Printable View
Flowrate is what matters, but pressure leads to flow rate (to a certain extent). As pressure allows the water to overcome obstruction. Ie a pump with 30000l/h max flow and 6" max head is useless, but 1000l/h max flow and say 16' max head is useful.Quote:
Originally Posted by Herr Brun
No1 answered my question :(
beating a dead horse.
Beating a dead horse
From Wikipedia, the free encyclopedia
In American English, "beating a dead horse" is an idiom which is most often used as a retort used to make clear that a particular request or line of conversation is already foreclosed, mooted, or otherwise resolved.
The linguistic roots of this phrase draw on an allusion to literally "whipping" or "beating" a deceased horse in order to make it get up and go. Figuratively, when one is "beating a dead horse", one is flailing at a dead or useless idea.
Yeah, but I thought more pressure at the start = more flow, but I'm probably all wrong. Really new to wc (just read my thread about the dead mobo and gpu :stick: ), and still doing my first year of physics in high school.Quote:
Originally Posted by nikhsub1
That's exactly what I thought. Logic seems to work to some extent ^^Quote:
Originally Posted by Atomicpineapple
It depends on how much pressure and how much flow you're dealing with since these pumps are not positive displacement.
For example you have a pump with 10000gph and 16' of head. Now imagine what happens when that pump gets restricted by 50%. A 50% restriction on a centrifugial pump means ~50% of it's energy gets transfered into the water through friction or whatnot.
That's a lot of energy being thrown away and more importantly being put into the water. Now logically one would think "but hey we still have ALL that flow" well the amount of heat the pump puts into the loop becomes more than the radiator can dissapate and heats up the loop. This works on a smaller scale in a similar way i'm just not sure of the actual impact. It would probably be calculated by the amount of force needed to move the amount of water head the pump loses a certain distance in a given time.
Seems we need some positive displacement pumps to make WC a lot less annoying.
Tero, sure, but then you get into pump design, which I know naff all about. With my example above I wasnt thinking about including the thermodynamics/ fluid mechanics of the pump into the equation, just using a mysterious force on the water to explain my point about head leading to flow.
I am not sure I understood you correctly, are you saying a restriction leading to 50% less flow, would increase heatdump ~50%? Lots of the the electric energy is still converted into kinetic energy, the water is just pumped inside the housing instead of entering the loop and with increased restriction the pump's rpm will drop (up to 100% restriction where all the water is pumped inside the housing), reducing power consumption.Quote:
Originally Posted by Tero
@nik: Well, if you use two individual (modded) tops, yes. I don't have enough space for that, so I had to use the one piece design tops, which don't give an improvement of 35%.
No, no way. Maxxxracer.Quote:
Originally Posted by L'enFer
oh god, you are going to try to correct maxxxracer? :rolleyes: have you seen who is the guy that made most of the stickys here? if he says the water should be black for better cooling is because he knows it :stick:Quote:
Originally Posted by Vikodemous
now please use some logic here we dont buy better pumps with more flow or head preasure because we want to sound cool its because it improves temps.... so like maxxx said go and read the stickies before you spit out more nonsense
But you explanation is bogus(or whatever they say in english) and you obviolsly dont understand waht is happening in the loop...Quote:
Originally Posted by Vikodemous
I dont mean to be harsh, man we all started out knowing nothing. just read a bit more and dont be so ":banana::banana::banana::banana:y" when coming with thoose false statement.
I cant see how virtualrain running the pumps at 7v then at 12v is going
to cause a 3deg temp swing as adding a northbridge block in the loop will
dump more heat than that in the water.
You never hear of people saying that they have had a 3deg temp increase
when adding a chipset block as the rads can dump 100's of watts of heat.
?
That's not what i said. I said that the friction of the impeller would result in the kinetic energy of the moving water being changed into heat. It has to go somewhere. This has nothing to do with with the amount of electricity the pumps motor uses to achieve the kinetic energy. The only way to figure that out is to figure out how much power the pump converts to physical energy(its efficeincy) then figure out how much of that energy is lost.Quote:
Originally Posted by Fairydust
The rpm drops due to friction and if the pump is being fed power to the motor at the same rate that power has to go somewhere. Meaning that when the rpm drops it either drops because energy is being changed into heat or the efficiency of the motor drops.
What do you suppose that water inside the housing is doing? I'm sure it isn't cooling down.
It's not as direct as i'm probably making it sound but, it does happen. This is just the reason why head is better than flow for WC.
I spent half the afternoon reading the HVAC Pump Handbook by James B Rishel and still cannot wrap my head around some of the concepts.
