Content Tools

Box 146, County Homes Road, Mt. Royal, New Jersey.

The injector is often thought of as a brass casting with mysterious powers that somehow propels water from the feed tank into the boiler. Often they fail to work at a very important moment and often the common remedy of the bucket of cold water dousing does little to alleviate the problem. Hauling the fire and shutting down is then the only alternative, although sometimes a pressure water hose applied to the injectors suction could save the situation, but not always.

There are many kinds of them including the inspirator which is only the same thing, except the jet is started first through the overflow before opening the discharge valve to the feed check. I think the make most associated with 'Iron-Men' is the Penberthy, though the Metro and the Sellers may come onto the scene. There were many others made by earlier companies that have long since fallen by the wayside, but their product lives, on, and often lay covered with dust perched up in the rafters of a barn. A little understanding of their 'mysterious inards' could possibly bring some of them back to work.

Our magical box was first invented by a French engineer named Henri Giffard in 1858 and it appears to have been introduced into this country by a William Sellers around 1860. In the original Giffard the steam nozzle was made to advance and retard into the combining cone and the jet was only established by fiddling with the steam valve and water valve, plus adjusting the steam nozzle to the correct depth. Sellers overcame this by his self-adjusting injector in 1876 and later the fixed nozzle type. I have no history of the Penberthy Company, but if someone has, I'm sure 'Iron-Men' would gladly print it.

Why does an injector feed water into a boiler? That is the question! (Seems I've heard that before somewhere.) Well, from an old 70-year handbook I have, may I quote: 'An injector works because the steam imparts sufficient velocity to the water to overcome the boiler pressure.' If we have a boiler running at 180 p.s.i. and we placed a thermometer inside, we would find that the steam and the surface of the water were much the same temperature, say 375 degrees F., but steam contains lots more work units and can expand many times its volume. One pound of water by weight, has far from the energy of one pound of steam because of the heat stored within. This is why our injector will not work on air pressure.

The whole secret to an injector is velocity, or feet per second of the steam. Fixing a nozzle to our previously mentioned boiler of 180 p.s.i., we would find that the speed of the steam leaving the nozzle would be around 3600 ft./sec. Now suppose we use water instead of steam at the same 180 p.s.i. we would find only a velocity, or speed, of 160 ft./sec. Divide 3600 by 160 and we get a ratio of 22 to one.


Whenever a jet of gas or liquid comes into contact with whatever, it naturally drives it away, like blowing the chips off your lathe with an air hose. This force is called the 'momentum' and is equal to the propelling jet in one sec. by its velocity in one sec. To obtain this action in the injector, steam leaves the nozzle and enters a funnel, or to give it its correct name, 'combining tube.' This causes water to enter all around the steam jet which cools the steam and causes it to reduce in size. Now the magic box has been given its sucking ability. At the same time this tremendous velocity of 3600 ft./sec. is being given to the water. Each particle of water entering the combining tube is driven faster and faster on by the atoms of steam moving much more rapidly until the steam has totally condensed and combined with the water into one fast and solid moving jet. This is the reason for the funnel of combining cone, and its taper is equal to the length of time it takes to condense all of the steam jet.



Water now leaves the combining cone orifice, as we have established, at a high velocity. O.K. now we nave to get it in the boiler against our still 180 p.s.i. We know that our steam jet travelled at 3600 ft./sec. Now if the suction created by the condensing steam jet draws in say 10 lbs. of water per sec. at a speed of 40 feet, its momentum is 10 times 40 which is 400. Add this to our 3600 and we get 4000. We have also to remember that the condensed steam jet is also an addition to the 10 lbs. of water we sucked in, so with the weight of the jet per sec. it could be 12 lbs. Now if we take the figure of 4000 and divide it by 12, we get 333.3 ft./sec.

