injector design - a few musings

[jma1009]

Roger raises a very interesting point.

so by way of example virtually all of Gordon Chiverton's medium injectors have their combining cone starting 3/8" into the injector body. that means the steam cone is a tad over 3/8" length plus the width of the flange on the steam cone. on one of Michael's 'problem' injectors, the combining cone started near enough 1/2" into the body of same effective length. so 1/8" is lost on the combining cone and delivery cone somewhere!

if you examine the successful designs, the proportions are readily apparent. Bob Bramson wrote at some length of the importance of the proportions of the 2 halves of the combining cones for different sizes of injectors and how these needed altering for high pressure injectors in miniature. the successful commercial products comply with Bob's analysis and proportions for the combining cone, perhaps by accident. unfortunately there are many inferior commercial products that do not comply and give injectors such a bad name.

i tested 3 of Michael's injectors today, as altered by me. 1 was excellent (the John Cashmore injector with new steam cone to replace the damaged original) but the other 2 worked of sorts but not to the required standard i would be happy with and would be 'bin jobs'. the new parts were 'spot on' but the proportions and details of the original combining cones were all wrong, despite being altered internally in the case of the Polly injector.

i do despair for our loco building community if 2 commercial products as obtained by Michael recently are sub standard and wont work to the required standard expected, even when 'corrected' by me. the design parameters for a successful injector in different sizes are well known and published yet it appears that only Len Steel via Pavier Steam now supplies reliable commercial injectors, having taken over Gordon Chiverton's products. usual disclaimer.

my advice is to have a go at making your own to Derek Brown's designs! buy Derek's brilliant book and 'have a go'. i have no interest with Derek, and usual disclaimer.

there is something rather nice (OK, an ego trip!) to say 'yes, i made these injectors and they work perfectly' etc.

it is rather a shame Derek isnt a member of this forum. he has a superb 5"g LNER A4!

cheers,
julian

[oldnorton]

Ahh… I answered my own question about vertical injectors a few minutes after posting. The primary air valve ball has to sit vertically on its seat, so all injectors have to be oriented to allow that.

[Roger]

Is it necessary for the ball just to rest on its seat, or could there be a spring? In other words, how important is the free flow of water from the middle of the combining cones?
I'm still unsure of the dynamics of that part of the design. I'm guessing that solid combining cone would choke because it would be too small at the outlet until it was running? Does the gap in the middle allow the build up of momentum through the larger area at that point until there's enough to pass straight though the smaller half?
I seem to remember someone saying that once it's started, you can remove the ball valve and see the jet through the aperture. I also seem to remember seeing a design where the overflow from that area was separate from the overflow after the combining cone. I'm not sure if there was a ball valve in that place though. Is the ball valve only necessary because it shares the same overflow on most designs?

[jma1009]

hi Norm,

when the steam exits from the end of the steam cone nozzle it is at below atmospheric pressure. also there is a sort of 'ejector' effect between this jet of steam and the water that enters at this point.

on a well designed lifting injector the steam cone and first half of the combining cone act as an ejector. if the overflow passage isnt big enough and the parts not correctly proportioned this useful (i would say vital) function is lost.

until the injector 'picks up' this mass of steam and water needs to escape - hence the check valve overflow. the only purpose of the check valve is to aid the starting of the injector till it works ie 'picks up'. once the injector starts working this part of the injector is working at below atmoshere creating a vacuum hence the need for the check valve.

a well made injector will work with the ball seat upside down or on one side. the Chiverton vertical injectors have the ball seat at an angle. Derek Brown's 10X injector design has the ball seat on the side.

Roger is quite correct about the overflow from the check valve. LINDA has the overflow from the check valve venting via a pipe, and not back through the injector body, and there is no overflow between the combining cone and delivery cone. it works quite ok. due to their being no overflow between combining cone and delivery cone it would be a design ideal for steel boilers.

cheers,
julian

[Roger]

Hi Julian,
Just to confirm that I understand you correctly... on LINDA, am I right in thinking that there isn't a ball valve on the overflow between the two halves of the combining cone?

[jma1009]

hi Roger,

there is a ball valve on Linda's 'special' injector

cheers,
julian

[jma1009]

here's a pic - the 'special' injector is the one on the side of the saddle tank. there is an ordinary 'standard' injector under the cab

[Roger]

Well, the "All you need to know about miniature injectors and ejectors" book arrived today and it looks to be an interesting read. It's a real pity that it uses obsolete Imperial measurements throughout, it's so much more difficult to work with these units when all my Thermodynamics were done in SI units, and that was over 30 years ago. I mean, who uses Farneheit these days!
Oh well, I'm sure it's still a useful document when it comes to understanding the principles.

