Kevan
Seasoned Member
Posts: 138 
|
Post by Kevan on Feb 8, 2021 22:45:40 GMT
I remember when I was learning to fire full size locos the water valve always needed adjusting to dry up the overflow irrespective of boiler pressure.
That seemed the same across all the locos I worked from Barclay pugs to a black five. Are you expecting too much for scale injectors?
|
|
|
|
Post by Roger on Feb 8, 2021 22:54:53 GMT
I thought I'd add a curved base to the inside of the delivery end of the development Injector. I didn't like that being flat. I guess it ought to have a peak in the middle too, to guide the flow outwards, but that's a stetch too far.  20210208_112333 by Georgia Montgomery, on Flickr I've been doing a few experiments using pockets to give the flow more room to expand into, with some success. This latest iteration isn't the keenest to pick up, but it does go from 90psi to 40psi with a dry overflow once it starts. It will go down to 30psi with water regulation, but not below. There are a couple of things at play here which are fighting each other, and it involves the overflow valve. The overflow valve has to be lightly sprung else it will settle at its lowest position and stay there. However, even a light spring requires a significant amount of pressure to open it. That's not a problem, there's tons of Steam pressure and it always opens. However, that pressure adds back pressure that the Ejector has to overcome, else the Injector won't start. If you open the steam valve before water is flowing through the Injector, it tends to rapidly heat the body, suppresses the water back down the feed pipe and you get Steam out of the overflow. If the valve wasn't there, it would probably just pick up. I'll have to test this tomorrow if I remember! The valve is a pretty good seal, so although it does allow a little water to flow with the Steam turned off, it's not a lot, and if that large Development Injector body gets roasted, it doesn't cool quickly. Obviously that's not going to be such an issue on the scale Injector. Now, to get more Ejector suction, the overflow holes have to be more generous or numerous. However, doing that makes the flow more unstable. So it's a juggling act between maximising the Ejector effect while minimising the instability, and the two tend to require changes that adversely affect the other! Anyway, I'm gradually making progress, I just need to try to optimise these two conflicting properties. It may be that I can change the overflow springing so that it only brings the valve up to a point where the suction and reverse flow of water will lift it onto the seat. That might make all the difference because it could potentially reduce the back pressure and make starting easier. Out of interest, I have found that if I carefully turn on the Steam just a little while the water is on, the Ejector works fine and it causes plenty of water out of the overflow. Once that's started, I can then fully open the valve and away it goes. That's not really a satisfactory way to leave it, but it's a handy thing to know if the injector needs cooling down.  Cone39 by Georgia Montgomery, on Flickr This is how that design looks in practice. There's still plenty of scope for experiment, it's just a matter of being patient and trying to read the signs as to what's going on.  20210208_220509 by Georgia Montgomery, on Flickr |
|
|
|
Post by Roger on Feb 8, 2021 22:59:29 GMT
exactly Roger, I would say that it probably goes down to around 50 or 60 before regulating with the water valve. I will try to run an experiment for you... Chris. Hi Chris, I was about to way that's not very low, but then you're starting at a much higher pressure. To be honest, if it was bullet proof and dry from 90-40psi I'd say it was job done. If I can get it to go to 30psi regulating the water then all well and good. I'm pretty sure there's more to be had from the design with more experimentation. |
|
|
|
Post by delaplume on Feb 8, 2021 23:00:47 GMT
I remember when I was learning to fire full size locos the water valve always needed adjusting to dry up the overflow irrespective of boiler pressure. That seemed the same across all the locos I worked from Barclay pugs to a black five. Are you expecting too much for scale injectors? Yes-- same for me on 3205---Manor and Hall locos....Even my brand new 5 inch 1366 Pannier needs a slight tweak on the water valve for optimum performance..... I must admit to quietly thinking the same re}-- scale injectors....After all the water itself isn't scale...and at these microscopic sizes it might well start to behave different to expectations ??........... Just a thought.. |
|
|
|
Post by Roger on Feb 8, 2021 23:01:53 GMT
I remember when I was learning to fire full size locos the water valve always needed adjusting to dry up the overflow irrespective of boiler pressure. That seemed the same across all the locos I worked from Barclay pugs to a black five. Are you expecting too much for scale injectors? Hi Kevan, All I'm going by is what I know is possible from the injectors that have been made to date. They certainly can and do run with a dry overflow over a pretty wide range of pressures, so I think that's what I ought to be aiming for. It's worth going the extra mile to figure out just how good I can get it, having come this far. |
|
|
|
Post by delaplume on Feb 8, 2021 23:04:21 GMT
I remember when I was learning to fire full size locos the water valve always needed adjusting to dry up the overflow irrespective of boiler pressure. That seemed the same across all the locos I worked from Barclay pugs to a black five. Are you expecting too much for scale injectors? Hi Kevan, All I'm going by is what I know is possible from the injectors that have been made to date. They certainly can and do run with a dry overflow over a pretty wide range of pressures, so I think that's what I ought to be aiming for. It's worth going the extra mile to figure out just how good I can get it, having come this far. Agreed----crack-on Mc Duff !! |
|
|
|
Post by chris vine on Feb 8, 2021 23:31:05 GMT
Hi Roger,
You mention starting the steam slowly to establish a flow.
I spent many hours on the footplate of steam locomotives, large and larger, travelling all round the Ukraine some years ago. The drivers operated the injectors in the same way on all of them (probably same style of injector).