Going back to the jet of water discharging at 180 p.s.i. we found that the velocity of the water was 160 ft./sec. So, all we have to do is establish a jet of water over and above 160 ft./sec. and it has got to overcome the p.s.i. of the boiler. Theoretically, we have a jet moving at 333.3 ft./sec, but of course, like everything else, we have losses and water friction within the cones etc. The velocity would be more like 200 ft./sec, but still well above the 160 ft./sec. required.

Having absorbed a little arithmetic, it should now be quite evident how our injector works and simplify some of the mysteries that surround them. In my sketch of the Giffard the parts are not entirely complete but the main parts are shown so that its function can be ascertained. To go over it simply, the steam entrance is marked and its path is through the pet cock and down the opening where it meets holes in the sliding tube. It now travels through these into the inside of the sliding tube and by cranking open the steam adjuster, it enters the steam cone. The remainder of its action is the same as stated above. The only other movement not mentioned is the cone adjuster which moves the steam cone back and forth.

Getting to the Penberthy, you will notice that there is really very little difference in principle than that of the Sellers, except the steam cone is now fixed and a plate valve is fitted half way up the lower section of the combining tube. This has to be a good fit to make a fairly good air tight seal.

Previously I said that Penberthy was the type most widely used by us 'Steam Blokes' so I have also submitted a drawing of some of the steam nozzles I checked, and give their measurements. Four inch size were taken apart, 2 of them from our roller 10541 and 2 supplied by Bob Hartzell. In each case the combining and delivery cones were the same taper and the same throat size, this being .118. The taper as near as I could figure, was 9 degrees in the combining cone and about 5 degrees in the delivery.

For the steam cones, 2 were the same, hence only 3 drawn. For the life of me I cannot remember which of the first two I put back in the roller, although I believe it was the cone with a throat size of .158. Cone #3, I did put back on the right hand side. It works well with only 40 p.s.i. and the other works also, but does not feed so low. By the way, the thread in the nozzle is not to cause the steam to twist as I have heard, its merely for screwing in a bolt so it can be withdrawn easily.

To make a few of these cones in the home shop would not be a difficult task should anyone want to experiment, and by using the formula I have given, the tapers would be easy to cut. To explain this, big D is the diameter of the large end of the bore, small 'd' is the diameter of the small end, and 'L' is the length of the taper. Looking at cone #1, big 'D' is .296. Small 'd' is .158 and 'L' is .593. Subtract i.e. .296 minus .158 is .138. Multiply L .593 by 2 and we get 1.186. Now divide .138 by 1.186 which equals .116. Most machinists have a book containing natural trigonomical signs, so looking up .116 we find it equals 6 degrees 40 mins. Swing the compound of your lathe around to 6 degrees and you're not far off.


Looking at the Penberthy sketch you will find 'X' at various point, and these are joints that must be air tight. The screwed joint at the top 'B' tube usually fits fairly well, so not much to worry here. The flapper 'E' valve is also usually good unless a bit of rust gets under the seat and becomes stuck. The plate valve 'D' should be a nice sliding fit up and down the tube, but if worn a light lathe cut would clean it up. By doing this though, you would have to make a new plate valve with a smaller bore to suit the freshly turned tube. The two X's at the right of 'C' are the most critical locations and must be air tight especially at the lower X which applies the pressure to hold cone 'C' in place. I suppose one good way of checking this would be by using a piece of lead wire between at location lower X and screw in the cap. Even a pencil mark on the edge would do. If a joint is made the pencil mark would disappear. If it did not, then a couple of thousandths would have to be faced off at Cap shoulder Y, but not the bottom of the injector.

Lastly, I would like to say that I am no expert on injectors, but having had trouble with our roller, I thought I would carry out a little research and seek out as much knowledge as I had books available. Books used: 'Machinery's Handbook,' 'Audells Power Plant Guide,' 'Maxims and Instructions for the Boiler-room' by N. Hawkins, and 'Sellers Handbook of Injectors.' It was my enjoyment writing this article and drawing the sketches, and I trust you chaps will find the information interesting.