[joanlluch]

Oct 14, 2015 20:25:23 GMT Roger said:

Well, the "All you need to know about miniature injectors and ejectors" book arrived today and it looks to be an interesting read. It's a real pity that it uses obsolete Imperial measurements throughout, it's so much more difficult to work with these units when all my Thermodynamics were done in SI units, and that was over 30 years ago. I mean, who uses Farneheit these days!
Oh well, I'm sure it's still a useful document when it comes to understanding the principles.

Hi Roger, 

Such a thing of the imperial units is certainly intriguing. I certainly have heard about inches (25,4mm) during my career, but I only learnt that they were actually used one year ago when I received the 5"G Britannia plans from Reeves. What the... I thought.... That was the very first time for me that I had seen such a thing.

What's almost hilarious is that all new steam locomotives that are currently being built in the UK from scratch by the A1 trust team in Darlington, make use of the latest 3D CAD technology and manufacturing techniques and yet all plans are being drawn in "imperial" units, which of course can only be expressed in the computer with lots of decimals, (49/64" is 0.765625" for example). They claim that one of the goals of the initiative is to motivate and bring steam technology to future generations, and yet they have lost the opportunity to design it all in millimetres, metres, kilograms, litres, watts, and centigrades (??). Incredible but true.

Anyway, let's not be so picky about that. Although imperial units are a bit annoying for those not grew up in the UK, I learnt to accept them as a unavoidable temporary fact, because there's no doubt to me that all this stuff will be all gone at some time, possibly as soon as when the current generation of engineers is replaced by the next one. We are all still afflicted by this, but will not be our children or grandchildren.

[jma1009]

hi roger,

i think you will find Derek Brown's book more to your taste

www.teepublishing.co.uk/books/operation-valve-gears-injectors/miniature-injectors-inside-and-out/

it is quite excellent.

you will also like Derek's approach to making them as real 'toolroom' stuff.

Derek's book is based on a number of well proven designs of his, some of which match very closely the best commercial products. my own copious old file of notes has largely been replaced by this book.

here are are a few tips for making them which Derek covers.

1. an ordinary centre drill of smallest size is too big for starting the very small drills required for injectors for 5" and 3.5"g locos. something else is required (and there are better and easier methods than suggested by Derek)

2. you need extremely sharp well honed lathe tools for the outside of the cones

3. you need a good set of quality small drills

4. the tapered reamers must be accurately made and when used impart a very good finish without any ridges. this is perhaps the hardest part of all

5. the second part of the combining cone needs a slight chamfer on the entry side (the 'secret Linden chamfer')

6. the entry to the delivery cone of bell-mouthed shape must be highly polished with no machining marks

7. the combining cone in 2 parts must be concentric after pressing in and have a constant taper

8. the check valve above the gap in the combining cone must seat very well. this can easily be tested, though i think the method i use of a rubber bung is better than Derek's cocktail stick method.

9. if the combining cone (in 2 parts) is wrong in sizes, concentricity, constant taper, or of positioning in the body, nothing you do by way of altering the steam cone or delivery cone will be of any effect.

if you can make the reamers to the required degree of accuracy, and decent cutting finish, then everything else is quite easy by comparison.

Don Young gives a very good description of making the 'bits' for his County Carlow injector in ME in april/may 1970, and although the injector the sizes of which detailed wont work, the manufacturing description is very good, and is worth having as a companion to Derek's book, as is the Basil Palmer articles of the Eric Rowbottom injectors in ME 1976. i never quite got the hang of the Laurie Lawrence articles.

Don's description is pretty much how Arthur Grimmett made them commercially with a rather well worn old Atlas lathe with turret and stops, but if you leaned on the saddle or turret they would shift over what i thought at the time was about 1/4"! John Helps has restored this rather lovely lathe of Arthur's back to 'spec' and is a lathe im incredibly fond of!

before making mniature injectors for Kennions and Reeves, Arthur was a toolmaker at De Havilands. his own injector reamers were 5 sided affairs made out of HSS on his rather basic grinder. John Helps has all of Arthur's tooling. it is rather nice to think that although Arthur died in 1997 at the age of 86, his Atlas lathe and tooling are still being looked after by a fellow model engineer and miniature steam enthusiast.

cheers,
julian

[ejparrott]

I wouldn't worry about the imperial sizes Roger, everything is convertible in the end!

