Pump mod for High Head

fhorst · 12-01-2003, 04:04 AM

I want to start a discussion about pumps.

Most of us use aquarium pumps, which are made for "normal" usage. A few of us bought an expensive Iwaki pump, which is made for High Head.

There is a nice graph, where all the pumps are http://www.procooling.com/~phaestus/temp/flowcurves.jpg and from that graph, we can see that the MCP600 is "the best".

I have not seen a discussion on why that one is the best, or why one pump is better than the other one.

First some theory:

The function of the pump's motor is simple: outside magnet and inside magnet. The outside magnet will due to the power (220v) become a north or a south pole magnet. The inside magnet will move due to the pole change. (50 or 60 Hz)
Or as Iwaki tells it: "The drive magnet is magnetized barium ferrite for 4-, 6- or 8 poles. It rotates the impeller by using the magnetic field from the outside of the rear casting

No brushes, so nice silent, and long lasting. Downside is that it is a weak motor.

All motors seem to run at about 3000 Rpm.

The motor is connected to the impeller:

So, what makes the head? And why?

As far as I can see, when a pump comes to its max head, the pressure is to high for the impeller to turn around. Actually, as I stated, the motor is weak, and it first will "slip", missing a few pole changes and finally will stand still. (With a bigger motor, having bigger magnets, this will take longer, and therefore give more head)
When the pressure drops the impeller starts moving again, till the pressure is again at the top. This goes fast, so you won't even notice. Except that there is less water flow.

From the Iwaki website, we can see that the pump motors stay the same, but they change the impeller for “Normal usage” high Head, or High flow.

This got me thinking: what if we modify the impeller of our “Normal usage/High flow” pumps? Simply look at the High head design. The impeller wheel is a bit bigger, and has no fins. The impeller housing has changed also, but still……

If my theory is correct, all we need to do is to take the impeller, and take of the outside of the impeller fins, (cutting it diagonal). This will kill our free flow rating, but it should improve our head a lot. And for water cooling, we are not interested in free flow, but in head… That’s why the 600 l/h MCP600 pump out performs a 1400 l/h pump. We are having only a 5 l/m in our setup.. (300 l/h)

Please share your ideas and theories.

bigben2k · 12-01-2003, 07:09 AM

From what I've seen discussed, there are two things that you can try immediately, to get more head:
1-restrict the outlet

2-given an open impeller, cap it with a doughnut shaped plate.

Here's the impeller to my Little Giant 2-MDQ-SC:

fhorst · 12-01-2003, 07:33 AM

Thanks BB2K for the fast reply.

I'm reading some other posts also, and there is a lot to discuss!

I'm lost with "2-given an open impeller, cap it with a doughnut shaped plate." = flat plate with hole in it? or realy capsulate it..?

the restriction of the outlet is what we all do with our nice waterblocks...

it looks like that a pump with a head of about 4 meters giving at that head a flow of about 400/500 l/h is the best pump to get. and silent, and small, and low wattage...... and at a normal price ;-)

That pump just is not there.... (for what I can see) The "cheap" ones all use the "wrong" impellor.

I liked this post: making a 12v pump...

And that one also sure makes you wonder. one watt of power, 170 l/h with max 20/30 cm head...

BillA · 12-01-2003, 01:07 PM

the MCP600 has a closed impeller per the 'high head type' shown in the Iwaki dwgs
- and that is why it is 'better' (for WCing applications)

understand that cheap AC motors can, and will, start and rotate in either direction
this is why aquarium pumps have straight vanes, reducing their efficiency even more

you will not have any 'luck' converting an open impeller to a closed one
no free lunch

#Rotor · 12-01-2003, 01:37 PM

the way I understand it.... head it a result of a couple of things, or characteristics of the pump.

1: tolerance in the impeller casing, IE. how much space is there for liquid to squeeze by the impeller.

2: The diameter of the impeller.

3: rotational Speed

#1 is obvious, so is #3

#2 includes a number of other factors too... Such as the power/torque of the motor. As the diameter of the impeller increases, the amount of kinetic energy transferred to the liquid by the impeller, as a result of the centrifugal force exerted on it as it goes through the pump, increases dramatically. This in turn requires more power from the motor (nothing is for free).... Torque in Particular becomes an issue, as the diameter of the impeller increases.

I think of it this way, you have a ball and attached to it, is a piece of rope. You want to through this ball as high up into the air as you can....
What will give you the best result.... holding the rope at 10% , 30% 70% or 100% of it's length, while you swing it.....

by the way.... this only pertains to centrifugal pumps....

Blackeagle · 12-01-2003, 02:49 PM

This topic brings us to the question of, 'How would the Iwaki pumps have compared to the Swiftech 600?

