Help with 'Speedy' valve gear and other issues...

[jma1009]

5 thou annular gap for a 16 oz per minute injector. Somewhat different for the Chiverton type with shorter and therefore less in diameter steam cone and Gordon's larger combining cone throat.

[Roger]

Jan 3, 2021 22:47:57 GMT jma1009 said:

Hi Roger,

Thanks for the reply. Your throat sizes are to proven sizes.

You might, for testing like to go down a size on the steam cone throat, just a few thou, to get your upper range - you can always open out the throat of this if I am wrong.

You could also go up a few thou on the end of the combining cone throat as part of a vigorous and methodical testing regime.

You could separate the delivery cone from the combining cone in your test body - and I personally would not rule out the traditional press or very close sliding fits if appropriate to design.

Concentricity is absolutely paramount throughout.

Cheers,

Julian

Thanks for confirming that Julian, and suggesting some useful things to try. At the moment it's barely working at all, so I think I need to figure out what the main reason for that is. It might be that the drillway style of gap in the Combining Cone is to blame for that. I'll have to run quite a few experiments to see what's happening. I'll double check all of the throat diameters that I've ended up with too to make sure there are no errors there.
The Concentricity issue is the one of the driving forces behind making a one piece cone if at all possible. The spacing of the cones is another. If I can eliminate those two problems then I think Injectors would be much easier to make.

[92220]

Hi Roger.

This is one very interesting experiment!! If concentricity and spacing is critical (which it obviously is), and that is the reason for trying to combine all cones in one, could you not produce individual cones with very accurate plug and socket ends to fit tightly into each other, to form one long assembly? After all, this would be done on the lathe so would automatically be dead concentric on each cone, so machining accurately to diameter would be the only critical operation in fitting them together. The cones could have end features to fix the spacing too, which, again, could be produced on the lathe and so be concentric and accurate on length. Does that sound feasible?

Bob.

[tony9f]

Morning Roger,

Just thinking about your cone design and it occured to me that just possibly the problem may revolve around the overflow cross drillings. I daresay you've considered this and I'm no expert but in most injectors a lot of emphasis is placed on creating smooth flows and the use of lead ins. Is it possible that the cross drillings are causing turbulence and upsetting the flow as it is not possible to create the 'Linden secret' in your design and disrupting the injectors ability to pick up?
I would suggest that larger size injectors are tolerant of these sorts of issues as presumably the boundary layer issue doesn't scale down and therefore multiplies any causes of malfunction.

I may also be talking absolute twaddle.

Tony

[Roger]

Jan 4, 2021 9:36:17 GMT 92220 said:

Hi Roger.

This is one very interesting experiment!! If concentricity and spacing is critical (which it obviously is), and that is the reason for trying to combine all cones in one, could you not produce individual cones with very accurate plug and socket ends to fit tightly into each other, to form one long assembly? After all, this would be done on the lathe so would automatically be dead concentric on each cone, so machining accurately to diameter would be the only critical operation in fitting them together. The cones could have end features to fix the spacing too, which, again, could be produced on the lathe and so be concentric and accurate on length. Does that sound feasible?

Bob.

Hi Bob,
I see what you're getting at, but I don't think that would be any easier to make than conventional cones. Ok, the reaming of the holes in one assembly makes that easier, but making them in the first place is still complex. You would also probably need to mark the cones to make sure they went back together in exactly the same angular position they were reamed in.
I think it would work, I'm just not convinced that it would be significantly easier to make.

[Roger]

Jan 4, 2021 10:18:22 GMT tony9f said:

Morning Roger,

Just thinking about your cone design and it occured to me that just possibly the problem may revolve around the overflow cross drillings. I daresay you've considered this and I'm no expert but in most injectors a lot of emphasis is placed on creating smooth flows and the use of lead ins. Is it possible that the cross drillings are causing turbulence and upsetting the flow as it is not possible to create the 'Linden secret' in your design and disrupting the injectors ability to pick up?
I would suggest that larger size injectors are tolerant of these sorts of issues as presumably the boundary layer issue doesn't scale down and therefore multiplies any causes of malfunction.

I may also be talking absolute twaddle.

Tony

Hi Tony,
I think it's perfectly possible that the cross drillings are a problem, however you can certainly do that on the delivery end where you traditionally have a large bell mouth.

