Centrifugal Pump Impeller

[Jmccrack]

Has anyone had any experience in building a
Centrifugal Pump Impeller like the one J galba uses on his beam engine?

 

[gbritnell]

All of my I.C. engines have centrifugal water pumps on them. Do you have a specific question about them?
gbritnell

 

[Jmccrack]

I was wondering how to setup to machine one

 

[Kaleb]

I'm also in the process of building J. Galba's beam engine, and had been wondering whether I could get by with a straight-blade impeller. I've seen such a thing in a full-sized pump I've got laying around(albeit with one blade missing!) If I could have straight blades, that would make my life a lot easier since I don't think I'm up to making Jurgen's curved blades without a milling machine.

 

[vederstein]

This thread illustrates the building of a steam turbine and centrifugal pump. It may give you some ideas.

http://www.homemodelenginemachinist.com/showthread.php?t=22430

...Ved.

 

[gbritnell]

The physics behind the operation of a centrifugal pump are that the liquid is drawn into the center of the impeller and through the angle of the fins on the impeller the liquid is pushed to the outside of the cavity and through the outlet port. It's not necessary to have curved fins but only that the fins be angled to the axis of the pump. When spinning this causes the liquid to slide along the blades. If the blades were at right angles to the axis they would just push the liquid around in a circle.
In any case a dividing tool is needed to index the hub so that the metal can be removed between the fins.
Attached is a sketch I did for the pump on my Holt engine.
gbritnell

View attachment WATER PUMP.pdf

 

[Jmccrack]

Ved
I see. I assume the impeller blade is milled just by offsetting the rotary table chuck.then indexing accordingly. Are there any calculations or just an eyeball set up for the first blade.then the rest follow.

 

[vederstein]

Ved
I see. I assume the impeller blade is milled just by offsetting the rotary table chuck.then indexing accordingly. Are there any calculations or just an eyeball set up for the first blade.then the rest follow.

I actually CNC milled the blades, but if I remember correctly, the design drawings illustrate how to set it up for manual milling.

...Ved.

 

[Charles Lamont]

Very small centrifugal pumps are appallingly inefficient (in the proper thermodynamic sense) even when well designed. At this scale the losses are mainly due to friction, and viscosity effects predominate. This means that, as gbritnell says, it is particulary important that the blades are swept backwards. (Radial blades would work to some extent, because of the centrifugal force on the liquid, but not well, especially at this size). It may help to imagine the pump working in a high vicosity fluid such as a heavy oil, or some kind of syrup.

Straight blades probably won't make a huge difference at this size. I would go for four blades, as being easy to mill with simple equipment, and I would make them tangential to the eye diameter. You could impove matters by then filing a bevel on the blades at both ends, on the leading side, almost to a sharp edge. This will approximate to a curved on the leading side of the blade and will assist the geometry at the eye as well as effectively increasing the back sweep at the periphery. I would prefer to see a proper 'snail shell' spiral volute housing, with only a small clearance between the outlet tongue and the impeller.

 

[barnesrickw]

Diaphragm pumps work well at small scale. Peristaltic do too, but it takes quite a bit of torque to squash the tube. My uncle made a linear peristaltic for some paint sprayer company that required less torque, but it's hard to figure out how it works.

 

[johnnyo]

When I used to work on my old boat more then fishing out of it I had the bilge pump quit on me again. I pulled up the panel and noticed the housing had seperated. Inside was a small 1" dia. Impeller with four offset straight blades. I remember thinking "thats it" this thing throws a full tube of water out the back. I needed a pump for my Holt engine so I went with this design. Centered in a close comcentric housing it still worked well. Programmed one with four curved blades and same dia. On Traxx Mill but didnt notice much difference. Still moved coolant through Holt passages. Use a close clearance at blade face cover to reduce recirculation. Gbritnels design is nice but you can step off four blades without indexers and still make it work. Nice pump design George.

 

[bobden72]

Hi I have made and use a small centrifugal pump on one of my engines to circulate water around the block. Nothing fancy is needed, straight blades in a well fitting housing and it works a treat. All fabricated out of brass and soft soldered together, then carefully turned to fit the housing. Don't get to complicated and back yourself in a corner.

 

[Jmccrack]

Thanks Guys
But I was wondering if there was a simple way with a milling machine and rotary head that one would be able to mill a six bladed impeller like Galba,s engine.You must be able to turn the blank on a lathe then offset the blank in a rotary able to mill the blades. If you can lay the blades out with a protractor you must be able to mill them.

 

[barnesrickw]

Most of the aquarium pumps I've seen have straight vanes. The simplest have a volume that is circular, but eccentric of the rotor. The vanes slide up and down in a slot being flung against the volume as they spin.

 

[Charles Lamont]

Most of the aquarium pumps I've seen have straight vanes. The simplest have a volume that is circular, but eccentric of the rotor. The vanes slide up and down in a slot being flung against the volume as they spin.

But that would be a vane pump, not a centrifugal one. A different animal entirely.

 

[Ken I]

This could get way technical - but for simplicity you don't have to look any further than the coolant pumps on virtually all car engines.
Appallingly inefficient but they get the job done.
I sell water cooled welding equipment and those pumps have straight radial vanes (not even offset or curved) and it does the job.
What do you need ? If its just to circulate coolant almost anything will work.

Regards,
Ken

 

[gbritnell]

The original posting was about a centrifugal pump. Yes it's true that straight fins will move liquid but they aren't as efficient as curved vanes/fins. With the curved shape the liquid slides along the curve to the outside of the cavity whereas with flat blades the liquid has a tendency to just be pushed along between the blades. Naturally centrifugal force will cause the liquid to be forced to the outside with both types.
https://search.yahoo.com/search;_yl...b-top-search&fr=hp-avast-s&type=avastbcl&fp=1

 

[Charles Lamont]

This could get way technical - but for simplicity you don't have to look any further than the coolant pumps on virtually all car engines.
Appallingly inefficient but they get the job done.
I sell water cooled welding equipment and those pumps have straight radial vanes (not even offset or curved) and it does the job.
What do you need ? If its just to circulate coolant almost anything will work.

Regards,
Ken

I spent a couple of years at work on centrifugal pumps. My boss on the project had come from a truck diesel manufacturer. He pointed out quite early on that (full size) engine coolant pumps actually work under quite difficult conditions, because the water they are handling is near boiling. The pump suction design has to be carried out carefully or water will vapourise near the inlet edges of the blades, and then, as the pressure increases, suddenly condense near the exit of the impeller, a highly undesirable phenomenon called cavitation - same as you get with marine propellers.

 

[Rustkolector]

Something else you might want to consider. Gear pumps are relatively simple and easy to make. They work especially well on non-pressurized model engine flows and temperatures. Jerry Howell's design is a great performing pump. Gear pumps move coolant with relatively little effort and volume can be designed into them accurately. Although they are thought of as positive displacement pumps by design, they are not normally built to tight tolerances for cooling service. They are also quite compact for the volume of coolant pumped.
Jeff