Point of using high-flowrate pumps?

[DeNs]

Well, considering the volume of your desired coolant will probably be no more than a couple of litres, I'm finding it hard to see the point of using 1200lph pumps within a WC system. Of course, there is a certain amount of restriction of flow within the system at certain points (WB, etc.), but I don't see how it justifies this.

If for instance you're using a closed-loop system and you're using a Eheim 1250 with 2/3 blocks, the water will not be able to cool down while going through the heatercore/rad since the waterflow rate is so high. Without restriction, 20litres of water can be pumped around the system in one minute - taking into account

If someone could prove me wrong i'd be really pleased becuase I don't see the point in spending $150+ on a pump that could do the same job as a $60 one. I'm just being cheap and trying to save money

True, reliability comes at a cost, but I don't believe they're going to design an aquarium pump that only lasts 6 months

---dens

[FaTs]

You want a pump with god head rather than flow rate. Iwaki's are about the best pumps for watercooling atm. They are alittle pricey

What do you have now ?

[Etacovda]

ooo here we go. Wait for a lot of posts here.

www.procooling.com

Have a look at the waterblock tests by Phaetus in his work log, or on the main site.

How 'long' the water stays in the heatercore/rad has little to do with it. I'll use cathars example here.

Say you've got a race track, with cars going around it. Take a section of that track for flowrate. If the cars are going 60km around it, they'll stay in that section for X amount of time per hour. If they're going around at 100km, the percentage of time (X)they spend in there is the same per hour, they're just going faster... I think thats followable, its 1.30am here essentially, it does nothing 'bad' by going through the radiator faster.

faster is better, because more water contacts the waterblock. This has diminishing returns, of course, its not linear.

Also, with a 'fast' flow pump, they generally have higher pressure, which many new waterblocks require as they are quite restrictive. A good flow rate to aim for is 1.5+ gpm through your entire loop.

Because your coolant is only a couple of litres max, the faster you have it going around and removing heat, the better.

Edit:
Once again, i apologise for my terrible spelling, im tired.

[DeNs]

Currently I don't have anything - which is why I asked actually I actually dont know a lot of terminology - one thing I don't know is what "head" is - care to explain?

Yeh, I guess I might not be able to skimp on the pump after all...

---dens

[Etacovda]

Its a rating of pressure, I think its related to torque. If you look at any pump graph, it'll be rated by head and flow rate

[jamesbond]

If the coolant is travelling to fast, it wont have enough time to absorb the heat in my opinion. That is why you need a thermosat on a car, it slows down the coolant flow.

[cyclonite]

Each individual water molecule might not absorb as much heat with higher flow rates, but more water molecules will be absorbing the heat.

It's like saying 1 + 1 + 1 + 1 = 4 is the same as 2 + 2 = 4.
Simple analogy which more than likely has flaws, but that's how I think of it.

[Ibkikn]

The heat transfer coefficient is directly related to the fluid flow, when it comes to water cooling through metal blocks.

Hence the higher the flowrate through the system, the better the potential of the system.

It usually the heat removal side of things that let most systems down.

[Long Haired Git]

Lots of posts on this subject, including a recent one.
JamesBond's thought re the thermostat has been debunked in that recent thread - the thermostat is there to allow the engine to get up to operating temperature quickly by reducing cooling performane by slowing coolant flow. Once at operating temp, the thermostat then opens up, allowing full flow.

Head is a measure of a pumps ability to beat resistance to flow. The measure is how high a pump can pump water vertically and still get any flow. The LPH measurement of pumps is with zero vertical height and zero restriction on the outlet. So, a watercooling rig will flow something in between zero and the max LPH, depending on where its resistance is between no resistance and the equivalent work of pumping the fluid up the head height quoted for the pump.

Example: A BA heater core, LRR waterblock and small quantity of 1/2" tubing is known to exhibit 1.1m of head at 5 LPM.

Head increases as flow rate increases by approx the square of the flow rate. So that same rig should show about 4.4m of head when flowing 10LPM.