There are many different hole drilling arrangements which I can try to tease out what, if anything, can be made to work. The current arrangement has ovelapping holes which leaves a wavy edge. I could use smaller holes that do the same and use two rows. I could potentially use three of four rows of tiny holes. There's a lot of scope for experiment.

Clearly what happens when the flow is required to jump some sore of gap is very important. At the moment, all we know is that a straight edged gap certainly works. I'm of the opinion that the Linden Secret just accommodates any misalignment of separate cones. I'm not convinced it's necessary when the alignment is perfect.

It's certainly an interesting topic, and your ideas are as good as anyone's in my opinion.

[andyhigham]

You could make a teeny weeny boring bar to cut a straight edge and maybe even put a radius on the corner

[tony9f]

Hi Roger,

What I was trying to say is that because the cross drillings are perpendicular to the cone there may be eddys set up as the flow goes over them potentially deflecting the jet alignment out of true to the bore or maybe even causing it to lose velocity. I am no fluid dynamics expert but sometimes bizarre things can occur against all logic.

Tony

[Roger]

Jan 4, 2021 15:32:18 GMT tony9f said:

Hi Roger,

What I was trying to say is that because the cross drillings are perpendicular to the cone there may be eddys set up as the flow goes over them potentially deflecting the jet alignment out of true to the bore or maybe even causing it to lose velocity. I am no fluid dynamics expert but sometimes bizarre things can occur against all logic.

Tony

Hi Tony,
I see what you mean. There's usually a perpendicular edge though when there's a gap, so I thought it wouldn't be any different. What is different it that the edge isn't a clean straight one on both sides of the gap. There are scallops where the holes overlap. I can arrange it so that the holes have gaps between them though, but then you get into issues with getting enough cross sectional area for the water and steam to exit. I'll be surprised if there isn't a combination that works as well as a plain gap.

I'd certainly agree that just because something seems sensible, this is what will actually happen. There's a lot to learn for sure.

[Roger]

Odd jobs today on the test injector...

This is a new smaller radius former for 5mm Copper tube...

20210104_191709 by Roger Froud, on Flickr

... so I could make the overflow and outlet pipes.

20210104_192416 by Roger Froud, on Flickr

This is one of the covers for the overflow valve.

20210104_192434 by Roger Froud, on Flickr

This is one of the bodies with a standoff at the bottom to get it in at the right height for Silver Soldering.

20210104_203209 by Roger Froud, on Flickr

The cross hole is going to be tapped so that the Copper pipes can just catch a turn or so to make Silver Soldering easier. There will be a lot of parts to assemble in one session.

The drill shank is to stop the Copper tube from collapsing while it's being threaded.

20210104_204234 by Roger Froud, on Flickr

I've actually made two of these, but unfortunately one of them shot down the back of the headstock on the lathe and I can't retrieve it.

20210104_205057 by Roger Froud, on Flickr

[tony9f]

Hi again Roger,

Your scientific and methodical approach to this design is impressive to say the least and hopefully you will crack this one. Injectors have always carried a certain mystique to them akin to the dark arts which are unwarranted provided you have the means to build them accurately.
I am building my injectors to DAG Browns 12oz design to fit the 10X injector body and my own class K body. I am using the same cartridge in both of them thereby standardising the cones.

The issue with the scalloped edges is that unlike the traditional design which has a sharp edge on the upstream side of the gap and it just may be that as the jet jumps the overflow gap there may be a slight eddying created which causes the jet to start expanding into the overflow holes more than would happen in the conventional design. It may be that the turbulent jet is hitting the downstream scalloped edges and breaking it up so that by the time it gets to the delivery section it is losing energy which could be causing the pressure drop, if you see what I mean.
This is pure supposition and as I said before could be total rubbish but please don't take this as in any way a negative comment, I would hope that I might have contributed in some way to making it successful.

Tony

[Roger]

Jan 4, 2021 22:49:19 GMT tony9f said:

Hi again Roger,

Your scientific and methodical approach to this design is impressive to say the least and hopefully you will crack this one. Injectors have always carried a certain mystique to them akin to the dark arts which are unwarranted provided you have the means to build them accurately.
I am building my injectors to DAG Browns 12oz design to fit the 10X injector body and my own class K body. I am using the same cartridge in both of them thereby standardising the cones.

