Digicam is back!
 

This is the current state of the pump. The cap was molded against the base for a perfect fit. I dremeled out a pocket for the outlet pipe, and fixed it with epoxy putty.

Well guess this info comes a bit late, but the fact is that default expoy doesnt tolerate water! It wont take any noteable damage at first. But in a while or two joints will start to fail. Epoxy decompose in contact with water. Alltrough I guess there is special epoxy which will tolerate water as well.

perhaps you can put a thin layer of silicone very thin layer to protect the epoxy if this is the case.. --Josh

This is the end...
 

Well, this experiment finished with a partial success only.

I've been trying several impeller variations, and found a strange problem: When the pump fhas too much flow restriction, it enters a pulsating cycle where it quickly stops and start spinning. It don't happens in free flow condition, but when I try a max head test, it always happens when it reach about 35 cm.

To avoid it, I had to reduce the impeller dimensions, up to a piont where it can't generate more flow than the critical level. With the straight radial blades configuration, I ended with something similar to the impeller os that water-X shit. With the two disk option, I had to reduce the diameter to 28 mm.

At first I thought it was something with the motor electronics, where it automatically cut off the power when there's too much resistance, but the strange thing is that this happened at very different rpm levels depending on the impeller shape.

At the end, the performance at the best "stable" configuration was 300 L/H and 32 cm head.

This can still be useful to move water on a secondary loop in the system, but definitely not enough for cpu cooling.

An interesting side effect: With straight blades, this made an excellent air blower!

I don't know an easy way to measure airflow cfm, but for pressure I sunk the hose in water, and the level inside it was 30 mm lower than outside. As I remember, most fans can do only small fractions of an inch of static pressure.

Here are a couple pics of the two initial impellers I made. That was the starting point, and then I was progressively reducing the dimensions with the lathe.

Common epoxy softens very quick with water, but I’ve made things with this stuff that have been more than 6 months underwater, and are as strong as the first day, with no color alteration.

A bonus pic: The pump working at free flow.

wanna send that to me .. =) --Josh

Re: This is the end...
 

That measure is based on a 1 square inch opening, I believe, so you can recalculate, if you know your tubing diameter.

I think you'll find that it's lower than spec.

Nice work. It could act as a backup, if it was possible to bypass it easily.

hey, all you need is a more powerful motor :)

Ben, the pressure shouldn't be dependent on the cross sectional area. If you used a larger tube it would take longer for the pressure to stabilize, but it wouldn't change the maximum value that the pump would produce.

or make a couple more and run them in series :D

Lol, but seriously:
Thats not bad!
What fan did the motor come off of?
I bet it didnt come off of a 60mm delta screamer, and that draws a whole ~4.5watts!
I mean a cheap VIA Aqua 1300 consumes 20 watts.

You gotta get you a nice big 120MM fan motor to really try this out on :) a whopping 6 watts!!!!! :)

For the size of the motor I'm impressed, it does'nt hit home until you see it in nicos paw how small it is!...

What did it sound like nico?, was it audible at all?...

Of course it wouldn't. I just observed Nico state that he gets a 30mm head, but didn't specify the diameter of the tube he used, to measure that 30mm.

30 mm is well over 1 inch, and as a fan it shouldn't be able to come close to that. Given that the fan was replaced with a paddlewheel, it should perform even worse.

I was merely pointing out that where fans are spec'd to say, less than 1/4" H2O, it is a measure that is based on a specific diameter, which I believe is 1 in. ^2.

Heck, with a tube small enough, I could state that my fan has a head of 1 meter!

I like the idea of using a fan motor as a pump. It's a very efficient design. The best option would have been to choose a 120mm fan as they rotate at a low rpm and could give you enough torque to move quite some water. ;)

Ben, pressure units are force/area, so area size is irrelevant. This was really a powerfull blower; but noisy at 9K rpm.

This was a 60 x 10 mm of intemediate power. It was rated at about 4 watts, but I don't know how precise that rating was.

