performance related to down sizing tubing, and flow related to smallest ID in loop

[DigitalChaos]

i was having a bit of a heated debat with a guy in the chat... i would like to see other peoples input on these 2 topics.

1: This is the question i am mainly interested in... When using 1/2" barbs, their ID is ~3/8". Silicone tubing with a 3/8" ID stretches over these barbs nicely. It would be easier to manage smaller tubing, and I do not see much potential for 1/2" ID tubing because your barbs are limiting the flow. In the majority of systems running 1/2" barbs (with ~ 3/8" ID), wouldnt 3/8" ID tubing be more ideal? would 1/2" tubing make any real difference here?

2: (this is basically the same question.. just a slight difference) If you have all 5/8" barbs (~1/2" ID), EXCEPT your water block which has 1/2" barbs (~3/8" ID)... wouldn't it STILL make sense to just use 3/8" ID tubing because the block would be limiting the overall flow. would 1/2" tubing make any reasonable difference?

(of course gettin a different block would be ideal, but i am looking for an answer on the setup AS IS)




*note that all of the above are assuming that you compare the exact same system, but just different size tubing

please make sure we know if you are referring to question #1 or #2

[decodeddiesel]

Well I deffiniatly agree with your conclusion on your first question. I plan to use drilled 1/2" ID fittings and 3/8"x5/8" silicone with my new system as the smaller tubbing is easier to route and I find it has a slightly smaller bend radius than the similar thick walled 1/2" ID silicone.

As to your second Q I would be inclined to agree with you there as well. The age old saying comes to mind "Your system is only as good as the worse component." I believe this holds true here as well. The only gains I can see with using 1/2" ID tubbing in this situation is it would be easier to fit over the 5/8" fittings. Regardless I believe you would be correct in assuming the 1/2" x 1/4" NPT fitting would be the limiter. It would be cool to see an experiment with flow and the impact in the situation you described above.

[pHaestus]

The head would be (I believe) the sum of all the individual resistances. So assuming that you have barbs on the pump, the res/air trap, the radiator, and the water block then there still is an advantage to using 5/8" barbs wherever possible coupled with 1/2" tubing. Doesn't the head get affected by the diameter of the tubing as well? I would think the inside walls would provide some resistance to flow.

You are also operating under the assumption that 1/4" NPT barbs must be left stock. In fact they can be drilled out substantially so that their ID is much closer to 1/2" than it is to 3/8".

[Brad]

pH is right, if you run 1/2" ID tube throughout, and have one small spot of 3/8", it will have more flow than 3/8" the entire way

[DigitalChaos]

true, you can bore out your barbs... but that just adds more variables to the whole question... making it a mess =)

as for the friction that 3/8 adds over 1/2... what about the fact that the pump would then have MORE water to be pushing up the length of the case... so i wonder which would reduce flow more: friction, or water mass.

[EMC2]

1) would 1/2" tubing make any real difference here?

In flow rate, a definite yes. In system performance - depends on the block and partially on the total length of tubing (which affects net gain from the increased tube size).

2) would 1/2" tubing make any reasonable difference?

Same answers as for the first.

The head would be (I believe) the sum of all the individual resistances.

Correct Key word on all. That includes bends, fittings, etc..

DG - to understand why I gave the answers I did - a given size inner diameter - be it tubing, a fitting, or a bend, has a "resistance" that is based upon both length and diameter (and the velocity of the water and the radius of any bends). In the case of the fittings on the block, while they might only be 3/8" unbored out, they are also quite short in length. The tubing on the other hand has significant length, and actually more restriction per unit length than the fitting does for the same diameter (due to the tube's inner surface compared to the fitting - unless the inside of the fitting is rough - but then if you bore out the fitting, it can be very smooth, lol).

Regarding your last question about the weight of the water - for the lengths of tubing in a puter system - you would benefit more from the reduced tube restrictions than you would loose from any issues in that regard.


I think the real questions you should ask yourself are (let's really complicate the issue)-

Will the block you have benefit from increased flow enough to warrant the increased cost given your intended use?

If the answer to the first is no, then will the benefit be enough to allow you to go to a smaller pump (less $, less heat, less electricity) than you would otherwise use?

Will the reduced flow restriction provide you other benefits that offset the cost of the larger tubing - be those benefits reduced wear on the pump, reduced maintenance, or possibly other things nobody has brought up yet.

Other than the cost of the tubing itself, are there any negative affects of the larger tubing that make it not worthwhile to use? (kind of a why not since you know the flow will increase).

If the plusses and minuses are pretty balanced - then what about future use of the tubing on your next system (i.e. pay for 3/8 now and need/want 1/2 later)

[DigitalChaos]

Originally posted by EMC2

"resistance" that is based upon both length and diameter

First time i have heard someone say that length has to do with it... but that does seem very logical. and thank you for explaining it.


if anyone has a flow meter... it would be really cool to see someone measure flow through a 1/2" barb with like... 5 feet of 1/2" tubing and of 3/8" tubing to see how much restriction happens.

