• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

a little flowrate question...

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

Suma

Member
Joined
Feb 3, 2003
Location
Israel & USA
Hey,
I'd like to have an advice about choosing the right flow rate for my new homemade system.

I'll be using a system with 1/2" tubing, execpt the radiatior which is 3/8" :/
I thought about fitting the outlet of the radiator to 1/2" while using the outlet of the pump and the inlet of the radiator in 3/8".

would you think that would be good? or maybe the other way around?


Thanks,
Suma.
 
Suma said:
Hey,
I'd like to have an advice about choosing the right flow rate for my new homemade system.

I'll be using a system with 1/2" tubing, execpt the radiatior which is 3/8" :/
I thought about fitting the outlet of the radiator to 1/2" while using the outlet of the pump and the inlet of the radiator in 3/8".

would you think that would be good? or maybe the other way around?


Thanks,
Suma.

Combining 1/2 and 3/8 in your system is probably not going to be much of an improvement over using only 3/8, if you are able to change the radiator to 1/2 I would just do 1/2 all around.
 
I got a different approach... I noticed the restriction on High Pressure pumps from 8mm and up tube was minimum and or non-existant.

In fact, the only thing effecting my flow rate is my CPU block. That is the bottle neck, and it is so great, that it just doesnt make a difference if I go with 1 foot tube or 8mm. Ive tested it..

Anyways, Low Flow is not bad... And 1/2" tube is a PITA to deal with. The 5/8 OD kinks, and the 3/4 OD is crazy thick. and hard to twist.

However, 1/2" Tube looks AWSOME!

Mike
 
you should probably go with 7/16" tubing, it beats 3/8 on flow rate and it kills 1/2 on flexibility.

also ever here the saying, the chain is only as strong as its weakest link, meaning dont use different sized barbs, try and get it all the same.

:)
 
Phrenetical said:
you should probably go with 7/16" tubing, it beats 3/8 on flow rate and it kills 1/2 on flexibility.

also ever here the saying, the chain is only as strong as its weakest link, meaning dont use different sized barbs, try and get it all the same.

:)


couldn't of said it better my self.
 
Correct

GreenJelly said:
I got a different approach... I noticed the restriction on High Pressure pumps from 8mm and up tube was minimum and or non-existant.

In fact, the only thing effecting my flow rate is my CPU block. That is the bottle neck, and it is so great, that it just doesnt make a difference if I go with 1 foot tube or 8mm. Ive tested it..

Anyways, Low Flow is not bad... And 1/2" tube is a PITA to deal with. The 5/8 OD kinks, and the 3/4 OD is crazy thick. and hard to twist.

However, 1/2" Tube looks AWSOME!

Mike
You are absolutely correct. It isn't noticed much but the ports in almost every cpu block are 1/4 inch anyway. So we screw in various size fittings to those ports and it is reduced to 1/4 thru the cpu and then opens back up to whatever size tubing we are using.

For example, the fittings for my block were Perfect seal G 1/4 fittings - 3/8. I chose to expand all the fittings to 3/8 for looks only. It wouldn't have made a bit of difference in the flow or performance because where you screw in the fitting is only 1/4.

In my small amount of experience, I would say that the thing that will make more difference is not using barb fittings. They are not needed and they reduce each fitting size due to the barbs shape.

I did an enormous amount of research when I set up my system and I learned a lot from the germans. They use all 1/4 tubing for the reasons I stated above. They are convinced the only benefit from 3/8 and 1/2 inch tubing is looks.

IMO of course.
 
jph1589 said:
You are absolutely correct. It isn't noticed much but the ports in almost every cpu block are 1/4 inch anyway. So we screw in various size fittings to those ports and it is reduced to 1/4 thru the cpu and then opens back up to whatever size tubing we are using.

For example, the fittings for my block were Perfect seal G 1/4 fittings - 3/8. I chose to expand all the fittings to 3/8 for looks only. It wouldn't have made a bit of difference in the flow or performance because where you screw in the fitting is only 1/4.

In my small amount of experience, I would say that the thing that will make more difference is not using barb fittings. They are not needed and they reduce each fitting size due to the barbs shape.

I did an enormous amount of research when I set up my system and I learned a lot from the germans. They use all 1/4 tubing for the reasons I stated above. They are convinced the only benefit from 3/8 and 1/2 inch tubing is looks.

IMO of course.


G1/4" is equal to 1/2" diameter. Don't believe me? Remove those barbs and measure that hole ;)

The measure of the barb hole (in this case G1/4") is always double the size AFAIK. I know that's the case for G1/4"
 
I think he meant the actual opening for water to flow through. On an unmodified perfect seal it's about 1/4". It would probably be beneficial to drill a larger hole.
 
I just measured the inside of the barbs that came with the MCW30 that I have sitting here (same as the barbs for the Storm, Apogee, MCW60, and many others I'm sure) and they are all 3/8" ID at the base of the screw. The opening at the top of the 1/2" barb is just a hair (my term for 1/32" ;)) more than that (10mm?), while the 3/8" barb is ~5/16" at the top (8mm?). So 3/8" barbs are slightly more restrictive than necessary - 1/2" is the only one that keeps a 3/8" ID or better all the way through the barb. IMO - Unless you have some long runs in your loop as I do, 7/16" tubing is ideal as has been the trend recently. The only thing holding back 7/16" from becoming universal in North America Computerland is it's availability - not every local hardware store carries it as they do 1/2".

Of course many water blocks have their own restriction. A quick check of the cross-sectional area of the Storm's 35 micro-jets at 0.029" each (according to Robotech over at system cooling) yields a combined ID equivalent slightly less than 3/32". Using some of the logic presented here you may as well run 1/8" tubing everywhere if you have a Storm in the loop ...
 