I'd have loved to see a MD10, MD15 & a MD20rlzt pumps included in the testing.

Even the MD10 should, on paper anyway, offer more system flow than the Swifty 600.

I will grant you the higher cost of the Iwaki's will turn many water coolers away, but they are supposed to be the best. And the three Iwaki's compared to the Swifty 600 would have also shown which of them would be best bang for the buck.

One of the biggest advantages of the Swifty pump is that it hooks up to the PSU.

bigben2k · 12-01-2003, 03:41 PM

Quote:

Originally posted by fhorst
Thanks BB2K for the fast reply.

I'm reading some other posts also, and there is a lot to discuss!

I'm lost with "2-given an open impeller, cap it with a doughnut shaped plate." = flat plate with hole in it? or realy capsulate it..?

...

Yeah, a flat plate with a hole in it.

Yes, you have to have the clearance for it.

Otherwise, I'd recomend getting that free book from Alfa Laval, which explains a lot different pumps (sorry, no links).

In short, a lot of the pumps are designed for peak efficiency (even though mag drives are still overall inefficient) at a specific flow and pressure range, often somewhere in the middle between min and max.

If you want to experiment with this stuff, I'd recomend picking a pump, and a few extra impellers and housings, and try different variations. If you really want to get into it, then try different motors. The trick then becomes matching the mag drive system to the pump specs you're trying to get to.

Otherwise, stick to the small mods I mentionned above, but don't expect any miracles: sometimes all your best tinkering still won't improve anything, and if it does, it's pretty marginal.

A fun exercise...

BillA · 12-01-2003, 04:03 PM

Ben
you are an unending source of bad info
do you actually know what a closed impeller is ?

so then how do your instructions yield such ?

and you are the WBTA, god spare us from your ignorance

Levish · 12-01-2003, 04:40 PM

I'm not the foremost expert on pumps but aren't fountain pumps working against a high head?

Perhaps looking at the design of those (or maybe even using one instead of the aquarium style pump) would be appropriate if higher head was desired?

fhorst · 12-01-2003, 04:58 PM

Quote:

Originally posted by unregistered
Ben
you are an unending source of bad info
..........., god spare us from your ignorance

LOL BillA, Grumpy again?

I guess we are all an unending sourse of ignorance. You know that one fool can ask more then a hundred wise men can answer.....

(and some times I think it's quite foolish to play with this water stuff, but boy, loads of fun, and nice temps!)

Getting back on topic again... IMHO it's a pain in the ... I've been looking though posts about pumps, specifications, etc. for about 2 months now. (and I've seen a lot!)

There are good pumps out there, but at a very high price.
There are also cheap aquarium pumps, with one downside, the impeller and housing ain't made for high head.

Sure, the MCP600 has the right impeller, but there is no MCP1200.

If I look at the stuff I want to cool, and the number of rad's I use, I need 2 MCP600...... (and with the current problems and availibillity in the Nehterlands, I gues I need to skip on them)

One nice thing I found, is that a lot of pumps are produced by the same factories, just a different label, and sometines a differend impeller, and there we have a new brand, with all different prices. I've seen the MD15 for 80 euro's, but also for a 175..... just a different direction of usage. Who makes the MCP600?

If I look at the prices of aquarium and pond pumps, there must be a "cheap" one, with the right "high head" impeller.
(I just haven't found it...) The ones I've seen are in normal aquarium type about a 50 euro's, and with High Head about 175 euro's! (same motor, other casing and impeller)

I hate this!

There must be some way to get arround this.

I don't mind spending a 150 bucks on a pump. Currently I have 2 L20 in serial but that is about half the pressure I need.....

fhorst · 12-01-2003, 05:02 PM

Quote:

Originally posted by Levish
I'm not the foremost expert on pumps but aren't fountain pumps working against a high head?

I wish you where right. But you have a nice fountain, if it is about 60 cm high, not 4 meters ;-) so all aquarium/pond pumps work a bit the same.

I have been thinking also to get me a central heating pump, but I'm affraid that will kill my blocks and radiators (to much presure)

bigben2k · 12-01-2003, 05:04 PM

Yes, I know that the closed impeller doesn't have curved "fins" in it. Something else to try.

Like I said, get a few spares, and experiment.

