation Crimedog. I still have a question though. When you mention flow, I am assuming that you mean the speed at which the water moves through the system. My question as it relates to this discussion is how does the volume of water relate to cooling performance. There is significant difference in the amount of water a 1/4" tubing loop and a 1/2" tubing loop is going to hold and pump. Where does the amount of water and size of tubing in the loop work out? If you theoretically just kept using larger and larger tubing, does that mean that the system would cool better? The larger the tubing, the lower the pressure would be, right? What factor does pressure play in the whole mess? Does the tubing size make that much of a difference?

 

[Jas]

As I undertsand the whole concept, With less restricive, larger diamter tubing, and components, less head pressure is needed to acchive higher flow rates. With lower diamter more restricive components more pressure is needed to acheive higher flow rates.

I beleive, less restriction, flow rate on a pump becomes more important, More restriction, the head pressure becomes more important.

As for the ammount of water in a system, in theory, I would think assuming you have the same components, and flow/pressure, only say one system has twice the volume of water, it would require more energy to heat ALL the coolant in the system with more water.

So, in practice, does that translate into lower temps? I honestly have no idea. However If we look at system that are "radical" in design, say using a 35 gallon trash bin a resevior, it woudl be difficult to not think that by simply having that much voluem of water, it would be abel to acheive good temps, potentially without a radiator.

 

[voigts]

What you are saying makes sense. There has to be some kind of balance between volume of water, flow, and pressure, with somewhere a combination that works best for a particular setup depending on the blocks used, etc. In reality, if you were using a 35 gallon water tank, would not the tank be acting kind of like a giant radiator? I mean, at some point, the tank would be transferring heat into the surrounding air, right? So is the question really that there is 35 gallons of water, or that by having such a large tank, it kind of replaces a rad?

 

[crimedog]

voigts basically there are very small temperature differences between tubing sizes, however this depends on your pump. if you have a weaker pump the higher restriction will have a bigger impact on flow rate (also, think of flow rate as volume/time not speed). however, the pump generally used for watercooling are designed for high pressure. these will give a much smaller flowrate loss with added restriction. that's why there will typically be a tiny temperature difference between 3/8" and 1/2" id tubing when you have a good pump.

now when you're using "low-flow" gear (blocks that don't gain much from higher flowrates) it doesn't really matter if you "upgrade" the tubing. you will decrease the restriction of the loop a bit, but the blocks are already huge flow killers. so even though your flow will increase, the c/w of the blocks will barely change and you won't notice a difference in temperatures. however, increasing the tubing size couldn't have a negative effect on temperatures.

so how much temperature difference are we talking about, say from 1/4" id to 5/8" id? with a koolance kit or similar it'll be small, with a medium pump and good block it'll be noticable, and with an extreme high performance pump it may not be noticable.

keep in mind that what i post generally isn't 100% technically correct, but i just want to convey the idea.

 

[voigts]

Thanks for the good info Crimedog. What you are saying makes sense and is the kind of answer that I have been looking for. I have Swiftech blocks which are meant for higher flow with a 50z pump, so changing the tubing from 3/8" to 1/2" might make some difference, but probably not more than a couple of degrees c at best. If I get you right, if the pump is good and has a good head/pressure rating, then it will get the water through the tubing whether you put 3/8" or 5/8" enough to give good cooling. How the system performs is more about the pump and the waterblocks and how restrictive they are vs. how well they transfer heat. Is that about the gist of the answer?

 

[OutS|der]

my understanding of it all is this
more flow = better cooling, with smaller tubing or mutiple blocks theres more restriction meaning less flow, to achive higher flow higher pressure is needed
say ya pump is rated to do 125 liters/h with 6 feet of pressure/head, thats what you'll get with 1/2" tubing but with 3/8" or mutiple blocks you'll need say 7 feet of head to reach that flow rate
these aint even close to excate but i hope ya get what i mean
more flow is better but may need more pressure or head to get it
least thats my understanding

 

[Albigger]

How the system performs is more about the pump and the waterblocks and how restrictive they are vs. how well they transfer heat. Is that about the gist of the answer?