A couple of things I did discover though was never dump excess flow back to the suction without sending it to the radiator first, and yes there is a such a thing as excess flow or churn which is water just spinning in the pump building up heat and not going anywhere.
It's just a forum stop taking things so personally. You were wrong deal with it.Quote:
Originally Posted by Vikodemous
Ok this is going to be a long post.
In response to Reinvented, Hound53, Fairydust, and Tero
As I will mention below, the heat dump into the water from the pump is mostly not from the impeller friction but rather the electronics and stator.
Next up, the effeciency of a pump is pretty easy to calculate if you know the delta P (pressure), flow and input wattage of the pump. With the pressure and flow you can calculate the hydaulic pumping power the pump is exerting on the water. This is the amount of work the pump is doing check out this review of the MCP655 by Robotech for more info on hydraulic pumping power. (Hydraulic Pumping Power ) . Now that we have the amount of work the pump is doing we just divide that by the input wattage which can be measured with any good digital multimeter. Equation goes as follows.
Effeciency = HPP/Wx100
HPP = Hydraulic Pumping Power and W = Input wattage.
In Response to virtualrain:
1.
Please let us know which flowmeter (the exact model #). Looking at the models you linked two, the only useful ones would have a pressure drop of 1.6 to 4.5 feet of head (depending on model, and at 5litre per min which is 1.3gpm). that is ALOT for just a little flow meter and the pressure drop gets logrithmically (i cant spell so sue me) higher as flowrate increases. This might partially explain the issues you have been experience. My reccomendation is to remove the flowmeter
2.
I am not entirely sure what you meant by "heat dump theory", but it is a fact that the pumps input a good portion of their heat into the water. BUT the heat is, for the most part, NOT from friction (KE) of the impeller and water inside of the impeller housing. Most of the heat comes from the electronics and stator. A perfect example is the Iwaki pumps. Their design insulates the impeller housing from the stator, thus keeping the heat dump in the ~50% range. The DDC, and D5 on the other hand have the stator housing in direct contact with the water and their heat dump percentage is 80+ (dont have the exact numbers off hand).
3.
If I remember some correctly some relatively high accuracy tests were done with calibrated platinum RTD probes with 0.01C resolution and the deltaT (difference in temp) between waterblock inlet and outlet on an 80watt source. the deltaT on that setup was between .5C and .15C depending on flowrate. So you .3C is probably relatively accurate although I am unaware of what sort of temperature measurement equipment you are using. Check out the info hereQuote:
In my case, the water temperature as measured just before the rads is often only 0.3 degrees C warmer than the water temp after the rads.
4.
As you pointed out as well as others, it is very likely that your system is just mismatched. The incredibly high pressure of the dual DDC's combined with a relatively low restriction system may be leading to bizarre issues with the DDC's. My suggestion is to try the loop with one rad (you can keep both in if your lazy) and one pump, with the inline temp and flow meters removed. Your system will probably work alot better like this. Also, some pictures of how you have your rig set up would be very beneficial in determining the cause of your issues as well as enabling us to help you redesign your loop for maximum effeciency.
In Response to Firelord-OCHW:
Loop Order does not affect flow so long as your reservoir is right before the pump. Other than that the loop order does not affect flow so long as all other things are kept constant (ie: tube bends and tubing length)
In Response to Vikodemous
I said it once and I will say it again.. Go read the guides. The "information" you are providing is a complete reversal of the truth and has no basis in reality. cooling effeciency in radiators and waterblocks increases per molecule of water as flow rate increases up to a certain point (much higher than any normal watercooler will achieve) and then will stay relativley constant or slowly drop off. The very smart, but rude Bill Adams demonstrated this property to the watercooling community in his radiator tests. If you like I can dig up some links to show you.
In Response to Herr Brun & Atomicpineapple
The actual pressure of the water in the block does NOT matter so long as its not negative, which to be frank is not really possible in a watercooling loop (barring the pump inlet). To add some more detail to that statement; the closest thing to the pump gets the highest water pressure and the pressure keeps dropping off until it hits the end of the loop (the reservoir) where it then becomes the same pressure as standing water (assuming the reservoir is open to the air). But while this is true, it does not play any role in the performance of the system, its just a little useless factoid (well almost useless ;) ) because the flow will remain constant at any point in the loop (sure water velocity changes but now overall flow).Quote:
But that [pressure] doesn't matter, if I understand correctly? Pressure does not matter, flowrate does? Or is it different with impingement (sp) style blocks?
Now head pressure of the pump on the other hand is VERY important. We are not talking about water pressure here, but rather how much pressure can the pump exert on the system. As long as you stay within the confines of Cathars guidelines for pumps (located HERE), looking for a pump with the highest head pressure is an excelent idea, and is the reason why the DDC is so popular. You just have to make sure that the pump has a properly proportioned PQ curve so you dont end up with a pump like a fuel pump for a car (insane pressure but very low flow).