WAJ_UK · 12-02-2003, 06:58 AM

You can get turbine style impellers where there are stator vanes in the housing around the outside of the impeller. They curve the opposite way to the blades on the impeller and expand towards the outside of the housing slowing the flow down thus increasing the flow pressure. Sorry no links just heard about it somewhere. I don't know if there are any pumps around with housings like this. They work a bit like the stator vanes in a gas turbine engine

Edit

just thought of an example. the vanes on the delta focused flow fans. just with a pump impellor the flow is coming from the centre (Radial not axial) so the vanes are around the outside of the impeller

Ewan · 12-02-2003, 09:51 AM

In order to tackle the problem one has to have an understanding of how centrifugal pumps work and why they are designed the way they are.
Going by the name, a centrifugal pump works by applying a centrifugal force to the water. i.e water at the edge of the impeller has a centrifugal force which flings it ******ds away from the impeller. Since the pump is full of water, water is centrifugally forced towards the edge of the pump casing which in turn forces the water out of the outlet.
Centrifugal force is proportional to v^2/r where v is velocity and r is the radius. Since v in turn is proportional to rotation speed (W) and radius (the larger the radius the higher the speed) this boils down to centrifugal force being proportional to W^2*r.
Thus as radius increases, force increases linearly.
In other words a large diameter wheel will produce the most force at the tip of the impeller which in turn will cause a higher pressure at the outer pump casing and outlet.

Why the closed impeller?
If you have an open impeller then the water can leave the impeller and loose it's centrifugal force and thus it's pressure. Having a closed impeller makes sure that all the water that enters the pump leaves the pump at the outer edge of the impeller and thus has the most force (and resulting pressure). To maximise pressure you have to make sure the water stays in the impeller all the way.
Thus you should be able to increase the pressure of the pump by closing the impeller in the manner that bigben2K describes.
However when you have a closed impeller you don't ahev a lot of room for the water to flow whihc will limit how much they can pump. You notice how thin the gap is in the closed impeller for the water to pass through? That's why the max flow is so much lower.
High flow pumps have an open design with plenty of room for water to move through the pump. If water slips of the edge of the impeller, that's OK because in the short time it was in the impeller it may have developed enough pressure anyway. In this case, the aim is to simply get the water in and out.

But can't we have it both ways? Have high flow and high pressure by having a closed impeller with a much larger gap instead of the realy thin gap? Yes you can. Increasing the gap will allow more water into the impeller without realy negatively effecting the pressure since the impeller radius is unchanged.
But now the problem is that you've got tons of flow and heaps of pressure. Can you're little motor handle the extra load? Maybe.

As has been pointed out, the AC motor outputs at a constant rotation speed due to the frequency of the AC it's being fed (it rotates at the same frequency as the AC, thus 50Hz will yeild 3000 rpm). If it can't keep up, it will simply stall. So an AC powered pump has to be designed in such a way that the flow through the pump will never overwork it. This means that more often than not, a closed impeller type pump has already been sized in such way as to minimise this risk. You could try to improve upon it but you risk overloading the pump motor.

As one could guess, increasing rotational speed would also increase both flow and pressure, but then you have to change AC frequency which ain't so easy. Note that with a DC motor the problem isn't as bad. It will simply slow down if the load increases. Thus, you could overclock a DC pump by overvolting it. You might burn out your DC motor this way though.

correct me if I'm wrong.

fhorst · 12-02-2003, 10:16 AM

To me it looks like you are 100% right!

as all I need it pressure, and not that much flow, i could play a bit with the opening in the closed impellor. Start small, and then open it untill it's to much for the pump to handle.

Can you hear it if the motor stops spinning?

On the otherhand, I can not make my impellor bigger, so just a plate might do the trick. My pump can now get about 1.5 meters, at 700l flow (I have 2 of them in series) what would the max benefit be? Or should I just buy me 2 MCP600......

BillA · 12-02-2003, 10:27 AM

how do you intend to dynamically balance an impeller that spins at ~3000 rpm ?
(the MCP600 is 3500 rpm @ 12.0 VDC)

and if it is not balanced your bearing is gone, quickly !

I'm having real difficulty understanding the need/benifit of so much pressure/flow

fhorst · 12-02-2003, 06:35 PM

LOL Bill, why do you think you are selling the MCP600?

That's why................

And for me... I have 4 radiators (want to go as close as possible to ambient temps - I know the last 2 rads will most likely only give me an extra 3 degrees in temp drop, but it works for me..)

WW CPU, currently Inovatec GPU (thinking on changing) Dangerden Z block for the North bridge, 2 koolance HDD cool blocks (I'm gone make me some other ones, side cooling with 1/2" tubing the koolance are 6mm)
Got 2 other HDD's to cool (total 4), I want to cool the south bridge, mosfets etc. on my mobo also. If it is possible also my memory.

After I fried my antec PSU (150 euros) with a WC setup, I decided not to water-cool that one anymore but got me a nice silent Nexus PSU.

If it's gets hot enough to need cooling fins or a fan, I want to see if I can water-cool it. Why? Well, we all need a hobby ;-)

Currently I have the Asus IC7-Max3, it has something called OTES, with a fan that makes noise. No way to water cool that. ;-( So I’m looking for a other
Mobo...

