A little head...?

[UnloadeD]

OK, here's something that has always bothered me since I started looking into watercooling. Pump performance is ALWAYS presented in relation to the amount of head on the system. Waterblocks, rads, tubing & fitting specs NEVER say how much head the add to the system. I always figured it would take a genius with a super computer to calculate any meaningful figures in this area, and that is why they are never seen. Well after tinkering with my setup and thinking about some others a simple testing method has come to mind. Lemme show a pic first:

OK, the simplified version would be a setup like this. For arguments sake lets say the cylinder is 1"ID, tubing and fittings are 1/2" ID. You fill the system enuff above the top fitting so water is calm enuff to mark a reading. Now its calibrated. Next you add the item you want to test to the Test Point, lets say a waterblock. You add the same amount of water back to the system, plus the amount that fills the waterblock. Fire it back up and see how much the water level in the cylinder has risen. The difference between the two readings is the amount of head the waterblock has added to the system.
Well that's my idea, I know its simplified but I also know how well the PC staff can setup testing to eliminate variables and such. I also know they are working on a waterblock roundup. I woul love to see some kinda test like this done on them. What do you guys think of this idea? Are there flaws in my logic? Is it doable? and if it is...would anybody else besides me like to see some real numbers things like: series vs paralell, rad vs heater core, 3/8" vs 1/2", quick connect fittings, elbows..... not to mention how your pump is really doing in the system.
I know I would be interested in seeing this info. I don't think it would be too hard to set up a test comparing items. Figuring out the exact amount of actual head on a given system might be more difficult. I know if this can be done, PC could do it reliably and I would love to see them be the first to offer this info to the community.

peace.
unloaded

[MadDogMe]

That's not the head though is it?, that's displacement, it's the blocks resistance that afects head. not it's displacement...
Head is the same measurement as PSI, as miles is to kilometres.

[UnloadeD]

No it wouldnt be measuring displacement. The back pressure created by the waterblock's restriction would cause the level in the cylinder to rise.

BTW I'm suprised you are the only one to comment on this, seems its gotten lots of views but only one reply. If I'm wrong on this, please tell me. Also explain the flaws.

peace.
unloaded

[Skulemate]

I think you need to think more about how you're measuring the "head" in your system. A vertical column of water will work, but not set up like you have it (think along the lines of a manometer tube... it doesn't have to be a very large diameter tube at all). The way you have it drawn your cylinder is acting as a reservoir... the surface of this may ripple a little when you're running (mine certainly did) but the level is not going to change a whole lot (certainly not by the same amount as the resistances in the system)

Here's something to consider to help you find the flaw in your thinking: suppose that around 0.5m H2O resistance is caused by your waterblock for a given flow rate... that means that, by your reasoning, the cylinder level will go up 0.5m also... question is, where did that additional quarter litre (0.253L if I calculated correctly) of fluid come from?

Oh, and MadDogMe, measuring the pressure of the system with a column of water is the same as using a pressure valve... you're going to have different units, but you're measuring the same thing, and the two can be converted easily.

[murray13]

Try this out:



There are a couple of things to remember. The tee must be the first thing after the pump. And you must not let any air get into the pump. The smaller the tube the better you will be able to see a change.

If someone thinks I'm off base here I'd like to know why. This is how some flowmeters work by measuring pressure.

Hope it makes sense.

[Skulemate]

I don't believe the tubing diameter plays too much of a role in this, though I agree that there is no sense to use anything over 1/8". However, I think that it's best to do these tests with some sort of a reservoir, otherwise, where will the (albeit small) amount of water needed to fill the manometer come from? Furthermore, to do this correctly, you'll need a manometer directly before and after the thing you're measuring... otherwise all you have is pressure at a point.

[UnloadeD]

Hmmm, this seemed so less complicated when I was first thinking of it. I diameter of the cylinder shouldn't be a factor as one foot of head will be the same amount of pressure in a 1" or a 10" diameter column. Two problem areas come to mind tho. First would the suction of the pumps intake throw this test out of whack? Would it be better to have the test piece dumping into an open res and the pump intake also drawing from same res? Something like this?

