Impellor Question hmmmm?

lw395 said:
Yes, I realise that.
More blades should not shift more water, if the blades are sealing against the pump body properly.

The pump works because the ramp on the pump body reduces the volume on one side of the pump. So more water is carried forwards than backwards. Assuming there are always at least two blades in contact on both sides, adding more blades won't move any more water per rev.
It's just breaking the moving 'chambers' between impeller blades into smaller parts.
But if the sealing isn't perfect, more blades may help.

https://en.wikipedia.org/wiki/Flexible_impeller
Look at the moving diagram and imagine doubling the number of blades.
Twice as many pulses of water ever rev, but each is half the size.
Maybe there is a small effect, the blade thickness reduces the volume carried back more than it reduces the volume carries forwards?
Or maybe if the pump isn't well designed to work with the smaller number of blades?

More blades very likely to seal better?
Click to expand...

Thanks for the wiki link, I think I am now convinced that marginal performance difference and seal wear aside, there should be no real difference in pump output.
My mate Barry will be pleased :) Being technically minded, I feel I should have been able to work this one out for myself. Why it wasn't obvious to me, I don't know.
I blame the Adnams. :)
 
Habebty said:
Thanks for the wiki link, I think I am now convinced that marginal performance difference and seal wear aside, there should be no real difference in pump output.
My mate Barry will be pleased :) Being technically minded, I feel I should have been able to work this one out for myself. Why it wasn't obvious to me, I don't know.
I blame the Adnams. :)
Click to expand...

I understand Barry had back room support in his tech advise capacity :confused:
 
Last edited:
Habebty said:
Thanks for the wiki link, I think I am now convinced that marginal performance difference and seal wear aside, there should be no real difference in pump output.
My mate Barry will be pleased :) Being technically minded, I feel I should have been able to work this one out for myself. Why it wasn't obvious to me, I don't know.
I blame the Adnams. :)
Click to expand...

Nitrile impeller for Adnams?
 
Interesting discussion.

As I see it, at zero blades no water gets shifted. With one blade there might be a small amount of water gets pushed through but not enough to do the job. Two blades should increase the flow. At the other end of the scale obviously with an infinite number no water will get through.

At the low end of the count the blades would need to be quite thick as each one has a large volume of water to "push" through and so would need to be quite thick and inflexible so more likely to wear and failure. As you increase the number of vanes there is less pressure and so they can be thinner and more flexible and likely to last longer but cost more to make.

So we have a curve where at each end no water is pumped and in between some volume is pumped. My guess is that the volume of water pumped is fairly flat in the 6-12 vanes range and the trade-off is reliability and cost. So any experiment would also need to look at the manufacturers MTBF and costs. I guess manufacturers don't publish MTBFs so that makes any conclusions harder.

This needs a bit of work looking at tech specs.
 
SimonFa said:
At the low end of the count the blades would need to be quite thick as each one has a large volume of water to "push" through and so would need to be quite thick and inflexible so more likely to wear and failure. As you increase the number of vanes there is less pressure and so they can be thinner and more flexible and likely to last longer but cost more to make.
Click to expand...

I don't see the link between volume and pressure. The pressure difference between input and output seems to me independent of the number of vanes.

Unless there are several chambers in the flow path, so the pressure difference can be spread over the multiple vanes?

Mike.
 
SimonFa said:
Interesting discussion.

As I see it, at zero blades no water gets shifted. With one blade there might be a small amount of water gets pushed through but not enough to do the job. Two blades should increase the flow. At the other end of the scale obviously with an infinite number no water will get through.

At the low end of the count the blades would need to be quite thick as each one has a large volume of water to "push" through and so would need to be quite thick and inflexible so more likely to wear and failure. As you increase the number of vanes there is less pressure and so they can be thinner and more flexible and likely to last longer but cost more to make.

So we have a curve where at each end no water is pumped and in between some volume is pumped. My guess is that the volume of water pumped is fairly flat in the 6-12 vanes range and the trade-off is reliability and cost. So any experiment would also need to look at the manufacturers MTBF and costs. I guess manufacturers don't publish MTBFs so that makes any conclusions harder.

This needs a bit of work looking at tech specs.
Click to expand...

Even with 1 vane the same volume would be moved as if it had 6 vanes ( less the volume of 5 vanes).
pressure doesnt come into this question the OP posed
 
My 5ml worth....

A one blade pump would be very inefficient since the water is free to return to the inlet from the outlet.

To prevent that, there must always be at least one blade between the inlet and outlet. That means there must be a minimum of 4 blades. So the smaller the angle between the inlet and outlet, the more blades will be required to ensure there is always a seal between them.

After sealing is achieved, adding additional blades will;
- reduce the total volume and so reduce the output as others have pointed out.
- increase the blade to chamber seal efficiency. That would increase the volume pumped if the fewer blade seal is poor (as the impeller wears for example).
 
Last edited:
mjcoon said:
I don't see the link between volume and pressure. The pressure difference between input and output seems to me independent of the number of vanes.

Unless there are several chambers in the flow path, so the pressure difference can be spread over the multiple vanes?

Mike.
Click to expand...

Possibly the pressure is spread over several vanes.
 
Surely largely irrelevant since the thermostat will control the amount of water actually cooling the engine, unless the pump is designed with no margin of error of course.
 
If it helps, my Beta 20 with 10 blade Jabsco (20mm inlet) kicks out a much better flow of water than my 6 blade Yanmar's does (15mm inlet), and that's both at about 850-900prm with a 2 inch exhaust hose. That's both with new impellers.
 
For an elaboration of the bucket test, a friend of mine uses the '15' rule. If you run the engine at 15 hundred RPM, the bucket should fill in 15 seconds.
 
Aren't the vanes spaced out to exactly match the widths of the inlet and outlet gaps? This then determines the optimum number of vanes for that particular pump. I'm just guessing here.
 
Dipper said:
Aren't the vanes spaced out to exactly match the widths of the inlet and outlet gaps? This then determines the optimum number of vanes for that particular pump. I'm just guessing here.
Click to expand...

No of vane can make no difference at all,its down to the pump volume on one revolution X the RPM of the punp shaft
 
I recall briefly considering the matter on finding (with a similar engine to the OP’s) that I had both 6 and 8 blade impellors in my spares stock. The absence of Adnams (on that occasion) may have influenced my thought that it probably didn’t make a lot of difference, but I shall await a final expert view - or better, empirical evidence - with interest.
 
A couple of years ago I had to replace by 8 blade Jabsco with a 6 blade Johnson(supplied by Beta) and I also wondered whether it would pump more or less but have not noticed any obvious difference. The Johnson was a reasonable price from memory, from Beta for a Beta 25hp.
 
You must log in or register to reply here.

Share this page

Top