OK. I think I covered everything that needed addressing. And if you want to complain about my spelling, go right ahead and post it in the official complaint deparment located HERE
Hi, first of all thanks again to everyone for their interest and input!Quote:
Originally Posted by MaxxxRacer
The flow meter I'm using is the 46500-164-4F16,5 (it's the one that produces 700imp/l and is rated for 2-35 lpm and has minimal pressure drop below 10lpm. I doubt that it affects flow significantly... probably less than a block since it has an ID of 8.5mm which is reasonable and the turbine is very streamlined. Never the less, it is a novelty and although I like it, I may remove it (more on that below).
The link again for reference... http://www.remag.ch/vision.php?language=4&sub=1&tab=2
Thanks... no further comment.Quote:
Originally Posted by MaxxxRacer
In my case, I'm using Innovatek thermal probes like these...Quote:
Originally Posted by MaxxxRacer
http://www.highspeedpc.com/Merchant2...de=FanTempFlow
One probe is in the unused inlet port of Pump #2 to measure the temperature before the rads. The second one is actually epoxied into a hole I drilled in the outlet area of the second rad.
They are connected to an Aquaero which was used to calibrate the probes.
Both probes were calibrated at a variety of water temperatures on the bench before being used in the loop.
I agree. Please note that being busy is different from being lazy! Hence, I won't be making any changes to my loop in the near future other than just running both pumps at 7V. When I have time, I will likely start by adding my GPU and a Storm to the loop. Hopefully that alone will address the mismatch. If there is still a mismatch then again, I'll probably just run the pumps at lower speeds.Quote:
Originally Posted by MaxxxRacer
Next time I have the loop apart, I will use some buckets to assess the impact the flow meter has on the loop and also to test it's accuracy which I was unable to do before assempling my current loop. If I find it is negatively impacting the loop in a significant way, I will remove it.
For me, at least, this thread has been helpful, and I'm not sure what more can be done.
I'll try to post some more pictures and update my blog soon.
Going out on a limb here....if it weren't the added heat dump or some funny stuff with the pumps, could it be the TIM?
If it were a new setup, AS5/Ceramique would show that progression of temps as they break in. Even with the lower performance of the pump, the broken-in TIM could easily outperform fresh TIM. On top of that....the true difference between 12V and 7V may not be more than a fraction of a degree with your loop.
virtualrain: when you add your gpu block, you could also consider doing two seperate loops with one pump/rad for each loop - I have read a few places that two seperate loops can give better performance, tho i am truely a newb and havn't got any experience myself to back it up.
well then. you got that pretty well figured out. Most people run those innovatek probes with tin little T shaped compression barbs which kills the flow. Nice to see you thought ahead on this one and didnt restrict the flow.Quote:
In my case, I'm using Innovatek thermal probes like these...
http://www.highspeedpc.com/Merchant2...de=FanTempFlow
One probe is in the unused inlet port of Pump #2 to measure the temperature before the rads. The second one is actually epoxied into a hole I drilled in the outlet area of the second rad.
They are connected to an Aquaero which was used to calibrate the probes.
Both probes were calibrated at a variety of water temperatures on the bench before being used in the loop.
Based on the specs of that flow meter, you appear to be correct; It has very little flow restriction in the 5-6lpm which you are working in.
As to the wierd results you obtained with the flow meter, it is probable that the hardware you are using to read the pulses cant read the high pulse count of the meter at higher flow rates which causes it to give crazy results. you can pick up a pulse counter on ebay for around 20 -40 dollars that has high accuracy and a very wide range. Such an item might be of some use to you.
Quote:
Originally Posted by Hound53
100% aggree.
is like resistance after a while is just gets exponential and you get so little for so much...
For those interested, I added some pics to the OP. I also updated the blog tonight with more pics on how the rads were mounted to the fans and side door... (click the link in my sig)
MaxxRacer thanks for the info. Excellent link, I may actually get a decent grip on this stuff one of these days. A big maybe there but still possible I suppose.
exhasted mule - the part about dumping back to the pump suction was bad advice on my part, any dumps would have to be back to the radiator or a heat build up would occur according to Rishel. But if Maxxracer is correct, and I have no doubt that he is, the heat issue from restriction is moot anyway.
Wonder if the leading edge of Diamond shaped pins of the apogee at
high flow push the water to the outside edges of the block and cause
a dead spot at the back of the pins. ?
Abit like the front of a ship pushing through water.
Just a thought.