Not totally, just that if you happen to have a block that has low heat transfer capability, increasing the flow rate (with a better pump or larger tubing) won't make as much of a difference as it will with a waterblock that is better at transferring heat but is currently limited by flow rate.

In other words, if a waterblock's C/W vs. flow rate curve is rather flat, changing the flow rate won't have as much of an effect on temperatures as it will with a waterblock with a C/W vs. flow rate curve that varies much more with flow rate.

 

[LabRat23]

Good work crimedog. I second everything you said.

Back to the origional post: Nice to see another DIY guy. What do your blocks look like and how do they work/perform? Can you figure out how much backpressure you have with your flowrate?

My advice is to put the full power of the pump into the loop. Doing anything else is just not getting the full value from your pump. Do not put your cpu block and your gpu block in parallel. If you have other blocks besides your gpu and cpu blocks, you may want to run your loop through the cpu block then split the loop up to feed your other blocks, but always put the entire flow through your cpu block IMHO.

 

[crimedog]

How the system performs is more about the pump and the waterblocks and how restrictive they are vs. how well they transfer heat. Is that about the gist of the answer?

Yup, it's all about how you match it up. It's why you see european style watercoolers with high restriction blocks, low flow pumps, and small tubing. It's just another style of watercooling (and it's very quiet and good looking ).

 

[voigts]

Thanks for all of the info. This has been very helpful. Now if I can keep myself from getting into waterblock designs.....

 

[robjen]

Well, I bagged that configuration and built 2 new blocks that were 3/4 thick instead of 1/2 to allow more flow through the blockes. Seemed to work perfectly except for one minor detail. I CRACKED THE DIE ON MY RADEON 9800 PRO!!!!

It boots, but the screen is all artifacted.

 

[crimedog]

omg how did you crack it?

 

[robjen]

Instead of just mounting it with a lot of artic silver, I got smart :lol: and decieded to use a shim between the sink and die.

With only 2 mounting holes, Im guessing when I put it in my case, the pressure from the hoses caused the sink to wable a bit and crack the corner off.

Man I am ****ed.

I just ordered another 9800 pro, and this time i am going to skip the shim and used that cheap thermal tape. and let the shim take the brunt of the pressure. I know that they don't perform as well but, it sure beats shelling out the money for a new card.

So right now I am running on my old standby nvidia card with 8meg ram. Damn, that card has bailed me out more than once.

Right now I am rethinking my stratigies for cooling. One of my systems is using a cheap *** 1/4 tubing system that I picked up on ebay. It runs very quiet, and is super easy to route the tubing. This half inch stuff is impossible. Right now I am getting 3 degrees over ambient with the half inch tubing, but even if I could get the same performance as a highend fan sink without the noise I would be happy.

If anyone knows anything about small pump small tubing setups, let me know.

Oh, I build my own blocks, so If you know of any designs that work with this type of setup, I would also appreciate your input.

enough of my ranting, I can't talk to my wife about this (she'd kill me if she knew I broke the card)

 

[voigts]

Cracking a VC really sucks. I shelled out the bucks and went the boring route of using Swiftech blocks. But hey, they work.

 

[crimedog]

search mcmaster.com for part number 5233K44

it's 7/16" id 5/8" od tubing. cathar officially discovered/recommended it and it rocks! much easier than 1/2" and apparently you don't lose any performance. sits very tight on 1/2" barbs, and routes incredibly easy. cheap too

 

[Jas]

Getting back the the "Matching of water volume and components and flow/rates is exactly the reason you see different sizes of tubing, different types of water blocks, systems with Reseviors, and systems without.