Balancing the impeller will be a pain, so I gues I'll be looking for closed impellors in the right diameter and connection. (or buy me 2 MCP600.... If the problems are solved

Ewan · 12-03-2003, 03:28 AM

Quote:

as all I need it pressure, and not that much flow, i could play a bit with the opening in the closed impellor. Start small, and then open it untill it's to much for the pump to handle.

Do you already have a closed impeller?
The price difference between open and closed impeller pumps may be due to closed impellers simply being more expensive to make. An open impeller is a simpler construction.
Though it would be interesting to try an impeller converson, I think just buying a higher pressure pump will be the cheapest option in the long run.

Levish · 12-03-2003, 10:26 AM

Quote:

Originally posted by fhorst
Currently I have the Asus IC7-Max3, it has something called OTES, with a fan that makes noise. No way to water cool that. ;-( So I’m looking for a other
Mobo...

The Abit Max-3 OTES cooling can be removed and if you've got a bit of fabbing skill you can probably rig up something like this

http://www.overclockers.com/tips1104/

mad mikee · 12-03-2003, 12:29 PM

I feel a new project appearing during winter vacation.....

fhorst · 12-03-2003, 05:19 PM

Quote:

Originally posted by Levish
The Abit Max-3 OTES cooling can be removed and if you've got a bit of fabbing skill you can probably rig up something like this

http://www.overclockers.com/tips1104/

I have been thinking about that, but with the OTES off, I could not see how I can manage.

http://members.chello.nl/l.vanderhorst1/otes1.jpg

and also http://www.thetechzone.com/computers.php?i=281&p=1

this gives you an idea..... I just can't see how I can WC that....

RoboTech · 12-03-2003, 06:21 PM

Hey fhorst,

IMHO the old saying “you get what you pay for” generally holds true with PC water-cooling pumps.

You have to remember that the typical pond-pump is a culmination of design compromises. As has already been stated, most AC pond pumps use 2-pole AC synchronous motors (easy to water-proof => submersible, and inexpensive to manufacture) which will randomly rotate CW or CCW each time they start. Because of this, the pumps discharge port is always perpendicular to the axis of rotation (Eheim, Danner, Via, etc). The impeller in DC motor driven pumps and more expensive AC induction motor driven pumps only rotate in one direction. The discharge port on these pumps is mounted tangentially and captures both the centripetal AND inertial (rotating) forces of the moving water (Iwaki, Swiftech), which results in higher discharge head.

Pond-pumps are intentionally designed with loose clearances between the impeller and case so they can ingest debris and small pieces of gravel without jamming. (Looser tollerances also makes them less expensive to manufacture.) These large clearances lead to internal leakage - water flows from high-press discharge area back to low press suction area INSIDE the pump (commonly referred to as recirculation). This severely limits the pumps ability to generate much head. Apply a little resistance to the flow of water exiting the pump and the water will just quit flowing out, and instead will recirculate internally back to the eye of the impeller.

You mentioned in your initial post…

Quote: “As far as I can see, when a pump comes to its max head, the pressure is to high for the impeller to turn around… it first will "slip", missing a few pole changes and finally will stand still. When the pressure drops the impeller starts moving again…”

The impeller doesn’t actually stop turning – it may slip a little, but it doesn’t stop. The water just short-circuits inside the pump instead of flowing externally where it’s needed.

High-pressure pumps typically have tighter clearances between the impeller and casing to minimize internal leakage. This requires higher tolerances, which in turn costs more to manufacture.

High-pressure pumps may also incorporate a “closed impeller” like you noted on some of the Iwakis and the new Swifty. Enclosing the flow of water thru the impeller can add efficiency and helps minimize recirculation.

Depending on what kind of pump you decide to modify... one mod that has been successfully demontrated here at SC.c is close off a pond pump's axial discharge port and replace it with a tangential port...

(search "Volenti - Auto reversing pump?") This can significantly increase the pumps discharge head - like a cow pissin' on a flat rock - but ONLY when the impeller turns in the right direction!

One last thought... impellers and pump casings work together so if you modify one (like the impeller) then the other part should also be considered. Have fun and good luck...

Levish · 12-04-2003, 10:33 AM

Quote:

Originally posted by fhorst
I have been thinking about that, but with the OTES off, I could not see how I can manage.

http://members.chello.nl/l.vanderhorst1/otes1.jpg

and also http://www.thetechzone.com/computers.php?i=281&p=1

this gives you an idea..... I just can't see how I can WC that....