[EMC2]

It can be calculated (but don't ask ). Just to give you a rough idea, there are standard charts for head loss per 100ft of copper pipe. The numbers will not be the same for silicon/tygon tubing, but it will at least give you an idea of the order of magnitude head loss There are a fair number of plumbing related web sites that have those charts, you just have to find one that has 3/8" I.D pipe on it, lol. You can find some numbers on the affect of 90 degree bends as well (note the radius of the bend has a significant impact). The affect on your flow rate will depend upon how your pump reacts to the increased head. You'll have to get a curve for the pump your using to see that.

If you wanted to see for your pump approximately what the difference is fairly easily - get you some scrap or cheap vinyl tubing (say a few feet of each size at the hardware store), put water in something that can hold say 5 or 10 gallons, put your test hose on the pump output, prime the system, then measure how long it takes to pump 4 or 9 gallons (the larger the amount of water you use, the less affect your timing inaccuracy will have on it). Then multiply the length of tubing you used by the length in your system.

[ECUPirate]

here's my two cents...

Regarding your last question about the weight of the water - for the lengths of tubing in a puter system

What goes up must come down. If you have a vertical loop of water, and you try to move the water around the loop, wouldn't the weight of the water moving up be balanced by the weight of the water falling down? Yes, initial inertia (like that?) would be greater, but once you got it moving.... I don't think there'd be much difference between a vertical loop and a flat (horizontal) loop.
I thought head meant vertical lift w/ the tube open on the top end. (the ability of the pump to lift water w/ no return.

Your thoughts??

[Marco]

Originally posted by ECUPirate
here's my two cents...

Regarding your last question about the weight of the water - for the lengths of tubing in a puter system

What goes up must come down. If you have a vertical loop of water, and you try to move the water around the loop, wouldn't the weight of the water moving up be balanced by the weight of the water falling down? Yes, initial inertia (like that?) would be greater, but once you got it moving.... I don't think there'd be much difference between a vertical loop and a flat (horizontal) loop.
I thought head meant vertical lift w/ the tube open on the top end. (the ability of the pump to lift water w/ no return.

Your thoughts??

Yeah that's true.
Also note that pumps pump a constant volume per unit time. So if your cross-section increases your speed decreases, but volume pumped remains the same (this is of course ignoring head loss due to friction).

Examining head loss due to friction, you need to look at the cross-section and examine the circumference (where surface tension resists the flow) and the area. You need to examine the relation of area per unit circumference. Since circumference of a circle is 2(pi)radius and its area is (pi)(radius)^2, it is obvious that if the circumference is increased linearly, the cross-sectional area increases in a quadratic manner, meaning that area per unit circumference increases as circumference increases, thus reducing friction per unit area.

[DigitalChaos]

EMC2 you rock

you can actually lay down numbers (or atleast lay down the fact that you know them)

the straight science behind it is pretty undisputable, and was what i was looking for.

it would be cool to see someone test with a flow meter too, or the tests with how long it takes to empty 5 gallons.
i would do it but i dont have everything right now, plus i always manage to get opposite of what everyone else gets (and its accurate and repeatable when i do it.. eck)


and everyone else that has posted, thanks. its been putting new thoughts through. if you guys have more to add please do.

[futRtrubL]

Lets see if i can remember this right resistance is proportional to lenth and to radius squared, at constant fluid speed, and to speed squared at constant radius. Since cross sectional area incrase as radius squared then flow speed is inversely proportional to radius squared. This gives that resistance is proportional to radius to the fourth power at constant flow(not speed). Resistance is the sum of all resistances along the loop so changing radius of a part will have an effect even if it's not the smallest radius part. However since radius as sutch a huge effect on resistance the parts with the smallest radius tend to be the dominant parts.

As to greater volumes of water in a system causing greater resistance. Nope, it will take a bit longer to start going full speed but it will have no other effect.

Edward

[Brad]

just like a big flywheel, when accelerating there is more resistance, but once at full speed it has no more resistance

[pHaestus]

My wife and son are going out of town this weekend and I will test this all out for you guys. I have an Eheim 1250, a flowmeter (0.5-5gpm) and a bag of fittings already so I will just have to hit Lowe's up for some 3/8" vinyl tubing and some more 3/8" fittings.

Hopefully I will get the results posted Sunday or so.

[jaydee]

Originally posted by pHaestus
My wife and son are going out of town this weekend and I will test this all out for you guys. I have an Eheim 1250, a flowmeter (0.5-5gpm) and a bag of fittings already so I will just have to hit Lowe's up for some 3/8" vinyl tubing and some more 3/8" fittings.

Hopefully I will get the results posted Sunday or so.

Well thats all good about the connectors but what about the channel sizes in the water blocks? If you have a 3/8" channel what good is 1/2"ID barbs? If you have 1/4" channel what good is 1/2"ID barbs (namely the Z4 before they stole mine and fixitts design). There are waaaaaay to many variables in each system to come up with standard numbers.

I see alsot of people saying 1/2" ID barbs work better than 3/8"ID barbs but yet their block has 3/8" channels?

