Pipe Size Puzzle.

[Nigel Graham 2Participant @nigelgraham2]

Problem: Injector very hard to start flowing, with a very delicate touch needed, and when it does run it puts as much water on the ground as in the boiler.

Locomitive: 7-1/4″g NG outline,, the Ken Swan interpretation of the Kerr-Stuart ‘Wren’ so fairly large boiler.

Injector fitted but not prototypically, with its steam feed from the dome. Instead previous (indeed, “past”) club members gave it a manifold above the backhead, but with a rather weeny looking steam valve for the injector.

Injector: commercial but no maker’s name. Figure ‘5’ stamped in the top cap.

All three spouts 3/8 X 32 ME threads (for 1/4″ od pipe unions).

The steam pipe is only 3/16″ o.d. and to complicate things has a kink on one of its wriggles (I would use “bends”, but…). The injector end has a 1/4″ union, the gap between nipple and pipe filled with silver-solder; the manifold end has a 5/16″-pipe union to fit the valve.

 

Water feed: Controlled by standard 15mm plumbing ball-valve on outlet of a tank in the driving-truck. The bore through the ball in these is 8mm, severely cramping the flow in domestic plumbing but more than adequate for a small injector.

 

Operating: water full on. Steam full on. Slowly turn the water back until it is on the verge of stopping, and there seems a “sweet spot” where the injector suddenly picks up and puts some of the water in the boiler. It is very narrowly-defined: not quite there produces copious hot water from the overflow; slightest touch further stops the water.

 

I have searched both Evans and Brown. Neither book is very forthcoming about miniature injector sizes and pipe diameters. Nor do they explain the devices’ rather vague specifications and their selection; but Brown goes by cone diameters, relevant only if you are making them as his book instructs. I even had to use a cookery-book to find units conversions, but that didn’t help much more.

DAG Brown gives the main cause of water-dropping as incorrect steam-cone depthing, and states he finds this common on commercially-made injectors; but I assume with all the rest of the system correct. Obviously I am not going to try modifying the injector’s tiny, very delicate parts to within 0.001″. Not on a massive great Myford ML7. (Read the book to see why). They should be made the right size from the start.

 

So am I right making the steam-pipe the prime suspect: about 18 inches of undersize tube reducing the steam volume and throttling it even before it finds the kinky bit? (And undersize manifold valve too, of course.)

And that all three pipes should be 1/4″ o.d., with control valves to match?

I think the water and delivery pipes actually may be that size, but what happens if the delivery is also 3/16″ od.?

What is the number on the injector cap? 5 of… pints? litres? per minute? Some maker’s design-number?

 

We do have the drawings but the ad-hoc modifications to make the loco a bit easier to drive rather negate those. I suspect whoever fitted the injector simply used some pipe he happened to have in stock!

(On the full-size ‘Wren’ the driver has to dismount and reach up to operate the blower and injector valves, unless he can reach them, Twizzle-like, through the spectacle-plate apertures.)

[noel shelleyParticipant @noelshelley55608]

No 5 would indicate 2.66 pints/min. You need ample hot steam and  plenty of cold water. For this size 1/4″ pipe is required. Until these conditions are met all bets are off. 18″ of steam pipe – I would lag the pipe. I’m no expert and others will no doubt add to the above.  Good luck.  Noel.

[Nigel Graham 2Participant @nigelgraham2]

Thankyou Noel.

I wonder how they deem “5” means 2.66 pt/min.  It doesn’t seem to correspond to anything!

 

I thought the pipe diameter could be significant, but wasn’t sure as the actual passage into the injector is so small. The pipe needed replacing anyway due to being partly crushed near the injector.

I’ve now made a new pipe, using the original as a sort of template, but with a very short stub of that pipe’s top end and union to fit the valve. It’s not easy having only limited access to the locomotive, and the pipe will likely need forming to fit more closely. I intend to complete the work with the right size valve but that will have to wait for now.

The stub of smaller pipe will choke it to some extent, but nowhere near as badly as the entire length.

We intend repeating the steam test during the week, if I can refit the three pipes I’ve serviced.