Yes I've been wondering the same but where there's a will there's a way

if a separate circuit is involved it shouldn't be too hard to fab up something where either the tops would go up and over the Cap's and over to the other Mosfet's ... and using something with thin tubing should make things easy since going around them isn't much of a option in a few of the cases

_
M / \ M
__| |__

Maybe the Epox or Asus mobo's are a better option for w/c

fhorst · 12-04-2003, 06:24 PM

the only positive side from the Otes, is that there are some mouning points. That could help with a waterblock design.

My designing skills are for waterblocks low, and so are my welding skills :-(

Any of teh designers out there who want to give this a try?

Cathar · 12-04-2003, 11:25 PM

Quote:

Originally posted by WAJ_UK
You can get turbine style impellers where there are stator vanes in the housing around the outside of the impeller. They curve the opposite way to the blades on the impeller and expand towards the outside of the housing slowing the flow down thus increasing the flow pressure. Sorry no links just heard about it somewhere. I don't know if there are any pumps around with housings like this. They work a bit like the stator vanes in a gas turbine engine

Edit

just thought of an example. the vanes on the delta focused flow fans. just with a pump impellor the flow is coming from the centre (Radial not axial) so the vanes are around the outside of the impeller

Yes, I've been looking out for pumps like this. According to what I've been reading, they are more efficient than centrifugal style pumps for the same power input, even for the closed impeller designs. Dramatically more pressure head than even a closed impellor cent. pump at the same RPM, but for quite a deal less power consumption.

More information can be found here:

http://www.animatedsoftware.com/pumpglos/turbinep.htm

Here's a manufacturer of turbine pumps in the USA:

http://www.albany-pump.com/pumps/turbine-pumps.html

Further information here:

http://www.dynafloweng.com/regenturbinepumps.html

Something like the MT-2003 from that last link would make a nice WC pump.

12-01-2003, 01:37 PM	#5
#Rotor Cooling Savant Join Date: Feb 2002 Location: Dione, sector 4s1256 Posts: 852	the way I understand it.... head it a result of a couple of things, or characteristics of the pump. 1: tolerance in the impeller casing, IE. how much space is there for liquid to squeeze by the impeller. 2: The diameter of the impeller. 3: rotational Speed #1 is obvious, so is #3 #2 includes a number of other factors too... Such as the power/torque of the motor. As the diameter of the impeller increases, the amount of kinetic energy transferred to the liquid by the impeller, as a result of the centrifugal force exerted on it as it goes through the pump, increases dramatically. This in turn requires more power from the motor (nothing is for free).... Torque in Particular becomes an issue, as the diameter of the impeller increases. I think of it this way, you have a ball and attached to it, is a piece of rope. You want to through this ball as high up into the air as you can.... What will give you the best result.... holding the rope at 10% , 30% 70% or 100% of it's length, while you swing it..... by the way.... this only pertains to centrifugal pumps.... __________________ There is no Spoon....

12-01-2003, 04:40 PM	#9
Levish Cooling Neophyte Join Date: Nov 2003 Location: NYC Posts: 19	I'm not the foremost expert on pumps but aren't fountain pumps working against a high head? Perhaps looking at the design of those (or maybe even using one instead of the aquarium style pump) would be appropriate if higher head was desired? __________________ 2.4c @ 274mhz fsb / Abit Ic7-Max3 / OCZ Permier PC4200 mem / Radeon 9800 Pro

12-01-2003, 05:04 PM	#12
bigben2k Responsible for 2% of all the posts here. Join Date: May 2002 Location: Texas, U.S.A. Posts: 8,302	Yes, I know that the closed impeller doesn't have curved "fins" in it. Something else to try. Like I said, get a few spares, and experiment. __________________ WBTA (Water Block Testing Alliance) Nordic Hardware WC 101

12-02-2003, 06:58 AM	#13
WAJ_UK Cooling Savant Join Date: Oct 2003 Location: Sussex Posts: 109	You can get turbine style impellers where there are stator vanes in the housing around the outside of the impeller. They curve the opposite way to the blades on the impeller and expand towards the outside of the housing slowing the flow down thus increasing the flow pressure. Sorry no links just heard about it somewhere. I don't know if there are any pumps around with housings like this. They work a bit like the stator vanes in a gas turbine engine Edit just thought of an example. the vanes on the delta focused flow fans. just with a pump impellor the flow is coming from the centre (Radial not axial) so the vanes are around the outside of the impeller Last edited by WAJ_UK; 12-02-2003 at 10:00 AM.

12-02-2003, 10:16 AM	#15
fhorst Cooling Savant Join Date: Oct 2003 Location: Almere, The Netherlands (Europe) Posts: 262	To me it looks like you are 100% right! as all I need it pressure, and not that much flow, i could play a bit with the opening in the closed impellor. Start small, and then open it untill it's to much for the pump to handle. Can you hear it if the motor stops spinning? On the otherhand, I can not make my impellor bigger, so just a plate might do the trick. My pump can now get about 1.5 meters, at 700l flow (I have 2 of them in series) what would the max benefit be? Or should I just buy me 2 MCP600...... __________________ If it get's hot, it needs to be watercooled!