The issue with the scalloped edges is that unlike the traditional design which has a sharp edge on the upstream side of the gap and it just may be that as the jet jumps the overflow gap there may be a slight eddying created which causes the jet to start expanding into the overflow holes more than would happen in the conventional design. It may be that the turbulent jet is hitting the downstream scalloped edges and breaking it up so that by the time it gets to the delivery section it is losing energy which could be causing the pressure drop, if you see what I mean.
This is pure supposition and as I said before could be total rubbish but please don't take this as in any way a negative comment, I would hope that I might have contributed in some way to making it successful.

Tony

Hi Tony,
Can you fit the overflow in the 10X scale body? There's certainly more room than the 8X which has a smaller diameter. That means the cones have to fit directly into the injector body, there isn't room for a sleeve. The body bore is 6mm and that just leaves room for M7 x 0.75 (fine) threads on the end.

What you suggest on the drilled hole edges may well be true, I'll have to try a few different arrangements to see if it can be made to work. I might have to make a two part cone, or one with a deep slot on each side to check that it works with a conventional gap if it doesn't work. I'm introducing too many variables at once at the moment, so it's hard to know what to blame.

You model things in your head the same way that I do by the sound of it. I think that's a very useful way to try to picture what might be happening. All contributions are gratefully received. I can't think of everything, so it's helpful to have prompts from every angle.

[delaplume]

Just to add a tangential element---- have you taken into account variation in the working temperatures ??.....as I understand it one of the key factors to get an injector to either start or work properly is the temp. difference between the water supplied and that being delivered from the injector to the boiler clack...

[Roger]

Jan 5, 2021 1:50:49 GMT delaplume said:

Just to add a tangential element---- have you taken into account variation in the working temperatures ??.....as I understand it one of the key factors to get an injector to either start or work properly is the temp. difference between the water supplied and that being delivered from the injector to the boiler clack...

H Alan,
Yes, I'm keeping a note of the temperature of the feed water. The water has to be able to absorb enough heat to completely condense the Steam in that very short length, and that's the limiting factor with regard to temperature. Presumably hot water injectors have a shallower angle and a longer Condensing Cone to give hotter water more time to achieve that?

[tony9f]

Hi Roger,

In answer to your question the overflow is mounted on the back side of the injector body as there isn't room to incorporate it within the body. The 10X injector is a Doug Hewson item with internals as designed by DAG Brown, so I'm working to presumably tried and tested methods. I must stress that I have not ever built injectors before and haven't even tested them as the time for their use is a long way off yet, I just wanted to see what I could achieve.
The 10X body has room for an 8mm bore into which is fitted a cartridge reamed 5mm containing the cones. These are designed to give 12oz/min but the overflow valve is of the usual type. On my class K injector, the overflow valve is fitted in the same position as prototype.
As you say I do tend to model mentally and envisage what might be happening inside any given item, but of course conclusions can only be drawn based on experience.

Tony

[Roger]

Jan 5, 2021 11:45:01 GMT tony9f said:

Hi Roger,

In answer to your question the overflow is mounted on the back side of the injector body as there isn't room to incorporate it within the body. The 10X injector is a Doug Hewson item with internals as designed by DAG Brown, so I'm working to presumably tried and tested methods. I must stress that I have not ever built injectors before and haven't even tested them as the time for their use is a long way off yet, I just wanted to see what I could achieve.
The 10X body has room for an 8mm bore into which is fitted a cartridge reamed 5mm containing the cones. These are designed to give 12oz/min but the overflow valve is of the usual type. On my class K injector, the overflow valve is fitted in the same position as prototype.
As you say I do tend to model mentally and envisage what might be happening inside any given item, but of course conclusions can only be drawn based on experience.

Tony

Hi Tony,
Ok, that makes sense. An 8mm bore makes life much easier. There's an injector that uses this extra space to fit disc valves for the overflow, eliminating the need for an overflow on the back.
I don't think there's anything especially difficult about making injectors to standard designs as long as you're very careful with the concentricity issues, taper angles and setting lengths. Obviously you have to be careful at every step, but I'm sure it can be done with a bit of patience.

It's interesting to experiment though, so hopefully we'll all learn something from our experiences.

[Roger]

I wanted to press the overflow ball into the seat before Silver Soldering it into the body. I have to revisit this, but it's easy to do it now and it won't put any unnecessary force on the body.

So I quickly knocked up the pusher which has a countersunk hole to centre the ball. The big flat base makes it easier to get it to stay square.


20210105_123401 by Roger Froud, on Flickr



20210105_123517 by Roger Froud, on Flickr

I'm using the same tiny springs I bought for the scale injector overflows.