Oh sod this, find that other thread, its got pictures and everything.

[littlee]

If you want to know how a car's coolant system works read this:

http://www.grapeaperacing.com/GrapeA...ingsystems.pdf

[^catalyst]

my thoughts on the subject is that you really want your water to be the same temperature all the way around, i.e no 'hot' or 'cold' spots, this is best achieved by having water get around your rig pretty quickly. Sounds simplistic but i've had a long day at work

[Cathar]

Originally posted by jamesbond
If the coolant is travelling to fast, it wont have enough time to absorb the heat in my opinion.

I am honestly baffled as to how people think like this, especially when all around them the evidence points to the contrary.

A really simple example is cooling down a hot saucepan. If you just sit the saucepan on the stove, it will take a long time to cool it down. Pick the saucepan up and swing it around in the air so that the air moves quickly over it, and it cools down a lot faster.

Or better still. Walk outside on a cold and windy morning wearing a T-shirt and shorts. Stand in an area where there's no wind, and it doesn't feel so cold. Stand out in the wind and it feels very cold. The faster the wind, the colder it feels because the heat is being sucked away from you more quickly.

So why then do people think that with water it is suddenly the opposite of what they experience every day?

[jlrii]

More flow = better performance to a point as you will reach a point of diminishing returns both in gains in block rad performance and heat added from the pump. Blocks respond quite a bit to higher flows, radiators less so. If you look at the graph 11 here which uses flow and gives the "c/w" the radiator will perform at. "C/W" is how many degrees c the water leaving the radiator will be above room temperature per watt of heat added to the water . So if you have a rad with a c/w of .03 and a processor that puts 100 watts of heat out, the water temp coming out will be 3 deg c above room temp. A little change in c/w makes a big diff as it will be multiplied by 80 to 110w.

Curves are given for 4 differant fans. Note that water flow does improve performance but not that much. Air flow is much more critical on rads

Now look at the 3rd graph down here . Note that c/w here refers to the temperature the processor will be over the water temp entering the block. With a flow rate of 1 lpm the c/w is about .22 while at 10 lpm it is about .17. That means at 1 lpm a 100w cpu will be 22c above the entering water temp, and ath 10 lpm it will be 17 deg c above entering water temp.

Thats the basics, final performance can be guestimated by simply adding the c/w of the rad and block so with the components shown above at 1 lpm the rad c/w= about.032, and the block is about .221, total cw is .253 x 100w =25.3 deg c above room temp or a cpu temp of about 47.3c in a 22c room. At 10 lpm the rad wil give you about .025 c/w and the block about .172 c/w for a total of.197 which would give you a cpu temp of about 41.7c in a 22c room. A total differance of almost 6c from 1 lpm to 10 lpm.

Note that these are guestimates only ther is quite a bit more involved and unless you are running prime95 or something equally intense the cpu will not be putting out its max heat. With most aplications water cooling has a nice advantage, due to the mass of all the water in the system it takes quite a while for water temps to rise when the cpu loads up intermittantly. Also know this: take any component testing with a grain of salt, blocks are notoriously difficult to test. Also note that you actual cpu temp will probably not be accurate as reported by the motherboard or most other methods. Even different BIOSs on the same MB may report different temps for the same setup.

Hope that this helps

[DeNs]

Wow, thanks for everyone who posted- you've helped out immensely

There is one last question that I have (Although it is one that is probably asked the most). For example, if I have a CPU WB, GPU WB, Heatercore and Res, what (roughly) pump would be best suited for the job? I was thinking of a Eheim 1250, but they're pretty expensive.

Can you purchase Iwaki pumps locally here (Melb)? One other thing, I haven't gone to any aquarium stores yet, but are pumps generally cheaper from there instead of specialised WC Stores (i.e. Below-0)

Thanks
-DeNs

[jlrii]

Iwaki= really nice pump=$$$.$$ I wish I could justify the money to get one, Personally I'd choose a Swiftech MCP 600, or Laing D4. Nice performance at higher head, ok price.