12-01-2003, 04:04 AM	#1
fhorst Cooling Savant Join Date: Oct 2003 Location: Almere, The Netherlands (Europe) Posts: 262	Pump mod for High Head I want to start a discussion about pumps. Most of us use aquarium pumps, which are made for "normal" usage. A few of us bought an expensive Iwaki pump, which is made for High Head. There is a nice graph, where all the pumps are http://www.procooling.com/~phaestus/temp/flowcurves.jpg and from that graph, we can see that the MCP600 is "the best". I have not seen a discussion on why that one is the best, or why one pump is better than the other one. First some theory: The function of the pump's motor is simple: outside magnet and inside magnet. The outside magnet will due to the power (220v) become a north or a south pole magnet. The inside magnet will move due to the pole change. (50 or 60 Hz) Or as Iwaki tells it: "The drive magnet is magnetized barium ferrite for 4-, 6- or 8 poles. It rotates the impeller by using the magnetic field from the outside of the rear casting No brushes, so nice silent, and long lasting. Downside is that it is a weak motor. All motors seem to run at about 3000 Rpm. The motor is connected to the impeller: So, what makes the head? And why? As far as I can see, when a pump comes to its max head, the pressure is to high for the impeller to turn around. Actually, as I stated, the motor is weak, and it first will "slip", missing a few pole changes and finally will stand still. (With a bigger motor, having bigger magnets, this will take longer, and therefore give more head) When the pressure drops the impeller starts moving again, till the pressure is again at the top. This goes fast, so you won't even notice. Except that there is less water flow. From the Iwaki website, we can see that the pump motors stay the same, but they change the impeller for “Normal usage” high Head, or High flow. This got me thinking: what if we modify the impeller of our “Normal usage/High flow” pumps? Simply look at the High head design. The impeller wheel is a bit bigger, and has no fins. The impeller housing has changed also, but still…… If my theory is correct, all we need to do is to take the impeller, and take of the outside of the impeller fins, (cutting it diagonal). This will kill our free flow rating, but it should improve our head a lot. And for water cooling, we are not interested in free flow, but in head… That’s why the 600 l/h MCP600 pump out performs a 1400 l/h pump. We are having only a 5 l/m in our setup.. (300 l/h) Please share your ideas and theories. __________________ If it get's hot, it needs to be watercooled! Last edited by fhorst; 12-01-2003 at 04:10 AM.

12-01-2003, 07:09 AM	#2
bigben2k Responsible for 2% of all the posts here. Join Date: May 2002 Location: Texas, U.S.A. Posts: 8,302	From what I've seen discussed, there are two things that you can try immediately, to get more head: 1-restrict the outlet 2-given an open impeller, cap it with a doughnut shaped plate. Here's the impeller to my Little Giant 2-MDQ-SC: __________________ WBTA (Water Block Testing Alliance) Nordic Hardware WC 101

12-01-2003, 07:33 AM	#3
fhorst Cooling Savant Join Date: Oct 2003 Location: Almere, The Netherlands (Europe) Posts: 262	Thanks BB2K for the fast reply. I'm reading some other posts also, and there is a lot to discuss! I'm lost with "2-given an open impeller, cap it with a doughnut shaped plate." = flat plate with hole in it? or realy capsulate it..? the restriction of the outlet is what we all do with our nice waterblocks... it looks like that a pump with a head of about 4 meters giving at that head a flow of about 400/500 l/h is the best pump to get. and silent, and small, and low wattage...... and at a normal price ;-) That pump just is not there.... (for what I can see) The "cheap" ones all use the "wrong" impellor. I liked this post: making a 12v pump... And that one also sure makes you wonder. one watt of power, 170 l/h with max 20/30 cm head... __________________ If it get's hot, it needs to be watercooled!

12-01-2003, 01:07 PM	#4
BillA CoolingWorks Tech Guy Formerly "Unregistered" Join Date: Dec 2000 Location: Posts: 2,371.493,106 Posts: 4,440	the MCP600 has a closed impeller per the 'high head type' shown in the Iwaki dwgs - and that is why it is 'better' (for WCing applications) understand that cheap AC motors can, and will, start and rotate in either direction this is why aquarium pumps have straight vanes, reducing their efficiency even more you will not have any 'luck' converting an open impeller to a closed one no free lunch

12-01-2003, 02:49 PM	#6
Blackeagle Thermophile Join Date: Oct 2002 Location: U.S.A = Michigan Posts: 1,243	This topic brings us to the question of, 'How would the Iwaki pumps have compared to the Swiftech 600? I'd have loved to see a MD10, MD15 & a MD20rlzt pumps included in the testing. Even the MD10 should, on paper anyway, offer more system flow than the Swifty 600. I will grant you the higher cost of the Iwaki's will turn many water coolers away, but they are supposed to be the best. And the three Iwaki's compared to the Swifty 600 would have also shown which of them would be best bang for the buck. One of the biggest advantages of the Swifty pump is that it hooks up to the PSU.

