Martin Evans 5 inch gauge ‘Firefly’ Crosshead Pump – issues

[Keith ChartersParticipant @keithcharters25446]

Hi, I’m in the process of refurbishing our club 2-6-2 Prairie Tank to Martin Evans ‘Firefly’ design. One (of many!) of the issues I’ve encountered is the Crosshead Pump, which just doesn’t put any water in the boiler when running. The pump works, as when I open the bypass valve I can clearly see the water being recycled into the side tank! However, when the bypass valve is closed the water level in the boiler continues to fall and we have to resort to the hand pump. I’ve removed the pump, striped it down and descaled everything. The clack valves on the water input and output work as they should and disconnecting the top feed from the boiler shows water is being pumped to the boiler! The pump just does not seem to be able to overcome the boiler pressure and feed water. I checked the dimensions of everything as per the drawings and the only ‘issue’ I found was that the ram was over the specified length! Should be 2 and 9/16 inches but was 2 and 3/4 inches, oversize by 3/16 of an inch, which seems rather a lot. I have no re-machined the ram to the specified length and re-fitted it to the loco. It has not been tried yet as, the loco is still partly stripped down for other issues. Can anyone tell me what, if any, issues an overlength ram in the crosshead pump would cause? Obviously the stroke of the ram remains the same but would it affect the water pressure being delivered?

TIA, Keith.

[Clive Brown 1Participant @clivebrown1]

I can’t see the over-long ram reducing the effectiveness of the pump. If anything one could argue the opposite, the longer the ram then the less clearance volume there is. Thet extra 3/16″ might just allow the end of the ram to cover the water inlet / outlet ports at the end of the inward stroke, with no real effect on the pump performance.

Since the pump is positive displacement, the water must be going somewhere if not into the boiler. I can only think of an untraced leakage or a volume of air somewhere in the system up to the boiler clack that is not being cleared. Could there be back-flow via the hand pump perhaps?

 

[Nigel Graham 2Participant @nigelgraham2]

The only effect shortening the ram might have is slightly reducing the volume of water pumped per stroke.

This sounds like a non-return valve problem somewhere in the system, likely in the pump itself.

[duncan webster 1Participant @duncanwebster1]

Shortening the ram should not change the volume pumped per stroke. Just because it pumps with the bypass open doesnt mean it will pump against boiler pressure. If the in clack is leaking it could just send water in and out of the tank, and if the feed clack is leaking it could just shuttle water in and out of the boiler

[noel shelleyParticipant @noelshelley55608]

It’s GOT to be a feed or clack problem be it in the pump, on the boiler or a faulty bypass some how leaking under pressure. That is assuming that the pump ram seals on the cylinder – has the O ring failed ? Noel.

[Nigel Graham 2Participant @nigelgraham2]

I think rather similarly, Noel.

Let’s examine what each fault might do:

A leaking piston-ring would be fairly obvious by water squirting out of the end of the pump.

A badly-leaking bypass valve should show in the tank.

A feed clack stuck closed would cause a hydraulic lock if all the other valves work properly. If the water can find somewhere to squeeze through it may show as jerky running by the locomotive.

What if the pump’s inlet valve is not closing? It ought under boiler pressure, reflected by the water pressure within the pump, but the pump is working enough for it to return water via the bypass, so the inlet can’t be sticking fully open.

On the other hand…

A sticking inlet valve with the bypass closed would cause the water simply to oscillate in the pump and pipe. If the tank oulet is visible that should show if water is simply going backwards and forwards, although this would be hard to observe on a locomotive under way.

Yet the bypass returns a healthy flow? Well, the water will take the line of least resistance, and if the inlet valve is not fully closing more water will escape via the open bypass than will go back through the inlet.

.
Clive suggests:

Back-flow through the hand-pump. Only if they share plumbing – and again this needs non-return valves returning.

An air-lock. Unlikely. The air will simply compress to equal the water-pressure but if all else is working it will not impede the flow. It is more likely just to be pushed into the boiler, though a small bubble might occupy the top of the clack-valve chamber. In fact reciprocating pumps and hydraulic rams (the sort that pump stream water) were often fitted with large air-vessels on their outlets, to remove shock and smooth the flow.

…

Any other possibilities?

Is the pump working but simply not passing enough water to make up for that evaporated, due to some obstruction upstream?

E.g.

Tank strainer clogged or too small area?
Dirt, solder or flux in the pipe from tank to pump?
A kinky bend or other damage partially closing the pipe?
Inherently restricted inlet plumbing: sharp elbows, perhaps the pipe a bit undersize, inlet ball not lifting enough?
Or lifting too high so it loses some water on its way back down?
Inlet valve ball partially obstructing the valve outlet?

Fast running could affect it, if the pump is only just matched to the maximum steam demand, by the barrel not being able to fill fully at each inlet stroke. This is particularly so if the inlet pipework is too restrictive. Reciprocating pumps are not too good at fast operation.

[Author]

Posts