The other possible problem that comes to mind, as the level in the cylinder rises to indicate the amount of head added by the test piece, the performance of the pump drops. I'm not sure this would effect anything other than the flow.

The more I ponder this, the more apparent te limits of my knowledge becomes. I'm hoping some of the guys more versed in this area will chime in, explaining it better. Hopefully in layman's terms so I can grasp it 8)

[Blackeagle]

Why not add a psi guage into the line instead? Have at first only the guage and say 1.5m of tube in a loop, this will give you a base line reading. NOw add a rad and take a second reading. Then as you add each part of your system you see the growing pressure and know where it's coming from.

Or for testing WB only. Once the above set up is made and you switch WB at any point then the change in psi will show if the differant block has higher or lower pressure than the one it was subed in for.

And you don't need any tall tubes at all.

Now someone show me why this is screwed up please. I'm new but this seems like it should work, to easy but I can't see the fault.

[Skulemate]

Well, your setup is better, but still not quite enough. The cylinder should work well enough like that, though you are still going to get a fair amount of restriction moving through it. Why not use a tee instead of a cylinder with two fittings? Furthermore, with only one tube all you've done is measure pressure at one point, before the block. What you want to do ideally is have two tubes (cylinders) directly before and after the item tested, in order to get a pressure difference. Maybe something like I've shown below would work.

Blackeagle: It's hard to find pressure gauges that are accurate enough to read the low pressures that are present in our systems... the ones that are good enough are fairly expensive. Hence use the tubes, since they're much cheaper (though not really suited for use aside from testing).

[BillA]

Skulemate's setup is fine
google for 'differential manometer'
Owenator described this in the flowrate sticky on OC
- and I even think he wrote an article for JoeC on it

[UnloadeD]

skullmate, i think the only thing that second tube in your image will do is show the level of water in the res.

[Skulemate]

I'm sorry UnloadeD, but you're wrong. That second tube will show the amount of pressure at that point in the system (it's a simple application of Bernoulli's equation...). As BillA (unregistered) has suggested, you may find that Google is a good source to read more about what I have suggested.

[airspirit]

This topic is blatantly mislabelled. Where is the pr0n? "A little head ... " ....

The second tube, if set up right, can show a difference in the overall volume of the res, but nothing more, methinks. All you need is the first tube in order to get a measurement, since the measurement difference is all you're after, anyway.

[Skulemate]

The second tube will measure the pressure at that point. If the pressure there was equal to the pressure in the res (which you're implying is the case by saying that the tube will read the same level as the reservoir) then there's no driving force behind the flow from that point on. Flow occurs because of a pressure differential (well, energy really, but whatever...). Really, honest... I've used a differential manometer before to measure the permeability of concretes and soils, so I am sure about this. But hey, if you don't believe me, ask Dave or BillA.

[MadDogMe]

Sorry, I should have read your post better UnLoaded...

I can see how Skulemates pic would work, even if the level was the same in both tubes and the res to start, it's the rise/diffence once the pump has engaged that matters is'nt it?, and the level WILL rise because of the restriction of the CPUblock won't it?. *EDIT* if the reservoir is closed/sealed it's moot anyway is'nt it?. I like this idea cause it's measuring componants in a 'sealed closed loop', not a pump pushing only scenario...

My question is how do you translate your result to equate with the pumps 'head rating'?. the amount the level rises is entirely dependant on the diametre of the (test) tube is'nt it?, the wider it is the less it will rise?. yes/no?. if so there has to be a 'standard' for head measurement does'nt there?...

Or is it the 'volume' that rises?...

If my question is'nt pertinant or well off base just kick me and I'll shut up and stick with stuff I understand like picking my nose and rolling fine cone shaped hand made cigarettes in liquorice papers ...

[Skulemate]

No, the diameter of the tubing won't matter one bit. Once the system has reached equilibrium there would be no change in the readings (though it would take longer to reach equilibrium if the tubes are a lot larger).

To convert a measured column of water into a pressure you need to multiply the column height by the specific weight of the fluid (specific weight is simply the mass density times gravity). Note though, if you're using imperial units, you'll need to make sure that you're working in the same units.

