A simple pump – is this design feasible?

[PaulRParticipant @paulr]

A novice question…

I'd like to build a simple cross-head driven pump something like one illustrated in an old book. I've made a very simplistic 3D model to illustrate the basic idea (but omitted all detail):

I'm thinking of using a piston valve rather than a slide valve and having a power piston in the region of 1/2 to 3/4 inch with the ram around 1/3 of that.

The trouble is I have no idea how much pressure (air or steam) would be required to actually operate it – I've seen the formula for calculating output but is there a way to calculate (or approximate) the minimum input pressure required to get it to work?

[PaulRParticipant @paulr]

[Neil WyattModerator @neilwyatt]

If the ram is 1/3 the diameter of the piston and the stroke is the same your starting point is 1/3×1/3 = 1/9 of the required water pressure. But then you need to allow for frictional losses, lifting valves and other losses.

I would guess you would be able to pump at least 4 or 5 times the steam pressure.

The real question is what water pressure do you need?

Neil

[PaulRParticipant @paulr]

Thanks Neil. I don't need any particular pressure at the minute as it's not going to be feeding anything, it's just for my amusement.

[Neil WyattModerator @neilwyatt]

Well if you have a bigger flywheel a small double acting steam engine should run on 5-10 psi of air, lifting water a couple of feet with no constrictions shouldn't require much more.

Neil

[Andy AshParticipant @andyash24902]

I don't know if you plan to implement a guide for the crosshead.

My thought was that if you load the pump, then everything will be fine, but if you load the crank then possibly not.

It looks to me from your model that if you load the crank then the piston rod will tend to rotate when the piston is pushing towards the crank.

I think you will need guide bars to prevent that behaviour.

Edited By Andy Ash on 20/01/2017 07:43:04

[PaulRParticipant @paulr]

Thanks Andy. Yes, there'd need to be a crosshead guide I just left it out for clarity. I'd prefer to omit the crank/eccentric altogether and make a Stuart style pump but I made a miniature prototype (~1/4" cylinder bore) a few years ago and found that even without the pump, it needed a lot of pressure (given its size) to stop it stalling:

Mind you the valve is a poor fit and the lumps on the valve rod are disproportionately large/heavy. Nevertheless, it works.

This model using a similar valve control is described in an old book, although I guess all the rods and levers could be omitted and the valve driven directly by the piston rod striker.

Trouble is I'm not sure if it would also suffer from a lack of flywheel, which is why I'm thinking of the version I posted originally. That said, I see some of the Stuart pumps on Youtube seemingly being driven at very low pressures. I guess its the fine points of design and quality of workmanship (and I'm lacking in the latter!).

Cheers

Paul

[Andy AshParticipant @andyash24902]

I don't think there's anything inherently wrong with simplex/duplex pumps, you see them fairly commonly.

I've never built a duplex pump, but I imagine it to be simpler than having to build up a crank and machine a flywheel.

Glands will always be a problem, and for a new build I would suggest o-rings, but you do need to be able calculate squeeze and machine to good tolerances.

I have used duplex pumps on other peoples engines, and you cannot fault them for ability to whack water into a boiler quickly. For obvious reasons this is a huge benefit if you suffer from injectors that refuse to pick up. On the other hand they use a lot of steam, and they tend to make quite a rattly racket.

Generally they're best when they can be coaxed to run gently and slowly, but with no rotating elements it is difficult to implement a governor. Without the governor and its feedback loop, no matter how good the throttle valve, they always seem to be "flat out" or "nothing".

I think you could make the pump rotative, by putting a scotch crank in the middle, and then you would have something to run a governor from. With a governor, you would get a very controllable pump.

Edited By Andy Ash on 20/01/2017 22:13:58

[julian atkinsParticipant @julianatkins58923]

Hi Paul,

There are a vast amount of designs in miniature for the type of pump I think you propose making, as there were in fullsize.

The Southworth pump designs in the UK are very popular.

Your old pic illustrates a non reciprocating pump as does your model, whereas your original posted pic suggests something quite different like a Weir reciprocating pump.

When you get down to a small simplex pump in miniature the steam side passages between shuttle valve and main valve are quite difficult to reproduce. If you have no rocking arm reversing the travel (LBSC used a rocking arm on his examples) then you have to have crossed ports which is not easy.

I have made a bit of a study of this and possibly the best summary is in ME 19th April 1956 E.G. Rix 'Simplex valve gears' which includes the very interesting Richardson type.

If you have a non reciprocating pump, then it must have a shuttle valve plus main valve.

There is a limit in miniature to how small you can make these work ok and tick over slowly.

A starting point is the duplex pump described in LBSC's 'Shop Shed and Road/The Live Steam Book. He did describe a simplex pump for a 2.5"g loco of non duplex type but I dont have the drawings for this.

The USA 7.25"g fraternity are very keen on scale Westinghouse pumps and in this size are practicable. Much much bigger than your proposal.

Cheers,

Julian

Edited By julian atkins on 20/01/2017 23:05:17

[PaulRParticipant @paulr]

Thank you both for the advice & suggestions. I did get part way through fabricating a pump (based on some old drawings but no precise plans) with a scotch yoke but it never came to fruition as I had serious problems with alignment that I couldn't resolve so all except the pump body ended up scrapped:

I made another simple 3D model which uses a combination of Stuart type valve for the engine and a scotch yoke to drive a flywheel (well two, for balance):

I think this might be the easiest design to make.get running and experiment with!

(I looked at the Southworth pumps at the October show and they're very nice but I'm not into building from castings.)

Thanks again.

[Jens Eirik Skogstad 1Participant @jenseirikskogstad1]

Better to make feeding pump self starting who is affected by a shuttle valve and a main valve without crankshaft and flywheel. In case crankpin is in tdc or bdc, this is impossible to self start feeding pump. The working piston diameter must be larger than feeding pump diameter to example 10 mm diameter working piston –> 6 mm feeding pump diameter. It will give high pressure to pump feeding water into the boiler.

[PaulRParticipant @paulr]

Posted by Jens Eirik Skogstad on 21/01/2017 13:29:12:

Better to make feeding pump self starting who is affected by a shuttle valve and a main valve without crankshaft and flywheel. In case crankpin is in tdc or bdc, this is impossible to self start feeding pump. The working piston diameter must be larger than feeding pump diameter to example 10 mm diameter working piston –> 6 mm feeding pump diameter. It will give high pressure to pump feeding water into the boiler.

Thanks for that. It's not going to be feeding a boiler, it's just for fun, so self-starting isn't that important

[Author]

Posts