12-01-2003, 04:03 PM	#8
BillA CoolingWorks Tech Guy Formerly "Unregistered" Join Date: Dec 2000 Location: Posts: 2,371.493,106 Posts: 4,440	Ben you are an unending source of bad info do you actually know what a closed impeller is ? so then how do your instructions yield such ? and you are the WBTA, god spare us from your ignorance

12-02-2003, 09:51 AM	#14
Ewan Cooling Neophyte Join Date: Jun 2003 Location: Sweden Posts: 30	In order to tackle the problem one has to have an understanding of how centrifugal pumps work and why they are designed the way they are. Going by the name, a centrifugal pump works by applying a centrifugal force to the water. i.e water at the edge of the impeller has a centrifugal force which flings it *****ds away from the impeller. Since the pump is full of water, water is centrifugally forced towards the edge of the pump casing which in turn forces the water out of the outlet. Centrifugal force is proportional to v^2/r where v is velocity and r is the radius. Since v in turn is proportional to rotation speed (W) and radius (the larger the radius the higher the speed) this boils down to centrifugal force being proportional to W^2r. Thus as radius increases, force increases linearly. In other words a large diameter wheel will produce the most force at the tip of the impeller which in turn will cause a higher pressure at the outer pump casing and outlet. Why the closed impeller? If you have an open impeller then the water can leave the impeller and loose it's centrifugal force and thus it's pressure. Having a closed impeller makes sure that all the water that enters the pump leaves the pump at the outer edge of the impeller and thus has the most force (and resulting pressure). To maximise pressure you have to make sure the water stays in the impeller all the way. Thus you should be able to increase the pressure of the pump by closing the impeller in the manner that bigben2K describes. However when you have a closed impeller you don't ahev a lot of room for the water to flow whihc will limit how much they can pump. You notice how thin the gap is in the closed impeller for the water to pass through? That's why the max flow is so much lower. High flow pumps have an open design with plenty of room for water to move through the pump. If water slips of the edge of the impeller, that's OK because in the short time it was in the impeller it may have developed enough pressure anyway. In this case, the aim is to simply get the water in and out. But can't we have it both ways? Have high flow and high pressure by having a closed impeller with a much larger gap instead of the realy thin gap? Yes you can. Increasing the gap will allow more water into the impeller without realy negatively effecting the pressure since the impeller radius is unchanged. But now the problem is that you've got tons of flow and heaps of pressure. Can you're little motor handle the extra load? Maybe. As has been pointed out, the AC motor outputs at a constant rotation speed due to the frequency of the AC it's being fed (it rotates at the same frequency as the AC, thus 50Hz will yeild 3000 rpm). If it can't keep up, it will simply stall. So an AC powered pump has to be designed in such a way that the flow through the pump will never overwork it. This means that more often than not, a closed impeller type pump has already been sized in such way as to minimise this risk. You could try to improve upon it but you risk overloading the pump motor. As one could guess, increasing rotational speed would also increase both flow and pressure, but then you have to change AC frequency which ain't so easy. Note that with a DC motor the problem isn't as bad. It will simply slow down if the load increases. Thus, you could overclock a DC pump by overvolting it. You might burn out your DC motor this way though. correct me if I'm wrong.

12-02-2003, 10:27 AM	#16
BillA CoolingWorks Tech Guy Formerly "Unregistered" Join Date: Dec 2000 Location: Posts: 2,371.493,106 Posts: 4,440	how do you intend to dynamically balance an impeller that spins at ~3000 rpm ? (the MCP600 is 3500 rpm @ 12.0 VDC) and if it is not balanced your bearing is gone, quickly ! I'm having real difficulty understanding the need/benifit of so much pressure/flow

