How does Planet's boiler get watered?

[What About The Bee]

As I plod through reproducing Planet in CAD, for my intended model, I study the Armengaud images to see how the parts interrelate. Along the way, I have noticed how water is introduced into Planet's boiler, even though Planet's boiler is under pressure. While this detail is extremely unlikely to migrate into my model, I think it interesting.

We begin at Park Side.  Park Side was the approximate halfway point between Liverpool and Manchester. The locomotives took on fuel and water here. The LMR detail appears in Colyer, and shows the water crane in situ, along with a Planet Class locomotive.

This does not offer us many details, but thankfully, Armengaud shows us the complete workings.  

The main valve is located below grade, likely to keep it from freezing. When open, water flows up the empty column and into the boom which can rotate. At the top is a pull chain valve, weighted such that it is normally closed. When the chain is pulled, the valve is opened and water can flow in the boom to the tender.

On the top of the tender, we can observe two rotary handles. These handles are on Hornby's Lion and Tiger tenders . I've also seen the video of the replicas, wherein the enginemen turn those handles. I did not understand, at the time, what they did. I do now. These are connected to two valves, located in the base of the tender

When the valves are open, water can flow from the tender.

The interconnection between the tender and locomotive is quite interesting. Here is the view of the device in situ, hanging from a chain off of the footplate.

And here is the cut away. 

The middle region, at B', is a linear extension, that is, one tube slides inside of another. To compensate for angular deflection, there are two ball and socket connections at either end of the linear extension. This allows Planet to move relative to the tender, without hoses! Note that there are two of these, one on either side of the footplate

Fantastic, the water is now onboard Planet, still at atmospheric pressure. The water is directed to cross head pumps. There are two, one on each crosshead

Armengaud provides us with a cut away detail of the internal workings of the pump.

When the pump piston retracts, negative pressure is created in chamber, lifting the lower ball from it's seat, drawing water up. The ball, contained in its cage, re-seats due to gravity. The pump drives forward, creating pressure in the chamber. When the pressure in the chamber is greater than the pressure in the boiler, the upper ball lifts from its seat, and water flows into the boiler.

Viola! Water is injected into the boiler.

A couple of details.  

1) The tiny upright pipe at the end of the pump is noted to be for the enginemen to check that the pump is pumping! When the valve there is opened, water will come out of the small pipe like a squirt gun, completely visible to the enginemen. Proof the pump is working.

2) Why all the duplication? When water is not present at the top of the firebox, the boiler can fail. Insuring the water remains at that level means a functional pump. Without it, loco go boom. Recall that the enginemen have a way to insure a pump is actually pumping. So two pumps implies that the designers were very concerned about pump failure.

3) What if the boiler is full? How to stop the pumps? The pump is hard connected to the crosshead. It is in motion when Planet is in motion. It cannot be stopped. What can be stopped is the flow of water. The valves in the tender tanks can be closed individually.

[threelink]

Fascinating, as ever, Bee. Do the drawings indicate what form of seal was adopted on the linear extension pieces and the ball and socket joints of the loco/tender connection? In the absence of an effective seal I would expect the connection to leak like fury if the water was under any pressure. If the water was not under pressure, the connection appears over-engineered in the sense that a simple hose would have sufficed. Was there a valve permitting water to be returned to the tender, which would place it under some pressure?

The concept of turning off the water supply from the tender raises all sorts of questions about the crosshead driven pumps running empty (and therefore dry), with all that would be entailed in terms of friction related wear. If the pumps ran continually and were always full, any surplus water being diverted back to the tender would pressurise the connection, necessitating the ball and socket/linear sliding pieces instead of a simple hose.

I have read that in the search for an effective seal, early engineers tried all manner of materials including rhinoceros hide (apparently very successfully in mining pumps).

[What About The Bee]

Hi ThreeLink 👋

I must admit that I do not speak or read French. In order to get the relevant text description, I use the plate number, figure and letter references to find likely candidates within the text of the book. I then use a photo translation tool which only permits a sentence or three at a time. It provides, at times, hilarious results.

You raise the point of a hose. To our eyes, this is so obvious! A flexible conduit for water. Why not?

For 1830s England, that would be a complete anachronism. That is, a water tight conduit, that was flexible, simply did not exist. Hancock and Macintosh, along with Goodyear, were in the process of developing the India Rubber patents. Gutta Percha was off in the future.

