Injectors 3


I haven't done much ME over the last week as I've decided to attack the garden and greenhouses with a view to getting ready for next year. I used to grow all my own veg etc. but this has slipped over the last few years and the garden has been sadly neglected. Also I'm still looking to fitting in a 100' test track along the fence! Pity I didn't think of this before putting a shed where I did as it's now in the way!

I did however manage a day in the kitchen last week playing with injectors, getting very wet in the process and doing my best to flood the place!

I did try the 4oz and one of the 8oz injectors on Helen when I went to Whitwick on the 28th October with mixed success (as reported in the Helen section). The 8oz was very temperimental and worked sometimes. The 4oz was much better but would not pick up at full boiler pressure so I tried them again on the test rig after reading all the information I could find (I've collated it all on the PC now for easy access).

The main problem seems to be that the injectors will not lift and have to have water flowing through them before they have any chance of working at all. This is quite a common fault apparently. The late Jim Ewins made injectors that would work reliably under any conditions, even when the water inlet pipe was heated to red hot with a blowlamp! It has been reported that he refused to tell anyone how he made his injectors but in my research I came across a letter in Model Engineer published by a gentleman who had been given a letter written by Jim which Jim had intended to send in to Model Engineer. In this letter Jim describes the 'Linden Secret' which explains how injectors can be made to lift under adverse conditions. E.J.Linden was a well known model engineer who, back in the 1930's (?) produced a range of miniature injectors which had the capability to lift, something quite rare back in those days. He also refused to reveal his 'secret'. Other manufacturers tried to copy the Linden injector but failed to get the same results. It would appear that Jim Ewins got hold of a Linden injector and spotted what others had missed and so found out the 'secret'. The Linden injectors have a small chamfer on the entrance to the second half of the combining cone which is what makes the injector lift properly.

In the Derek Brown articles (and indeed, other articles) this 'chamfer' is mentioned but not stressed enough in my opinion. Mention is made of just breaking the sharp edge of the entrance to the cone which I took to be just slightly deburr it. However, the illustration in Jim Ewins' letter shows a substantial (30 thou) chamfer at 45° although it is not clear which dimension the 30 thou refers to - is it the width of the chamfer, depth of the chamfer, diameter of the chamfer, or what?

Anyway, armed with this information I removed the combining cone from the 8oz injector and put a chamfer on the entrance to the second half with the reamer used for the bellmouths on the cones, re-assembled the injector and tried again. There was a slight improvement but still very little lift! I then repeated the operation on the 4oz injector with the same result - very little improvement. Hmmm.

Further reading on lifting problems pointed to the ball valve - either not enough lift or the passages too small. One article suggested trying the injector without the ball in place and see if the lifting is improved (yes, an injector will work without the ball valve!). I tried this and things did improve somewhat. I spent the rest of the time playing around with not a lot of success. I did try running the injectors without the delivery cones to see what the output from the combining cones was like - it should be a narrow jet of hot water which should travel a long way (about 30') before hitting the floor. This was interesting as my kitchen is considerably less than 30' and being tied to the cooker for a heat source I couldn't go outside. The kitchen window does not have a side opener so I couldn't direct the jet through the open window either! Eventually I cleared everything away from the window area, put some old towels on the window sill to soak up the water and just let the jet hit the window! Good job I do live on my own!

It was interesting to note how the jet changed as the boiler pressure changed. At low pressure the water was a sort of dribble and as the pressure increased the jet got narrower and more powerful until at a certain pressure and above the jet tended to break up and get much wider and contained a lot of steam. I presume the ideal working pressure range is where the jet is narrowest and coherent. Either side of this range, the jet will be too wide to enter the delivery cone cleanly and this is when some of the water and/or steam comes out of the overflow rather than go into the boiler. Observing the water jet as it leaves the injector can also show up alignment problems with the cones - the jet should leave the injector body on the axis of the body. If it leaves at an angle then one or more of the cones are not concentric or are tilted in the body.