Last edited:
yes

Moto7451 said:
I think he meant the actual opening for water to flow through. On an unmodified perfect seal it's about 1/4". It would probably be beneficial to drill a larger hole.
That is what I was trying to descibe, although I didn't seem to do it very clearly.
 
On an unmodified perfect seal it's about 1/4". It would probably be beneficial to drill a larger hole.

Fitting ID's according to the various measuring tools (digital and analogue calipers) in front of me of various G1/4" threaded parts... rounded to nearest 10th of a mm.

DangerDen's 1/2" Perfect Seal fitting is 10mm = 3/8", not 1/4" and remains one of the larger ID'd fittings available.
DangerDen's 3/8" Perfect Seal fitting is 9mm = just UNDER 3/8"
DangerDen's 1/2" HiFlow fitting is 10mm = 3/8"
DangerDen's 3/8" HiFlow fitting is 7.5mm = just over 2/8"
DangerDen's old white poly 1/2" fitting is 9.5mm
Swiftech's black 1/2" Poly fitting is 9.5mm
Swiftech's black 3/8" Poly fitting is 6.7mm
EK's 3/8" fitting is 7mm
EK's 1/2" fitting is [OUT OF STOCK SO CAN'T MEASURE]

For reference...
ThermoChill G3/8" threaded 1/2" fitting = 10.5mm ID
ThermoChill G3/8" threaded 3/8" fitting = 7.2mm ID

ID of a G1/4" threaded hole excluding threads = 11.4mm
TOTAL G1/4" threaded hole including threads = 12.3mm (1/2" being 12.5mm)

ID of a G3/8" threaded hole excluding threads = 14.9mm
TOTAL G3/8" threaded hole including threads = 16mm
 
Last edited:
I have spent hours in the frozen CPU warehouse looking at just about everything they sell (which is allot). And I can honestly say that no matter what you do, you will be restricted to 1/4" ID on at least one part of your system.

The Storm, is considered one of the best water blocks on the market. It comes with 1/2" and 3/8" barbs. When I attach the 1/2" tube to the VGA block (Tyee) which is relatively unrestrictive, I get a waterfall effect on the output. This is NOT the case with the 3/8" barbs. I assume this is due to the lack of flow coming out of the water block.

Also if you really want a nice setup, they sell 3/8" Compression Fittings, which they DONT sell for 1/2". Compression Fittings are the best, though changing fittings on the parts often requires some pipe tape to seal.

Germans are no idiots. They use 8mm almost exclusively. (Which is slightly larger then 1/4"). They know that water pressure and not water flow maters. It takes ALLOT of heat to warm water up just a bit, and flow only decreases the single cycle hot spot time. Since the single cycle time on the heat source (even at the lowest flow rates) is so slight, it doesnt change the water temp much. Thus the water will hit the radiator without any significant increase in heat (again even at the lowest flow rates).

Also remember that water takes 1 calorie of energy to heat one cc of water. When we talk about food we use Kilocalorie's which the food industry was so wise (and deceitfull) to define a single calorie as a "calorie" and a kilocalorie as a "Calorie" (they now even break this caps rule). Anyways, calories are a way to determine the amount of energy it takes to heat up different components. Luckily to us water coolers, and the entire field of science, water itself has the highest heat absorption rate then any other materials before it changes temps (though their may be other man made materials I havent read about though have heard about). This means that the water can sit on the hot spot and cold spot a long time without changing a degree.

I keep mentioning Single Cycle because it is important when the water temp rises and then cools dramatically on every cycle that it starts to effect performance. This is unefficent because it means the water is staying on the hot spot for WAY too long, thus changing the delta of water from cold to heat source temp. When this change in hot to cold decreases, the efficency of the water to cool decreases. This would require higher flow rates to fix. What makes things worst is that a 10 l/min pump may have enough flow, but because a low pressure, it only performs at 1 l/std. So flow rate and pressure are not independant

From my personal experiances, the level of flow that starts to effect a system is around 1 l/std(which is german speak for 1 l/min).

The way to test the flow rates efficency is to test the radiators input water temp and the radiators output water temp. If the temp is drasticly different (more then 1-2c) then you have to increase flow for better performance.

The interesting fact that I learned is that it isnt the size of the tube that effects this flow so dramatically, but the pressure the pump can exert. I came to this conclusion because of my fathers home improvement history. He showed me the water valves used for garden hoses. Inside the huge water pipe, there is a 1/4" inner pipe that feeds the water hose. Now water pressure from the water authority is much higher then what we want or need, it definately proves my point.

What makes everything even more complex is that many water blocks use techknowledgies that uses the presure to increase the water -> copper surface area. They use micro-channels or jets to do this. This mixes the water, and that allows the water to heat evenly. If you want your water on the stove to heat up faster, you can stir it. This mixes the water and thus increases the delta temp between the hot side and the cold side.

The Danger Den Tyee and its Ge-force equivelent is a high flow water cooler, that depends on pre-existing water turbulance, to ensure that the water is mixing within its wide and long passage. I find it works great, and is one of my most unrestrictive parts in my system.

Choosing a pump, means looking at the flow rate and the pressure. The pressure should be very high, while the flow rate should be 10 l/min at a minimum.

I hope this all makes sense and helps out things. I recommend you go to the home site and that you read the articles on flow rate, pressure and water cycles.
Mike
 
Last edited:
I like Jph's system too... Just dont like his pump... I love my pump, its silent, powerful, and works great in my system. I dont know why people by any other swiftech pump... Then again, I am totally ignorant to how almost any other swiftech pump works
 
Hard to go wrong with an MCP655 and a Storm. :D

But I must admit the new DDC+ (18w) is extremely quiet and with the right top just as good, though more expensive ...
 
Back