For a cheap pump, i don't believe that the inlet/outlet ratio really matters!

The main problem with what I suggested is that there would be a huge loss of pressure/flow, because the centrifugal effect is effectively bypassed: these pumps work by throwing the water around, and as above, it would only be paddling the water around. Not good.

As for flow reversing, I think that as long as the flow rate is high enough, it shouldn't matter in which direction it's going. If the flow is too slow, then yes, there would be that problem in the order of the components.

I see,

what if we put inlet up and outlet down not inline like the pic, to get half the circle, maybe some more centrifugal effect and get at least 1/2 of it?

Same problem. It would only work if what you're talking about is actually 2 outlets, with a center inlet.

The water must be thrown around.

GREAT RESULTS!!
 

Adding the second outlet and simply combining their output with a Y adaptor yeilds excellent results;
http://users.bigpond.net.au/volenti/PM_overview.jpg

http://users.bigpond.net.au/volenti/PM_closeup.jpg

http://users.bigpond.net.au/volenti/...seup_angle.jpg

I have measured static head as well as just raw flow rate, should make the results a little more relavant.

un-modifed pump static head; 155cm (60.5'') (rated at 200cm)
modified pump static head single outlet; 198cm (77.2'')
modified pump static head dual combined outlet; 210cm (81.9'')

raw open flowrate is ~1100L/H, which from memory is slightly down on what the pump could do in stock condition, but that doesn't matter seeing as we now have more static head to play with now.

I also blocked the center inlet to see if it would pump this way (turning one of the outlets into an inlet) it does actually pump, but not with a lot of flow, or pressure,I didn't bother measuring it.

**************************************
WOW!, I really expected it to be less when they were combined, I think some of it is the long legnth of the two tubes before they meet, stops the flow going out one side and back in the other, if they were shorter I think there'd be problems.

Would you try them with very short tubes Volenti?, to see what happens?, Please huh? :D ...

hmm, I shortened the tubes as much as I could (about half) reduced the static head by about 10cm (3'') but I did use 90 degree elbows to shorten the tubes, probably where the head loss came from...

however something else I observed while "playing" with the pump, was that as you progressively restrict the inlet, the water flow in the tube from the "weaker" outlet slows, stops, then goes backwards, re-circulating back to the pump, it appears that this mod may only be suitable for in-res use only, not in-line.

Don't forget that inline rigs push the water into the pump inlet, I also wrote a bit about the inlet at OCAus that I did'nt do here...

I'd also be interseted in what happens to the weaker vent when you introduce pressure/restriction to the outlet, whether it stays balanced or starts to do what you just mentioned. How would you test for it though, fit some blocks and rads inline and see what comes out the end?, if it's still got good pressure all's OK...

If it buggers up there's always the ratchet method, did you try it?, did it work and flip the right way?, or did it keep stalling the wrong way?...

PS, a valve/flap at the Y would work as well, when alls well it would be 'central, if the weak flow gets to low the strong one will force the valve shut. you'd have to make the Y & valve yourself but I'm sure it's 'doable' :) . worth it too if that's what it takes...

PPS, you could just use 'one way' valves in the two tubes that lead to the Y fitting ;) , I don't know how restrictive they are though?. I'm sure we could make a 'passive' one that works of the water pressure according to flow direction, rather than a spring type one. would have near on zero restriction if made oversized bore wize...

If we assume the combined water flow to be 100% the weak outlet is about 30% and the strong one 70% (roughly judging by eye in free flowing operation) block the weak one and the strong one goes to 95%, block the strong one and the weak one goes to about 55%.
[edit] only under special conditions will the "weak" outlet begin to become an intake, the spinning water in the pump still wants to escape, and if the que at the "strong" intake is too long it will do the almost 120 degree turn (loose energy) and leave via the weak outlet.[/edit]

I'm not sure how I'm going to test the flow in-line, mabie use the T peice thing after the pump to check pressure? may work...

[doh another edit] yea I'll look at the flap valve if the in-line testing turns sour, I'll be using a particularly restrictive in-line set up (like my current high pressure direct die block) so if there's an issue I'll find it.

To check for adverse reactions under system pressure/restriction:

Try pinching off the weaker one, if flow improves it means water is flowing back throught it. if flow decreases alot you've pinched the good one, if it decreases a little you've pinched the bad one but all's OK cause it's still giving pressure.

Just make sure you put a good bit of restriction in the 'line' (block & rad, more preferably) before the exit where you see the flow/water come out.