As far as a 35 gallon trash bin, even if it was in a plastic enclosure, the sheer volume of water would act as a cooler in itself. Fluid resistance will eventual cause the heated water molecules to slow down and loose heat. If you were to draw the water from the very bottom of the bin, I can pretty much garuantee you wouldn't need a radiator.

Is that practical? Heck no. Would it work, yes. But when it comes to putting a water cooling setup inside a computer space becomes an issue. In europe, smaller tubing is the norm, but at high pressures. Takes up less space. Throw a Res into the mix to increase water volume and a good rad, and you can have a great performing loop.

Its all about matachin up the right componetns for your setup. If you have a restricive water block (basically any jet impigent design) then 3/8 or 1/2 ID probably won't make a lick of difference, since you will need a pump with some good pressure to it.

 

[voigts]

And your discussion brings up the res vs t-line debate. Some say that a res hinders performance. Others say it helps performance. Others say the t-line helps performance. and still others say that it is mostly a matter of preference and space. And of course Jas, if the European systems function on more pressure, wouldn't putting in a res decrease pressure? All of this can be rather confusing to say the least, and knowing what really to get to achieve the best performance for the best bang for the buck seems more of an art form than just science. Slowly but surely I am getting the sense of this all. I guess it just takes time, reading, and messing with different setups to figure out. Thanks for the input.

 

[Jas]

Well, assuming you put the Res in the right spot in your loop, it will not decrease performance.

You will see mentioned many times that the order of components in your loop do not matter, as far as temps/performance is concerned. However, If you were to have a loop that had the Res, anywhere, except for directly before the pump, then you are not acheiving maximum flow/pressure on the components that would come after the res. Assuming you are using the Res properly. Now, assuming you have the lid/fillport on tight, and the loop is completely closed, the argument that since no medium (gas or liquid) can enter or exit the loop, pressure and flow should be constant is not entirely true. The larger mass of liquid and gas in the Res has an elasticity. This is easily seen in clear flat res (Bay Res is the easiest to view this), you will see a wave like motion, this is the liquid rebounding against itself and the container. This results in a pulse like motion, which if you listen to some pumps in a system with a Res you can hear a slite change in the pitch of the motors hum. A sounds of Vrum Vrum Vrum Vrum

This is because the pressure against the inlet is changing slightly on a pattern over time. Essentially, the res is producing a wave like fluctuation in the pressure. As the liquid is pumped into the res, liquid needs to be expelled to maintain equilibrium. But since there is a buffer of gas, the liquid has room to expand, thus allowing for more liquid to be pumped in, than is drawn out at that exact point in time. To acheive equilibrium, the liquid must then be drawn out faster than it is pumped in, hence the wave like motion.

Its like a suction affect, rather than a steady flow.

Mounting the Res directly before the pump, helps ease this a bit. As the vacum affect of the inlet to the pump has a direct affect on the coolant in the res, drawing it out as liquid is pumped in, rather than a more secondary affect should there be a component to the loop between the pump and res.

A completely closed loop with a T-Line would maintain the most even distrubution of flow and pressure. Since there is no large liquid repository or gas buffer, the elasticity of the liquid is minimized.

As a test. Try placing a decent sized Res in a system somewhere besides directly in front of the pump, listen to the Pump, you should hear a slight ocilation of the pitch to the sound of the pump. In a T-Line you should not hear an occilation. You will still have SOME wave motion to the flow even with the res directlyin front of the pump, but properly placed, it would be extremely slight and completely neglidgable.

 

[voigts]

I had a res, and took it out and put in a t-line mainly for the sake of room, but also the res makde a lot of noise. There was a very noticeable swooshing sound as water went through the res. Needless to say the t-line got rid of that. So as long as the system has a good pump (like MCP650) and is rather non-restrictive, the t-line should actually provide a more even flow and pressure, correct? And as long as the t-line is at or near the top of the loop like I have it, bleeding it is not hard at all, which is the major reason that I hear for using a res, right?