I’ll examine the injector’s other two pipes but I doubt I’d be able to replace them, if they need that, before we need the loco for a public running event.

 

I don’t think it will need lagging. It’s unusual to lag the auxiliary pipes. They were often exposed even on full-size locos and traction-engines. Also the lagging likely to be practical to fit on this pipe would be too thin to do much.

[Charles LamontParticipant @charleslamont71117]

Has the injector itself been dismantled, cleaned and inspected?

[bernard towersParticipant @bernardtowers37738]

surely your stub of smaller dia pipe restricts the flow the same as a long length?

[duncan webster 1Participant @duncanwebster1]

Mike Tilby did an article (I think in ME) which showed how adding too much insulation to a small pipe can make things worse by increasing the area from which heat is lost. The optimum thickness for a small pipe is surprisingly small.  No doubt some google meister will find the relevant issue.

[Nigel Graham 2Participant @nigelgraham2]

Charles –

Yes- and I discovered the damage to the pipe while disconnecting the injector for cleaning.

Surprisingly it was not particularly scaled, perhaps due to the water-treatment whose primary pupose is corrosion resistance in the steel boiler.

I almost lost the delivery cone too. By sheer chance I saw it on the dining-room floor after I’d packed everything to take back to the loco. It was an easy sliding fit in the injector body, so I’ve held it in with a spare union nut until the whole thing can be re-installed.

 

Bernard –

Yes, of course but far less than the original length being that diameter. The original pipe was also slightly kinked at one point.

I needed graft the original end, a little over an inch long, onto the new, 1/4″ pipe to suit the valve, as changing that will likely need a new manifold as well to accommodate the valve’s own 3/8″ fitting.

 

Duncan –

I remember that article too. It was quite a while back.

I think Mike Tilby also pointed to the flaw in painting the pipe-lagging. Prototypical as in full-size the insulation was often covered with canvas that was then painted or white-washed, but in miniature the paint soaks into the lagging and removes its insulation behaviour.

I don’t recall ever seeing a full-size boiler with its auxiliary pipes lagged. A main steam line to an engine some distance away, yes (e.g from boiler-room to valve-chests on P.S. Waverley), but not pipes feeding things like injectors only a few feet from the manifold. They were sometimes polished though, not for looks but to reduce radiated heat loss – a practice also sometimes used on the exposed cast-iron cylinder covers of very large stationary engines.

[noel shelleyParticipant @noelshelley55608]

When one grasps the way these tiny miracles work then many things become obvious ! Ample steam must be available and able to accelerate to and through the steam cone to meet with water, condense causing an increase in velocity that inturn creates pressure to re enter the boiler the steam first came from = a miracle or some black art ? Ideally the water should not only have some small positive pressure but also adequate flow once the injector “picks up ” ! If the flow is not adequate then a partial vacuum may be created that the injector cannot over come, or a situation will arise whereby the injector will stop/start, if this is the case then water will flow from the over flow.

All of the above only applies to a cold water feed and a properly designed and built injector. Restriction in any of the pipes or the smallest air leak will result in trouble. The pressure of the steam also plays a part, to low or to high may result in a failure to “pick up”

My reference to lagging the steam pipe  was in the context of it’s length/diameter and that it should not condense before reaching the injector.

Good luck and best wishes. Noel.

[bernard towersParticipant @bernardtowers37738]

Surely it does not matter how long the restriction is?

[bernard towers]

On bernard towers Said:

Surely it does not matter how long the restriction is?

There is a small loss at entry to the restriction, a larger loss at the exit, and there is a friction loss throughout the length of the restriction that is greater than it would be in a larger bore.

[Nigel Graham 2Participant @nigelgraham2]

Bends, especially elbows and tees, and valves also restrict flow. Generous curves offer far less resistance than a series of right-angles such as a T-shaped turret and right-angle valve.

On this application I (or rather the device) have to put up with the turret and valve but the 1/4″ pipe starts only an inch or so from the valve, then its right-angle then hair-pin bends close to the injector are sweeping; their radii being of the normal 3 X pipe diameter minimum.

There are formulae for friction losses in pipes and fittings but I don’t know them, nor how accurate they would be in small scales.