I am probably just missing something though as usual. On water block #2 and 3 on my site I tested 3/16"ID barbs up to 3/8"ID barbs and had no difference in temps at all. I just wonder how many people mod their block to 3/8 or 1/2ID barbs and end up mounting the block better the second time to show better results.

But then again this all falls on what block design you have. Maybe the designs I have made do not matter what size the barbs are? The only one it made a difference on was #4 when I used the center inlet at 1/2ID and 2 outlets at 1/4ID. But that was probably the duel outlets center inlet that helped. Also I never used a pump over 200GPH in anything I use and I get pretty good results.

Bahhh!!! To much rambling, got to get back to the code for my Spiral.

[pHaestus]

I can't test every block out there, but you are right. Unfortunately, you are getting very far away from something that can be controlled in a test then. I tell you what, in some tests I will also add either a 3/8" or a 1/2" length (maybe 6"?) to simulate waterblock's flow restriction. Seem reasonable? It's crude but it's controlled at least.


As anecdotal information, I noticed a pretty substantial performance increase (could o/c 15-20MHz more) by going from 3/8" fittings and tubing to 1/2" fittings and tubing on my original Maze2 (that thing has tiny channels and a ton of turns).

[DigitalChaos]

jaydee, yup thats another factor, but harder for people to know unless they made the blocks =). but if you know then im sure the same logic will apply to it.



pH- Thank you!!! very cool to see someone want to help out like that.

[DigitalChaos]

pH- idealy it would be cool to see a pump going ONLY through a 1/2" barb (3/8" id) or 2 of them.

and then 6 feet of each type of tubing


but if you have a water block with knows ID of the channels being 3/8" or bigger. that should work as well

[pHaestus]

I am going to do the barbs in pairs because you would never have only one oddball barb in the setup. I was planning to use 3 sets of barbs, and start with 3/8" barbs and 3/8" tubing then replace 1 pair with 1/2 and repeat then the other pair and then after having all 1/2" barbs switch to 1/2" tubing. At that point the test would split and I would insert a 6" 3/8" tubing to simulate a block with restricted flow for half of the measurements and then go from 1/2" to 5/8" 1 fitting at a time.

Seem reasonable?

[jaydee]

Originally posted by DigitalChaos
jaydee, yup thats another factor, but harder for people to know unless they made the blocks =). but if you know then im sure the same logic will apply to it.



pH- Thank you!!! very cool to see someone want to help out like that.

But if you don't know then non of the stuff in this thread applies if used right? I do not see anything valid in anyones comments if they do not know the size of the channel in the block they are testing all this with (not refering to anyone in this thread BTW, just generalizsation)as that is a major key part IMO. All else is a moot point unless that is known.


Will be good to see thge results of that connector test though as I DO know the sizes of the channels in my blocks.

[pHaestus]

You should be able to pull that information from the use of a 3/8" length vs a true 1/2" channel though Jaydee. Even still this is a simplification to the point of being incorrect; we know that elbows and turns in the waterblock will also affect the flow rate. It's a reasonable approximation though I think.

[DigitalChaos]

sounds good.. as long as you have 1/2" barbs (3/8" id) tested with a run of 3/8 and 1/2 tubing

and if this is gonna be any type of write up for procooling (non web board) make sure ppl know when you say 1/2" barb.. that they know if you are referring to its technical name or the ID of the barb

thats a big thing alot of ppl dont realy pay attention to.

[pHaestus]

Yea I always refer to barbs by OD since that is how you buy them at the store. I think I will add some 1/2" copper tubing to the mix as well; that is pretty common choice for DIY blocks

[BillA]

pHaestus

you're digging yourself into a hole
for the theory get an intro text to Fluid Mechanics
for a more practical run at it try Crane's Flow of Fluids Through Valves and Fittings
hit Google on 'flow resistance' and 'differential pressure'
here is a pdf of a educational test rig for exactly this subject

there are 2 parameters you need to measure:
the flow rate
the pressure drop (differential pressure across the item being tested)

and you'll crap your drawers when you measure the pressure drop across your flow meter
remember; each incremental flow resistance in the system is additive

[pHaestus]

There is an adage at the pH household:

"When BillA replies to your threads, testing becomes much more expensive."

Looks like to do properly I would have to buy a fair amount of equipment. Pressure drop transmitters and reading equipment doesn't sound cheap. I am aware that the flow meter I have seriously affects flow already, but I dont have the cashflow to buy any new testing equipment at the moment.

Before I go dig up the books tomorrow, how robust are the equations? Could I reasonably do some number crunching in Mathcad and extract the same sort of information? I suppose in the end what the people who originally started this thread are interested in is how large the contribution from using smaller bore barbs/larger bore barbs compared to using smaller/larger tubing. That seems like a solvable problem to me, but I live in the realm of the semi-quantitative.

What other parameters are there besides pressure drop and flow rate? Why can't you simply estimate the pressure drop from the difference in flow rate prior to and after addition of a restriction? I apologize if I am asking a foolish question; I will read up tomorrow (if the University library is open). I still have some required reading on turbulent vs. transitional flow regimes anyway