Watching the jets from both the 4oz and 8oz injectors showed that the jets were breaking up at pressures quite a bit lower than the required 90psi which probably explained why the 4oz injector worked well below boiler pressure but would not pick up at full boiler pressure - the operating range of the injector was not high enough. Whilst the maximum operating pressure can be increased by shimming out the steam cone from the body this is not the correct way of doing it. The correct way is to make a new steam cone with a smaller diameter to reduce the amount of steam through the injector. Basically, the water to steam ratio is only correct over a certain pressure range - above that range there is more steam entering the injector than the water can condense and you get steam blowing out of the overflow and the injector will not pick up. Obviously the amount of steam that the steam cone passes increases with pressure so the way to reduce the amount of steam is to decrease the size of the steam cone (I think!).

Anyway, I'd had enough of injectors for the time being and did not fancy turning new steam cones at that time so I shelved the job and spent a few days in the garden instead.

After a day in the greenhouse today I decided to get the test rig out and have another play (get wet again!). Just out of interest I decided to try the steam cone from the 4oz injector in the 8oz. Boy, what a difference! The injector dribbled water from the overflow all the time (as expected because the end of the steam cone is much too small for the 8oz combining cone) but worked amazingly well. It picked up perfectly at 100psi (screwed the safety valve down a bit) and it did not matter whether the steam or water was turned on first. If the water was shut off whilst the injector was running and then turned back on it picked up again straight away. Putting a finger over the water inlet showed quite a strong vacuum, in fact you could hear the air being sucked into the end of the pipe. The water flow needed to be restricted to enable the injector to pump over boiler pressure but again, that is to be expected as the cones are not matched. It looks as though we are getting somewhere and the steam cones are the problem. Next job - turning new cones!


For reference here are some further articles on injectors which I have come across and may be useful:

Miniature Injectors - C M Keiller - Model Engineer Volume 84 Issue 2094

Testing Injectors - W A Carter - ME Volume 141 Issue 3519

Miniature Live Steam Injectors - Basil Palmer - ME Volume 142 Issues 3536 and 3537

Laurie's Standard Live Steam Injectors - ME Volumes 156/157 Issues 3774 - 3787 (reprint)

Making Laurie's Injectors (The Easy Way) - Geoff King - ME Volume 179 Issue 4046

Messing About With Injectors - Chris Leggo - ME Volume 190 Issue 4187

Injectors - The Linden Secret (Postbag) - Fred La Roche - ME Volume 191 Issue 4200



Well, I made a new steam cone for the 8oz injector reducing the diameter from 0.028 to 0.025 inches and tried it in the injector. Result - back to square one ! No suction at the water inlet again with very poor starting. Methinks I'm going round in circles!

Anyway, I decided to do what I should have done a long time ago and check the fit of the end of the steam cone in the combining cone and hence the annual gap around the end of the steam cone. The diameter of the narrow part of the steam cone was spot on to drawing but checking the diameter of the entrance to the combining cone with a drill showed that it was too small although the exit diameter was spot on. Possibly my homemade 9° reamer is actually slightly less than 9° which would result in this. What was happening was that when the end of the steam cone was the correct distance inside the combining cone the gap between that and the bore of the combining cone was far too small, in fact virtually none existent with the result that no water could get through the gap and only steam came out of the overflow. Moving the steam cone further out by fitting shims was making the gap wider and allowing water through but the end of the cone was then not far enough into the combining cone. In fact the end of the steam cone was probably not entering the combining cone at all. This was probably the cause of the lack of suction or lift. This was why the injector worked much better with the steam cone from the 4oz injector - the end is a lot smaller diameter and also longer so it entered the combining cone further and produced a strong vacuum at the entrance to the combining cone.

Armed with that knowledge I then made another cone with a slightly smaller diameter end (0.01" less than the original) and tried that. There was a great improvement in the suction and the injector picked up at 100psi with no problems. However, it needed the water flow restricting for it to overcome boiler pressure and a lot of water was lost through the overflow. However it did pump pretty reliably and would start with either the water or the steam turned on first. In order to try and get it to run dry I gradually turned the inside shoulder of the steam cone down so that it would go further into the body and eventually got to the stage where only a slight dribble came from the overflow with the majority of the water going through the dummy load. I did notice though that the starting got a bit temperimental the further the steam cone went in. Measuring the length of the steam cone showed that it was now entering the combining cone by 0.05" which is probably far too much and so the end of the cone is now too small in diameter. Working the figures out from Derek Brown's drawings shows that the steam cone should enter the combining cone by 0.02" and give an annular gap of 0.005". These figures agree more or less with the dimensions given in some of the other articles.I'll make another cone tomorrow with a slightly increased diameter and try again! I'm determined to get there sooner or later!