 

(I wonder the throttling effects of the usual spear-point returns in the most common type of miniature locomotive superheaters. The full size ones have hair-pin bends.)

[ZanParticipant @zan]

The 5 relates just to the injector size. 53 oz per min or 2.2/3 pints, with a recommendation of 1/4” pipes.

I think the numbering originally came from Don(?) English who produced a wide range of steam fittings and supplied the trade, thus providing easy identification of the delivery.

look in any ME suppliers catalog and you will see the numbers.

no. 4 is mainly for 5”gauge,  no 3 for small 5” or 3”

[Nigel Graham 2Participant @nigelgraham2]

Thankyou for these.

We used the locomotive at a rally a week or so ago. (I had half-expected that to be cancelled or restricted for fire precautions.)

I found the injector certainly better with the new pipe, even with the short remnant of narrower steam-pipe and now-undersize valve. It was still a bit temperamental at full boiler pressure (90psi), although it did feed the boiler.

The injector’s best range appeared to be around 60-80psi.

Consequently we could keep the locomotive in steam throughout Saturday even with long idle periods enforced by events in the adjacent arena, so no input from the mechanical pump. At least a boiler a third of the size (1/27 volume) of the full-size prototype can simmer gently for an hour with a very dull fire. The driving-truck is fitted with a hand-pump, just in case.

On the Sunday there were more arena events preventing our operation, so we gave up and let the fire die out. The injector also realised it was the Sabbath, but the trouble was found simply a bit of dirt in the cones, easily blown out by mouth.

 

So I re-read everyone’s comments above, and studied D.A.G Brown’s book again.

It would seem over the Winter we need:

– fit a proper strainer to the water feed,

– examine the feed plumbing from the tank in the driving-truck, for possible air leaks,

– complete the full 1/4″ pipe-size steam supply.

That does not answer the point about the pressure-range, but Brown suggests this can be due to the setting of the steam-cone in the combining-cone, affecting the tiny annular water-inlet gap. That might be a matter of design: perhaps our injector was for the slightly lower steam-pressure (note – pressure not volume). Or of manufacturing tolerance as these things are fussy about “thous”.

 

Noel –

Water feed head: not quite so.

Injectors fitted with starting valves (the overflow valves) should be able to lift the water; although we normally gravity-feed them and turn the water on first so they are not heated by the steam before the water reaches them. Many industrial boilers and some narrow-gauge locomotives had lifting-injectors (our type), and delivery clack-valves directly on the backhead. I’ve often wondered how they coped!

Provided the water-pipe is reasonably large the injector will take in all what it needs, through an annulus of far smaller area than that of the pipe.

 

Very oddly, our Wren’s injector likes a very reduced cold feed. Turn the water fully on; steam on full; then throttle the water valve back practically to the point of closing!

I can’t explain that except perhaps an effect of using the injector above its intended working-pressure.

I did notice the bigger steam-pipe and using the injector at lower steam pressure, allowed setting the water valve a bit wider open. Altering the steam valve opening made little difference.

 

Zan –

Thankyou for those number clues. I don’t think miniature injectors have particularly standard numbering systems.

The one you cite, indicating the delivery volume in ounces/minute (or pints/minute), is logical. Ideally on a locomotive at least, you want to match the injector to the engine’s steam demand: fine on long continuous circuits but a moot point for back-and-forth portable-track duty.

D.A.G Brown says he designates his injectors by their delivery cone diameters; but that does not directly tell you the flow value.

 

Long-established miniature-locomotive practice uses an axle-pump as the running feed, and swears by (at?) the injector as for use only when stationary; likely by injectors’ reputation for unduly reducing the boiler pressure. Though using them on the move might be awkward. Are they usually over-sized for the system?

Surely it’s the pump that would reduce the pressure as it feeds cold water, and it made one 5″-gauge loco I knew, run quite lumpily when the reverser was notched up.

 

 

[Charles LamontParticipant @charleslamont71117]

Possibly of mild interest, full size injectors are designated by the diameter in millimetres of the delivery cone throat. Common sizes are No.8 and No.10.

[Author]

Posts