Imperial is here to stay and should not be avoided. I've been working on a shaft for a very large company, the drawings are only months old, and are imperial. Interestingly what appears to be a mating part is in metric!

[Roger]

I'm going to put Derek Brown's book on my Christmas list, it sounds very useful. The new booklet is very good in my opinion, an worth reading even though it doesn't cover making them, only the design parameters. What's particularly interesting though, is that there's no mention anywhere of the angles of the cones. He mentions the ratios of each cone's minor diameter compared to each other, but it would appear that the actual angle isn't important. Doubtless there are practical reasons for the angles being as they are, but he doesn't cover that.

There's an interesting part where he talks about the purpose of splitting of the combining cone into two parts....

"In order to start and injector, the initial excess of steam and the trapped air within the system have to be released"

If I understand this correctly, I think I would put that differently. I'd say that if there's too much restriction in the combining cone, the steam creates back pressure that prevents the ejector from drawing up water.

He does mention that once started, the valve is there because there's an area of low pressure that tries to draw air into the gap, and this would mix with the column of water. If the column of water isn't solid, it won't enter the delivery cone. The same argument goes for using hot water, when the steam can't be completely condensed, and that's why they don't work when hot. ie the column of water is 'elastic' with trapped steam or air bubbles.

There is no mention of a chamfer on the second part of the combining cone. I think this is probably just a work round to accommodate small inaccuracies of alignment between the two cones. A tiny misalignment with the jet touching a sharp edge is clearly going to cause issues.

Another thing that I found interesting, is the comment that the 'chirping' sound so often heard is the instability of the jet and this can cause the injector to suddenly stop. He says that the sound should be continuous and that it's air being drawn into the final stage of the injector. I'm not sure I believe that, because where does it go? surely it's not going into the boiler? I thought the column of water had to be solid, and that seems to argue something else. Maybe the air forms a thin layer on the outside of the delivery cone? Who knows. He does say that it should be eliminated by the use of a check valve when employed with Steel Boilers (see final section), but as far as I can see, he doesn't mention it again. If it does indeed draw air into the boiler, I can see that this at high temperature could be a cause for corrosion.
Does anyone fit a non-return valve on the overflow?

Anyway, it's a thought provoking booklet and worth reading in my opinion.

[Deleted]

Just to say Julian that I find this thread very interesting and very helpful...I plan to make my own injectors so will take note of all info given...

cheers

Pete

[ejparrott]

Interesting, I always though the chirping was air being drawn into the water supply and was a bad thing and should not be happening!

[4fbuilder]

Hello Julian,

Thank you for starting a most interesting and valuable thread; I've a couple of Arthur's injectors here, still bagged up, one of which has never been used, the other being used very occasionally as a base line to prove pipework etc. He always had a few when he came visiting with the IoW Society to Beechhurst; we never had anything other than injectors on our locomotives they were that reliable. I do have here a set of Bill Carters drawings for his injectors, Bill was an excellent draughtsman and meticulous engineer and I suspect the articles by Laurie Lawrence and the sizes were based on Bill's findings also possibly the work of Jim Ewins, all of which when you go back in time were probably based on the designs of Eric Linden and Keiller. It has to be considered all of this work was carried out without the aid of today’s sophisticated machinery, tooling and measuring equipment and as you said on Arthur's old Atlas, on which we often compared notes as in those far off days as I was also the proud possessor of a clapped out ex lease lend Atlas lathe, gather up the backlash and off you go!

All of the books mentioned in the previous posts are invaluable to anyone with an interest, owning an injector or starting out to produce their own injectors; what an acheivement, when you have completed one, fitted it to your pride and joy, turned on the steam, popped the water on and it picks up first time.