Do not miss the 1887 advertisement:

https://thegardenstrust.blog/2019/07/13/the-history-of-hosepipe/

So instead of an over engineered solution, the water interconnection appears to be a reasonable approach. The other thing which occurs to me is that Armengaud is simply showing off, ineffective hoses were used, but this seems quite unlikely to me.

The linear extension has the standard packing gland, right hand side between the ball and socket joints. This would be the same type of packing gland used at the steam chest and piston rods, so it would be reasonably effective vis un-pressurized water.

The socket, of the ball and socket, can be tightened down with the threaded rod and nuts present. I do not see any other means of making it water tight. No problem getting the parts to rotate relative to each other, Planet on one side, a full tender on the other. That will provide a large torque moment.

As to the theorized return: we are in full agreement! It, again, is so obvious. Yet try as I might, I could not find one. I thought I had found it with the "squirt", but that connected to nothing AND the translation, whilst hilarious, eventually became clear. Not the return.

I am left with either the pumps constantly pumping OR shutting the flow off with concomitant pump issues.

In the first analysis, the boiler will get water beyond safety pressurization. The full power of Planet from the cross head forcing water into the boiler is an issue. Consider the extrema: it will drive the pressure in that vessel very high indeed. To the extent that it may stop the crosshead, damage the mechanism, damage the pump or destroy the boiler vessel. Perhaps this is why the primary shut off valves are on the tender. The fireman could maintain the correct water level via observation of the glass and adjustment of the valves.

In the second analysis, once the water in the pump chamber is slightly reduced, it will not develop sufficient pressure to overcome the upper ball seat. That is, the water sloshes around but goes no where. The pump doesn't run dry, but is ineffective at driving water into the boiler.

I would like to emphasize that this is just my understanding at this time. Further discovery and analysis awaits. The entirety of the Planet class locomotive drawn by Armengaud is very rich with detail. I'm just scratching the surface. I did think it interesting to bring forth, in light of the recent discussion of your model of the Penydarren locomotive. Your solution is correct!

Bee

[Topcat2018]

Those early 19th century engineers weren't daft, coming up with a solution in the absence of high-pressure flexible hoses. But it seems that it would have allowed no fore and aft movement between the loco and tender, and what happened when they went round bends? Also the boiler pressure was only about 50 psi, so why couldn't they use canvas hoses like those on early fire engines (which did leak a bit)?

Or was it too early for those?

[What About The Bee]

Hi Topcat 👋

Please do inspect the drawing with the words "Linear Extension" prominently displayed.

The pipe betwixt the two ball and sockets is actually two pipes, one inside the other. So the inside and outside pipes slide longitudinally with respect to each other. The gland packing is to the right, just before the right ball.

As to canvas hoses, certainly that is possible. The vertical distance from the water level in the tender, down to the interconnection is a few feet. That would have linearly increased pressure with depth. If a canvas hose was used, it would likely leak like a sieve, draining the tender. Armengaud did not invent his connection from whole cloth. He drew something that did exist, albeit maybe not to exclusion of all other solutions.

@ThreeLink

After closer examination, I do think packing could be placed at the join line of the two piece socket. This would effectively seal the ball. Nothing definitive, just a thought.

Bee

[96RAF]

The double ball valve system is standard methodology, the top valve seats against suction when the pump is stroked to fill from the bottom valve, then the bottom valve seats as the pump is stroked to push that charge past the top valve into the boiler.

[Topcat2018]

@What About The Bee

Wasn't sure what the sliding link was until I zoomed in. Now I can see it more closely.

[threelink]

@Bee. I have to admit that I know nothing about the development of hoses and I am sure that Armengaud would have drawn what he saw. Unless the water was under pressure a hose would have done but if there were none capable of doing the job, the ball and socket /linear slider solution would be the only answer. I see the gland arrangement but am interested to know whether the packing would have been metallic packing or something else (such as the rhino hide I mentioned in my last post). What ever the packing, it seems a bit risky to rely on residual water to prevent drying out and consequent damage, but the Stephensons knew what they were about and the system clearly worked. It would be an interesting exercise to build a replica (full size). Until then I hope that you will keep the thread going - it's fascinating.

Best Regards

[96RAF]

I think the old engineers knew a thing or two about gland packing materials.

The tube within tube sliding link principle is still in use today. Some PTOs on tractors use a square section to provide drive yet allow expansion and contraction of the driven interface link.