Looking at Eheim pumps, they have a 4 watt model, the 1005 that makes 270 l/h and 75 cm head. So sure 4 watt can do more than 30 cm, but I don't know what makes that limit. Maybe head and torque are related in some way, and the low torque of the motor makes impossible to go higher.

It made some little noise, but it was from vibration of the poorly balanced impeller. I did'nt take too much care with that as it was a quick test.It was a bit tricky to center it with precision in the lathe, cause the soft plastic of the fan deforms under pressure.

the bearing was silent at UW speeds, so a carefull balance can make it really noiseless.

Maybe I'm just not making myself clear... or I'm being stubborn.

So what you're saying is that if the hose diameter was say, 3 feet, your pump could raise/lower the water by 30 mm?

That's what he's saying, and he's right.

Hum...

so if cross section is irrelevant, the water would be raised/lowered by 30 mm, no matter what the diameter...

...but if the diameter is say, 1 meter, the volume of water moved would be 0.024 cubic meters, which is 24 liters, which weighs 24 kg.

If the diameter was 2 cm, then the volume of water would be 0.0000094248 cubic meters, which is 0.009425 liters, which weighs 0.009425 kg, aka 9.4 grams.

Given that the fan will produce a fixed amount of force, the diameter very much makes a difference.

Am I wrong?

I think you're just having a temporary short in your brain, 'cause you know this already :)

Pressure is not dependent on volume. Your 100 PSI air pump will fill a truck tire to 100 PSI just like it will fill a bicycle tire to 100 PSI. Just takes a lot more pumping!

That little pump will fill the city pool up to 30mm deep. It may take a couple years, but it will do it. That is the nature of centrifugal pumps. These pumps maintain a constant pressure, regardless of flow rate. That's in direct contrast to displacement pumps, which move specific volumes of water regardless of pressure (to a point, of course.)

Oh yes! Thanks!

The city pool did it for me:D I always have a problem remembering that one.

So if the volume of water IE diamater is more , the psi doesn't increase? so 10 feet deep in the ocean is the same psi as 10 feet deep in a common pool , interesting.. Never gave that much thought. So I see what you're saying about filling the pool.. nice job breaking it down for us.. lol --Bikr

Thanks nicozeg and Graystar for chiming in ;)

Substitute force by pressure in that sentence and you're OK

Multiply pressure by area and you get force.

That's the key concept behind hidraulic machinery; Fluids transmit pressure, not force.

I tested this pump in my system today. I was a bit scared at he begining about overheating, but it did a good job!

Here are the results compared to my regular pump, a Shott 11.10.

first number is my 12V model

Pump power in watts:_________________4/12
In system waterflow (LPM)___________2.1/4.7
CPU-Water delta temp idle____________11/9
CPU-Water delta temp BurnK7_________18/16

This is amazing! despite the huge difference in power and flowrate my temps are only 2ºC worst; not bad at all!

more important for me is noise. My other pump makes a slightly anoying rattle some days (I guess it depends on the turning direction it choose on startup), and this one was impossible to hear. :)

I found this quite a success, given that I started testing ready to unplug the power in case of high temps. :cool:

nice work! now i wanna see you use a beefier 120mm fan motor :)

Perhaps a blower instead of a axial fan will give you more torque. It might also be a lot easier to convert.

What about something like this: http://comairrotron.com/pdfs/DD5236Series.PDF

How long did you give it to equalise?, my rig takes forever to reach the max temp(where it falls and rises for a while then stabalises), up to an hour...

I'm not too surprised though, I ran my PC for hours without the pump running before the Alarm sounds at 45DegC. The radiator was higher than the CPU so convection happened easily though...

My cpu to ambient temps take several hours to stabilize, and sure at that time ambient temp has changed. I have an open 5 galon res that cools by evaporation instead of a rad. However, I measured cpu to water temps, that takes about 10 minutes and I waited 15.

I Dont want to make a suicide test without pump, I'm prety sure that convection is going to have a hard work trying to move the water across 2 meters of hose and a fine mesh filter, that I use to keep the block free of debris.

you want a 120MM fan as a donation, to build another one?

Damn! I had decided to not make another one cause I have very little free time; but if you give me a push... :D