As for relating the results to the pumps head rating... well, you really can't without using the two manometers, one on either side of the pump (that will tell you how much head the pump is producing).

[MadDogMe]

Do you mean the manometre tube?, that's what I meant(assuming use of ghetto, not calibrated)...

this is all a bit beyond me I'm afraid, which is a shame cause it's quite an important aspect I think...

[Skulemate]

Yeah, I meant the manometer tube. Doesn't really matter what diameter it is. Though smaller will generally work better as it requires a much smaller volume of water (therefore less change in the surface of your reservoir, it'll reach equilibrium faster...). I think we used 1/16" ID tubing for the manometers in the permeability cell at school.

[murray13]

If all you want to do is measure the pressure drop of some item in a test, you can put a valve in the line after the second manometer and restrict flow which will increase pressure to both manometers allowing you to be able to see more clearly the change. It will also let you set up a common starting point for all tests. You just have to keep some flow going so the pump does not get to it's shutoff point.

The biggest problem you will have is that you will be looking at relatively small changes in water level in the manometers. I think...

The smaller the tubes used for the manometers the faster they will respond to a change. You don't have to wait as long for the system to reach equalibrium.

[UnloadeD]

Originally posted by Skulemate
No, the diameter of the tubing won't matter one bit. Once the system has reached equilibrium there would be no change in the readings (though it would take longer to reach equilibrium if the tubes are a lot larger).

To convert a measured column of water into a pressure you need to multiply the column height by the specific weight of the fluid (specific weight is simply the mass density times gravity). Note though, if you're using imperial units, you'll need to make sure that you're working in the same units.

As for relating the results to the pumps head rating... well, you really can't without using the two manometers, one on either side of the pump (that will tell you how much head the pump is producing).

Well, I invisioned using water as the fluid, as pump head ratings are usually given as units of water. BTW your last paragraph shows that you and I aren't after the exact thing. I was mainly interested in seeing how much head a given device or string of devices adds to the system. I really wasn't concerned with what the pump is doing. I'm assuming that any pump would show the same results as long as the amount of head isn't enough to stall the pump. Is this correct or would flow rate have an effect on the results? The main reason for all of this is that I'd like to something like this in a WB round up " WB-A will add x" of head to your 1/2" system..." and maybe a bar graph showing the amount for the different WB's. Its just more info at hand when deciding to build a system. If I were able to see the amount of head added for cpu-wb, gpu-wb & a rad and added them up to equal 2-1/2' of head, that would be very useful info to have while shopping for a pump for my system. The performance curves of the pumps would then hold more meaningful info for me. You've got to bear in mind that a lot of you guys on this board are waaay more knowledgable about this stuff than the average user(including me). So what I'm really after is meanful numbers that are easily digestible. I read Bill's articles last night and for me it was information overload. I also noticed he measured the pressure drops across the blocks. I'm assuming this is very close if not the same as what we're talking about with only the units of measure being different. I also assume that his numbers/units could be converted to or used to calculate the amount of head. But like I say, this is a bit beyond me atm. You guys did answer my question if it was doable. Now I hope somebody does it 8) Thnx for the input and if theres more to add especially towards making this practical, please do.

peace.
unloaded

[Skulemate]

UnloadeD, we are after the same thing. The reason you need to concern yourself with the pump is because it impacts the performance of everything else in the loop. The head loss of any object in the system will increase with increased flow. If you were to test the pressure drop for your block at several different flow rates you'd get a curve that initially started out quite low for small flow rates, but it would increase exponentially as the flow is increased (since the pressure drop is related to the velocity squared). Try to understand that the pressure drop for the waterblock is really a series of drops as the water moves through the block... first a loss at the entrance barb, then new drops for all the tight turns and the channel cross section, then another exit loss. It's the combined effect of all of these losses that determines the pressure drop of a block.

Unfortunately, what you propose is not as simple as it seems. You see, each of the components will have resistance curves similar to the block, and these need to be combined to create a pressure drop curve for the entire system. Only then can you use the pump's head curve to determine what your flow rate is going to be. I've illustrated this concept below (with some made up data, so don't be too critical of my values).