[What About The Bee]

Hi ThreeLink 👋

With regards to the packing gland, or 'stuffing box', Armengaud is silent.

In the image of Planet's cylinder, two packing gland cavities are shown. Both are at working boiler pressure.

I added the colors for clarity. The red packing gland seals the piston rod. The green packing gland seals the rod which drives the slide valve. In neither case does Armengaud have letter references or other demarcation to indicate relevant associated text. You would think this of some interest, as the initial packing glands, developed by James Watt, himself, were in the 1760s, a mere 70 years before Planet. Yet Armengaud is silent on the gland itself

Note the piston N. There appear to be two piston rings, I think we can see the split in the ring on the right. Piston rings are still used today in internal combustion engines. So this seal was understood

I'm not sure where we can go further with the Planet's packing glands. If Stephenson could seal the piston rods and slide valve rods against working pressure, he would have little issue sealing the water interconnection between the tender and locomotive at atmospheric pressure.

I poked around at historical packing glands, which brings us to James Watt, who used tallow and oil. I did see one reference to leather piston rings, lending absolute credence to rhino hide seals. Strings impregnated with tallow. All sorts of materials. But in the end, what Planet used on the LMR is currently out of my reach.

I hope that helps!

Bee

[threelink]

Informative as ever, Bee. The history of packing glands and seals is a subject in itself. When Newcomen perfected his atmospheric engine in 1712, his 2 main difficulties were boring a straight cylinder with truly parallel walls and creating an effective seal between piston and cylinder. The latter eventually consisted of leather kept moist by water contained in a reservoir at the top of the (open ended) cylinder from where it could soak into the leather, keeping it moist.

I agree with you about the piston rings shown in the Armengaud drawing. Keep up the good work.

Best Regards

Threelink

[threelink]

Correction. The reservoir was formed by the top of the piston being dish shaped to contain water which splash lubricated the leather seal/piston ring. The cylinder being open ended it could not contain a reservoir. Sorry to have caused any confusion.

[What About The Bee]

We previously observed how water arrived in Planet's boiler. We looked at the crosshead pump and there was some concern about that pump.

I spent quite a bit of time examining the Armengaud images for control features near to the footplate and cross referencing them to Pambour, another noted period author.

In doing so, I was able to understand Planet's water management scheme better. Firstly allow me to present the updated CAD Model 

Further, here is a schematic diagram showing the valves.  You can zoom this and any image

Annotated on OO Planet

I was always bothered by the squirt of the crosshead pumps. The squirt was used by the enginemen to confirm that the crosshead pump is indeed pumping water. When opened, a thin spray of water will be lofted into the air. The only time the pump functions is when Planet is rolling, yet the valve is buried down in the frame. How could the enginemen manipulate the valve? S1 and S2 are long levers that attach to the valves, easily accessible from the footplate.

Previously, we observed that the water from the tender could be shut off using FP1 (and or FP2). The concern was that the pump would be running dry.  If E1 (and or E2) is additionally opened, water will be extracted from the boiler and then that water will flow into the intake of the crosshead pump.  Where it is pumped back into the boiler! Note that E1 and E2 must extract water from deep in the boiler, not steam from high in the boiler.  

Next we come to L1, L2 and L3. Armengaud is quite specific that L1 is at the steam level, while L2 and L3 are at water level. These vent directly to the air around Planet. Pambour calls them "Faucet Gauges" [literally "robinet-jauge"]

I think that they could be used to fill the boiler at the start. Open the Tender valves (T1 & T2), open the Footplate valves (FP1 & FP2), and all three Faucet valves (L1, L2 & L3). Water flowing into the Tender will seek level, flow through the ball seats (depicted by the blue ball & semi circle), and into the boiler. When the water level hits L3, it can be shut off and then, when water hits L2, the tender valves (T1 & T2) can be shut, filling the tender to the brim. L1 can be shut at leisure.

I can see L1 being used as a way to dump pressure in an emergency. Consider a situation of high overpressure, with both crosshead pumps destroyed. Open L1. That will be an exciting moment, as high pressure steam and boiling water is spewed into the air. Yet Planet will be saved.

That is my understanding so far. You must know that I do not have an operators manual, and anything I can glean from the arrangement is purely by mental exercise. I could be wrong or have missed something.  

Your thoughts and views are encouraged! You may have a view that cracks the problem open.

Bee