Regards

Bob

[jma1009]

well i hope all this isnt getting too tedious or boring!

i agree with Bob Bramson as quoted by Roger about the cause of the 'chirping'. i quite like a chirping injector, but then i dont have a steel boiler to feed! in fullsize an extra check valve (spring loaded) is added to the overflow.

the 'chirping' is definitely not air being drawn in via the water supply. i have never found that a 'chirping' injector might suddenly stop.

if you have a 'chirping' injector put your finger over the end of the overflow pipe, and see what happens.

so far as taper angles are concerned, 9 degrees is the angle in the diverging taper of the steam cone nozzle, and 9 degrees is the taper inside the combining cone. Bob states 8 to 10 degrees for the diverging part of the steam cone, and in Fig.23 9 degrees is given as the combining cone taper by him. in practice you would obviously use the same reamer for both. Eric Rowbottom used a 10 degree taper for the combining cone on his high pressure injectors (ME May 1976)

the diverging taper on the delivery cone is subject to some debate and a quick look at Derek Brown's injectors on p.52 of his book shows 3 different angles on different sizes of injector. his 'medium' 26 oz per min injector has 13 degrees, his 18 oz per min 9 degrees, and the smaller injectors have 6 degrees. however he does state on p. 62 that 9 degrees will be ok if you dont want to make an extra reamer of 6 degrees

Bob Bramson says the angle on the delivery cone isnt critical but then recommends 6 to 8 degrees. Rowbottom used 6 degrees as specified by C.M. Keiller, these injectors being from medium to small.

i do not know myself why such a difference of angles for the delivery cone taper should exist, but i guess that it has more to do with making the length of the delivery cone fit into standard size bodies, but i might be totally wrong on this!

there has always been a bit of a debate about the entry side of the steam cone. Bob Bramson and LBSC said a plain drilled hole was sufficient without any need for a converging taper. however Derek Brown explains why he doesnt agree, and instead specifies a 13 degree taper. i have always used a taper on the entry side of the steam cone of 13 degrees, and every commercial product ive examined has such a taper.

for the various radii, bell mouths, and chamfers i use a clockmakers graver as this saves changing tools and upsetting everything and i find the use of a clockmakers graver quick easy and accurate. you dont have to go to some of the rather extreme 'toolroom' type lengths Derek Brown uses to make these gadgets.

cheers,
julian

[ejparrott]

Oct 15, 2015 11:45:23 GMT jma1009 said:

the 'chirping' is definitely not air being drawn in via the water supply. i have never found that a 'chirping' injector might suddenly stop.

And explains why after meticulous efforts to ensure air is not being drawn into the inlet, why it still happens!

[Roger]

It's definitely not boring Julian, I'm finding it illuminating and interesting. I think we may be looking at different sources of Bob Branson's though, I can't see the angle references you mention in the booklet. I accept that these are sensible values to use, so I'm not questioning those, I was just curious as to their omission in the booklet.

From the booklet, it's the ratios of the orifices that are absolutely critical, and he gives design criteria for those.

Bob indicates his preference for 'End reglation' of the ejector part rather than have the steam cone partially entering the combining cone to provide an annular orifice for the water. The argument centres around the issues of corrosion of the very thin edge of the steam cone that you find in annular regulation. The steam cone can have a flat end if end regulation is used. I like that idea.

[ejparrott]

+1 for the not boring!

I am intending on making my own for the MW so it's very useful stuff.

[jma1009]

hi Bob,

many thanks for your comments and reminiscences.

i was on the BeechHurst annual visit from the IOW from 1983 to 1996 although as the visit was June i missed a few years when i had to sit exams. i also remember the Southern Fed Rally being held at BeechHurst and them having a big bust up with Don Young and Gordon Chiverton on that occasion!

if you dont fit an axle pump or hand pump then you have to have injectors that are 100% reliable. we never gave this a second thought because Arthur Grimmett was always present with a spare in his grey toolbox or jacket pocket.

i have probably mentioned this before but in the summer of 1983 (my school holidays) i spent virtually everyday in Arthur's workshop using his Le Blond lathe making bits for my first loco. at that time Arthur had been coaxed out of retirement to make a final few batches of injectors for Don Young after Don had had a bust up with Gordon. so i was supposedly machining up wheel castings etc on the Le Blond, and Arthur was alongside shaking the first floor floorboards with the Atlas on top speed making his injectors with me learning more about how to make injectors than progress on my cast iron wheels! at the end of each day i wrote all i had learnt down onto paper. i was also shown all the drawings, which were quickly and carefully locked away if ever Don Young and daughter Elaine turned up.

the injectors were each tested on Arthur's gas fired test boiler in his kitchen at Apse Heath. Arthur had a lot of problems with those final batches because the brass castings Don had cast using Arthur's patterns for the injector bodies were hard and often porous. previously, Arthur had always used gunmetal castings same as Ted Linden, and the patterns were Ted's originals.

the stories about Arthur are legend and often very funny.

i did not realise at that young age how privileged i was to have been taught so much by Arthur.

cheers,
julian