20210105_124457 by Roger Froud, on Flickr

This is a thin strip of 400 grit Wet & Dry paper I'm gingerly working over as much of the internal sharp edges as I can. The O-rings only have a light nip, but given half a chance a sharp edge will shave a bit of it off. Since they will be in and out like a Fiddler's elbow, I might as well try to make that as painless as possible.

20210105_160258 by Roger Froud, on Flickr

The ends of the overflow pipes are hanging on by just over a turn of thread, but it's enough to keep it where it needs to be while Silver Soldering.

I've added a single turn of 0.5mm diameter Silver Solder wire to each joint. The olive was done separately since that's easy to do that way and it's one less joint to think about during the big heat up.

20210105_162135 by Roger Froud, on Flickr

I've used plenty of Tippex internally to protect the inside of the pipes and the various threads. I've also added a bit on the outside of the body where I've got a bit more Silver Solder. It's going to be laying on its side both ways, so it might want to run around rather than go where it's intended.

I've also added the Brass barb to the water inlet pipe to help get that to seal. That worked well on the other one I made for the scale body adaptor.

20210105_170717 by Roger Froud, on Flickr

I decided to remove those two rear reflector blocks because I thought it might end up overheating the fitting furthest away from the camera.

You can see that I'm gingerly warming it up, trying to get as much heat into the body while avoiding the pipes as much as possible.

20210105_170925 by Roger Froud, on Flickr

I managed to Silver Solder the overflow pipes to the overflow bodies and the Steam inlet union, at least I think that was ok. It was upside down, so it may have melted and gone away from the joint.

The pipe nearest the camera turned round a little and that needs to be corrected.

Anyway, I couldn't get it hot enough that way around to do the rest without overheating the parts I'd already done, so I've turned it over and added some more flux since that had already been hot for some time.

This side makes it easier to heat the body by running the torch along the edge while keeping the flame away from the thinner parts.

The pipe near the front was given a support to sag onto if it felt so inclined. As it happens, it did just that so I didn't need to ask it.

20210105_171459 by Roger Froud, on Flickr

What a mess. It's hard to see what I've got until it's cleaned up.

20210105_171729 by Roger Froud, on Flickr

This is straight out of the Ultrasonic tank after 15 minutes.

20210105_174726 by Roger Froud, on Flickr



20210105_174743 by Roger Froud, on Flickr

It looks like it's got everwhere it's needed without making a mess so I'm really pleased with that.

A few years ago before I'd started this build, I'd have made a right dog's dinner of that! You live and learn, usually the hard way when it's me!


20210105_174755 by Roger Froud, on Flickr

Hopefully this will clean up well now and all of the threads and seals will be ok.

[Roger]

Ok, here's the cleaned up body and the rest of the parts with their O-rings in place. It all goes together nicely with a nip on the O-rings, so that looks promising.

20210105_205541 by Roger Froud, on Flickr

However, there's a bit of a leak on the valves, so they're getting another pressing...

20210105_213954 by Roger Froud, on Flickr

... twice in fact, but that's still not fixed it. I suspect the drilled seats aren't perfectly round, I should have bored them really.

Anyway, here's the ball on a stick and some fine grinding paste. Hopefully that will do the trick.

20210105_215606 by Roger Froud, on Flickr

[92220]

Hi Roger.

Just a thought. It might be a total no-no, but could you not use flat PTFE valves to seal (a disc on the end of a shaft), instead of balls? All you then need is a flat face to seal on, and being soft, the valve is almost guaranteed to seal. Adam, I think, uses this method for sealing some of his valves. I also used it on my cylinder drain cocks, and it sealed well.

Bob.

[Roger]

Jan 6, 2021 9:24:30 GMT 92220 said:

Hi Roger.

Just a thought. It might be a total no-no, but could you not use flat PTFE valves to seal (a disc on the end of a shaft), instead of balls? All you then need is a flat face to seal on, and being soft, the valve is almost guaranteed to seal. Adam, I think, uses this method for sealing some of his valves. I also used it on my cylinder drain cocks, and it sealed well.

Bob.

Hi Bob,
I could do that, but I don't think it's going to be necessary. I think they just need a light grinding in to make them seal. The problem is almost certainly that I didn't bore or ream the through hole, and that's not perfectly round. That means that when I press the ball home, the seat springs back more in some places than others. It's only a tiny amount, but it's enough to make it leak slightly.