***************************************

The only way to check inline is to watch the temps I think, or watch for cavitation, it should'nt make any difference between inline and submersed I think...
**Edited cause I'm an Idiot:drool: :D ...**

I did that with the static head testing, both combined is about 3'' of head better than the "strong" outlet with the weak outlet blocked, but I will re-test that in the in line testing.

next question?:D

We're gonna have to stop this 'editing' confusion :D...

Did you try a headtest with restriction in the line?, put a really restrictive block and rad in the line after the Y, then try the 'pinch tests' and see how high the fountain is :) ...

Then when you try it in a 'closed loop' just watch for cavitation, if it's getting enough water pressure it'll be OK. I think it will as the output is pushing back into the inlet(I've heard it said it's better than submersed for flow because of this), as long as backpressure does'nt bugger things up...

Good luck dude!...

BAH! I hooked the modded pump up to my direct die watercooling, and as it fired up and the bubbles slowly cleared I could see a line of bubbles going back down the "weak" outlet tube into the pump (water circulating around and around the pump and the Y)

My direct die block is a pretty extreme restriction, and I'm still getting sufficient flow through it, but it looks like my work's not over yet...

I though this might happen under pressure! :( . the lenght of the tube to the Y is acting as a 'buffer', but as the pressure increases it becomes easier to pump back into the tube than to pump through the system :(...

one way valves on the tubes would work. a ball bearing type thing would work well enough for our needs would'nt it?, a flap would be better cause it'd open up the whole tubes diametre, or the 'ball' valve could be oversized...

Yea, oh well I'll look at a combined Y and valve setup tomorrow, see what I come up with.

Did you try stopping the impellor from starting the wrong way yet?, does it flip back or not?...

It would be nice to build a mini reservoir and have little rubber flaps that covers each inlet, the pressure would either flip them open for the main flow, or close down for the bad one, they could be very pliable as well so they'd open easy enough without restricting, the system back pressure of the water itself would hold it shut from flowing the wrong way, bit like how a bicycle pump works, so there'd be no positive pressure valve to restrict the flow, they could both open up as well if there was'nt too much backpressure...

I'm pretty sure there are no~return valves that work off of the back pressure rather than having a spring involved, a 'passive no~return valve' maybe?. it's just a job of getting the size/flow~through of the valve right...

RS is a good place to start I 'spose :) ...

hea I like the rubber flaps idea, will look into that.

I havn't looked into the ratchet yet, I feel it will prove to more trouble than it's worth.

/images of broken impellers...

You could make the ratchet out of rubber too :) , it'd help the 'bounce back' effect :D ...

Had any joy finding low resistance no~return valves?, there lies the easiest path :dome: ...

What do you think of having the outlet shaped as a funnel? O>-*, instead of a circle with the outlet flush? O-* . do you think it'd improve matters?, or would the bulk of the water in the > part retard flow?...
It'd work either way though, that's the important part...

*The - part (barb) is supposed to be central, as in a standard configuration...

Could something like this be made without being too restrictive?

assembled, with cuts in the rubber...

It might be restrictive if not oversized, if the rubber was'nt firm enough it'd leak, too firm and its not going to open up. I was thinking of a kindof 'catflap' that only opened one way, with just enough tensity to keep it in the closed position, you'd still have the full size of the bore with a flap. I'd try to incorporate it into an airtrap~fillpoint or small fitting, as I'm not keen on reservoirs...

It's a good idea though, never seen 'owt like it before...

Yeah, I thought so... This would be a little on the hard side to incorporate into tube, if not ovesized. I know something like this is not so good, bu just adding another O between center and outer ring might make it a little stronger on holding water not going throug backwards. Although, added restriction comes in the way...:(

I just got inspired by the heart (human or animal) blood valves, so if anybody got those artificially made, go for it...

Sorry, this just got too morbid...

Back on the bright side :cool:

Maybe something could be derived of this one... It could be made of two pieces of plexi, than glued together, but you would have to have the piece holding the rubber milled, glue the rubber, than glue second part (just with a barb) to first one...Now I see that that second one would have to be milled too, to give space to the rubber to flip open...hmmm....

In hydraulic systems i've often seen this:

Puzzdre!!, that's what I was thinking!, except have both outlets entering side by side, saves making two of them, there may have to be a lip between them, or have them both countersunk in abit, to stop the pressure from the 'good' one 'swirling' the flap up on the 'bad' one.

If the two outlets are side by side on the pump it could be 'built on' to the pump somehow. No!, scratch that!, I think a bit of distance from the pump is important to stop the impellor lifting the weak side, allow a distance for the pressure to diminish, the further H2o travels along the two tubes the greater the difference between the two heads becomes...

G~mat!, It'd be nice to have a 'passive' one, one that works on a 'slight' negative bouyancy of the ball and the backpressure, rather than a spring...