In his injector articles, Basil Palmer does describe a method of machining the steam cone to match the combining cone entrance which is fairly simple to do so I think I'll give that a try rather than my trial and error method!


The saga continues!

I've spent the last few days playing injectors again (too wet to do any gardening!). I've made several new steam cones (I can turn them up quickly now!) but still could not get the 4oz injector to run dry. In desperation I ran the injector without the delivery cone to check the water stream from the combining cone to see what was happening. This time I rigged up a funnel and a length of hose to catch the jet and direct it down the sink!

I noticed that at low steam pressure the water jet left the combining cone dead square to the axis but as the steam was increased, the jet veered to one side, the higher the steam pressure, the more the jet went off axis. This made me suspect that something was not right with the combining cones so I made and fitted a new pair. Exactly the same thing happened with the new set! Further investigation and a bit of thought explained what was happening. When the water to steam ratio was correct the jet left the combining cone dead square. As the ratio of steam to water was increased the jet changed from just water to water containing uncondensed steam and it was the steam in the water that was sending the jet off axis. Why it always veered to the same side was probably something to do with the way the water entered the combining cone. Turning the steam cone to a different position did not affect the direction of 'veer' so it was not caused by the steam cone being out of line as I first suspected.

Another experiment I tried at the same time as above was to fit the delivery cone with just a short length of pipe so that the exit was open to the air instead of going into the dummy load. When the steam pressure was low and the jet from the combining cone would have been inline, water issued from the exit pipe in a nice steady flow and only travelled a couple of inches. When the steam was increased causing the jet to go off axis and contain steam, the flow from the exit pipe broke up and shot out several feet in a spray.

Armed with this knowledge I then adjusted the position of the steam cone with shims until the water/steam ratio was correct at the required boiler pressure and the water jet still left the combining cone dead on axis. I figured that the steam cone was now in it's optimum position and any further problems would be with the delivery cone.

The delivery cone was refitted and the injector tried again. It picked up and pumped ok but there was a lot of water coming from the overflow still. In Laurie Laurence's articles he mentions that the gap between the combining cone and delivery cone is critical for correct operation so I tried varying that with great success. The relevance of this is that apparently the water jet is slightly bulbous in shape as it leaves the combining cone rather than a parallel jet and the end of this 'bulb' should just touch the flare on the delivery jet. If the gap is too small, the outer parts of the jet hit the flare and are deflected back causing overflow. If the gap is too large, the bulb does not reach the delivery cone and air is sucked in through the delivery cone and causes the infamous chirping noises (it is suggested elsewhere that this is caused by a leaking check valve but obviously this is not the case) I eventually got to the point where the injector ran dry over a range of from 100psi down to 70psi and below that it started to drip until it packed up at about 30psi. It was interesting to note that it did not start dry if the water was turned on first - it dripped, but if the water was turned off and on again it did run dry. It would start dry if the steam was turned on first!

A tip that I picked up from LBSC's writings is to open the steam valve slowly until the injector picks up and then open it fully rather than whack it full open straight away. I noticed this method worked well for the small injector. If I opened the steam valve fully straight away the injector will not pick up and just blows steam from the overflow.

The measured output was a little higher now at 6oz but this was no doubt due to all the water going out of the delivery cone instead of the overflow! Out of interest I made a new delivery cone with a slightly smaller bore and this did reduce the output to about 5oz. At this point I have decided to leave this injector as it is and refit it to the loco ready for the next run. The temporary shims made from rings of wire were replaced with permanent ones turned from brass bar and parted off to the right thickness.

The next job is to have another go with the 8oz and see if I can improve that. What I will also try and do is get some photographs to show the problems I have encountered to illustrate the above notes.

As promised, some photos of the water jets:

Output from the combining cone:

Not enough steam - water just dribbles out

Water/steam ratio just right - water exits in a straight coherent jet

Slightly too much steam - water jet breaking up and deviating off axis

Far too much steam - water just sprays out everywhere!

Output from the delivery cone:

Water/steam ratio ok - water exits in a continuous steady stream

Excess steam - water stream breaks up and sprays out


Previous Page Next Page

Page 1 2 4