12-02-2003, 06:35 PM	#17
fhorst Cooling Savant Join Date: Oct 2003 Location: Almere, The Netherlands (Europe) Posts: 262	LOL Bill, why do you think you are selling the MCP600? That's why................ And for me... I have 4 radiators (want to go as close as possible to ambient temps - I know the last 2 rads will most likely only give me an extra 3 degrees in temp drop, but it works for me..) WW CPU, currently Inovatec GPU (thinking on changing) Dangerden Z block for the North bridge, 2 koolance HDD cool blocks (I'm gone make me some other ones, side cooling with 1/2" tubing the koolance are 6mm) Got 2 other HDD's to cool (total 4), I want to cool the south bridge, mosfets etc. on my mobo also. If it is possible also my memory. After I fried my antec PSU (150 euros) with a WC setup, I decided not to water-cool that one anymore but got me a nice silent Nexus PSU. If it's gets hot enough to need cooling fins or a fan, I want to see if I can water-cool it. Why? Well, we all need a hobby ;-) Currently I have the Asus IC7-Max3, it has something called OTES, with a fan that makes noise. No way to water cool that. ;-( So I’m looking for a other Mobo... Balancing the impeller will be a pain, so I gues I'll be looking for closed impellors in the right diameter and connection. (or buy me 2 MCP600.... If the problems are solved __________________ If it get's hot, it needs to be watercooled!

12-03-2003, 12:29 PM	#20
mad mikee Cooling Savant Join Date: Oct 2001 Location: US of A!!!! Posts: 146	Just the thing to watercool!!! I feel a new project appearing during winter vacation..... __________________ His PCV2000 DFI ULTRA D (SLI Modded just because) 4200 x2 @ ~ 10 x 260 /1.49 Vcore , G5 + PA160 + 150mm Papst@5V + Maze 4 GPU + mcp350. Temp 36-37(load-2 x D20L) 2x1GB Ballistix PC4000, x800xl 256m, 74G rap + 160G Hitachi SATA , 21" mit/nec 1920x1440 Crt + 19 Neovo F-419 1280x1024 LCD, OCZ Powerstream 600 Hers: (my beloved who lets me do all this) LianLi PCV1000, SD3700 @ 9x300, 1GB GBLA, 250+ 120 SATA, X700pro, 19" Neovo F-419 LCD, another OCZ 600

12-03-2003, 06:21 PM	#22
RoboTech Cooling Savant Join Date: Sep 2002 Location: Cincinnati, OH Posts: 229	Hey fhorst, IMHO the old saying “you get what you pay for” generally holds true with PC water-cooling pumps. You have to remember that the typical pond-pump is a culmination of design compromises. As has already been stated, most AC pond pumps use 2-pole AC synchronous motors (easy to water-proof => submersible, and inexpensive to manufacture) which will randomly rotate CW or CCW each time they start. Because of this, the pumps discharge port is always perpendicular to the axis of rotation (Eheim, Danner, Via, etc). The impeller in DC motor driven pumps and more expensive AC induction motor driven pumps only rotate in one direction. The discharge port on these pumps is mounted tangentially and captures both the centripetal AND inertial (rotating) forces of the moving water (Iwaki, Swiftech), which results in higher discharge head. Pond-pumps are intentionally designed with loose clearances between the impeller and case so they can ingest debris and small pieces of gravel without jamming. (Looser tollerances also makes them less expensive to manufacture.) These large clearances lead to internal leakage - water flows from high-press discharge area back to low press suction area INSIDE the pump (commonly referred to as recirculation). This severely limits the pumps ability to generate much head. Apply a little resistance to the flow of water exiting the pump and the water will just quit flowing out, and instead will recirculate internally back to the eye of the impeller. You mentioned in your initial post… Quote: “As far as I can see, when a pump comes to its max head, the pressure is to high for the impeller to turn around… it first will "slip", missing a few pole changes and finally will stand still. When the pressure drops the impeller starts moving again…” The impeller doesn’t actually stop turning – it may slip a little, but it doesn’t stop. The water just short-circuits inside the pump instead of flowing externally where it’s needed. High-pressure pumps typically have tighter clearances between the impeller and casing to minimize internal leakage. This requires higher tolerances, which in turn costs more to manufacture. High-pressure pumps may also incorporate a “closed impeller” like you noted on some of the Iwakis and the new Swifty. Enclosing the flow of water thru the impeller can add efficiency and helps minimize recirculation. Depending on what kind of pump you decide to modify... one mod that has been successfully demontrated here at SC.c is close off a pond pump's axial discharge port and replace it with a tangential port... (search "Volenti - Auto reversing pump?") This can significantly increase the pumps discharge head - like a cow pissin' on a flat rock - but ONLY when the impeller turns in the right direction! One last thought... impellers and pump casings work together so if you modify one (like the impeller) then the other part should also be considered. Have fun and good luck...

12-04-2003, 06:24 PM	#24
fhorst Cooling Savant Join Date: Oct 2003 Location: Almere, The Netherlands (Europe) Posts: 262	the only positive side from the Otes, is that there are some mouning points. That could help with a waterblock design. My designing skills are for waterblocks low, and so are my welding skills :-( Any of teh designers out there who want to give this a try? __________________ If it get's hot, it needs to be watercooled!

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)