They had a lifting lever as a steam valve: They would lift it just a crack until they heard something, or maybe sensed it, or maybe just knew to give it a second or two. Then they would whack it up to the full open position to operate.
You are seeing something similar on yours. So that, in my opinion, would not be a negative in the design/operating characteristics of an injector.
Valve: You mention the lightness of spring required. Yes, you have a problem in that the head of water to push it open is tiny in your model - only a few inches of head. In the full size engine, it is several feet. So yes, I think you should look at making the spring super light, or not there at all?!
On a normal model injector, I think I am right in saying that the effective spring is just the weight of the ball. So, you might look at a design where the valve is similar, IE using gravity to shut the valve. Gravity is a very reliable force!!!
Or can you make the valve so light, and maybe with a chamber around it to give it a bit of a pop or servo action, so that the sucking above it causes it to lift up (pop up) and seat. Vacuum (atmospheric pressure, sorry to my physics teacher there) holding it shut.
Hope this helps.
Chris.
|
|
jma1009
Elder Statesman
Posts: 5,922
|
Post by jma1009 on Feb 8, 2021 23:41:49 GMT
Of course you need to regulate the water input for lower steam pressures as Roger's end regulation or annular gap is not then optimal.
What you would regard as an acceptable 'range' without regulation is what this is possibly all about.
I would want something that works from 90 psi down to 30 psi, then down to 20 psi with regulation of the water inlet.
We don't know yet whether Roger has a water valve on the inlet to the water into his test set up!
|
|
|
|
Post by Roger on Feb 9, 2021 0:02:28 GMT
Hi Roger, You mention starting the steam slowly to establish a flow. I spent many hours on the footplate of steam locomotives, large and larger, travelling all round the Ukraine some years ago. The drivers operated the injectors in the same way on all of them (probably same style of injector). They had a lifting lever as a steam valve: They would lift it just a crack until they heard something, or maybe sensed it, or maybe just knew to give it a second or two. Then they would whack it up to the full open position to operate. You are seeing something similar on yours. So that, in my opinion, would not be a negative in the design/operating characteristics of an injector. Valve: You mention the lightness of spring required. Yes, you have a problem in that the head of water to push it open is tiny in your model - only a few inches of head. In the full size engine, it is several feet. So yes, I think you should look at making the spring super light, or not there at all?! On a normal model injector, I think I am right in saying that the effective spring is just the weight of the ball. So, you might look at a design where the valve is similar, IE using gravity to shut the valve. Gravity is a very reliable force!!! Or can you make the valve so light, and maybe with a chamber around it to give it a bit of a pop or servo action, so that the sucking above it causes it to lift up (pop up) and seat. Vacuum (atmospheric pressure, sorry to my physics teacher there) holding it shut. Hope this helps. Chris. Hi Chris, That's interesting, one day we must have a long chat about your escapades on the footplate! I suspect they have the same issues as I have because they almost certainly have a lightly sprung valve on the overflow. It might just be something I have to live with, but it would be better if that wasn't the case. The spring is incredibly light, but there's nothing to stop me from setting it so it's not fully closing the valve. I don't think I can go much lighter. I could always try it without a spring. There must be a position from which it will suck the valve shut. I guess the larger the valve the better, I might be able to make the hole in the seat slightly bigger, but there's no much room. The idea of gravity closing the valve is certainly appealing, but that's very difficult to arrange within the confines of a scale body. It might already have some degree of 'pop' in the action because the valve diameter is pretty close to the bore of the housing. Again, it's something that can be experimented with. This is also slightly different in the scale body, so I ought to go back to trying a plausibly good cone in that before going too far. Thanks for these thoughts, it all helps. It's easy to get tunnel vision and it helps to step back and look at the bigger picture. |
|
|
|
Post by Roger on Feb 9, 2021 0:08:20 GMT
Of course you need to regulate the water input for lower steam pressures as Roger's end regulation or annular gap is not then optimal. What you would regard as an acceptable 'range' without regulation is what this is possibly all about. I would want something that works from 90 psi down to 30 psi, then down to 20 psi with regulation of the water inlet. We don't know yet whether Roger has a water valve on the inlet to the water into his test set up! Hi Julian, Hopefully I can get it to work down to 30psi without regulation. I assume you mean that it's bone dry do that, no dribbling allowed? So far dribbling hasn't been an issue. Yes, I've had a water valve on it for ages, I have mentioned it and shown it in the videos too. I fitted a new one recently which is more like the one on the locomotive in terms of size. So far these most recent cones are looking the most promising, but there's still some way to go to get the best out of it. One thing I have noticed is that these cones are working with significantly smaller regulation gaps, and that seems to help. They are a little more tricky to set the gap just right, but the smaller gap allows them to work down to much lower pressures. |
|
|
|
Post by delaplume on Feb 9, 2021 3:44:35 GMT
Hello Chris Vine,
Yes, that's how the Ex-BR Drivers on the SVR taught us to operate the injector as well ie}---- Water full on and visually check for flow......crack the steam valve on the manifold open a tad---- count "two--three" and then open fully ( about 2 full turns ) making sure that the handle was pointing downwards.....This was done to quickly show which valve was open as the fitters would have adjusted the handle to be vertically upright when shut....
I use the same idea on my own locos with all control handles upright when shut....easy to just cast a glance and know instantly what's going on....
Similarly when in the Engine Room during my RN days, the dials would be rotated such that all needles were pointing upwards at correct operating temp. or press..........so when it came to doing the hourly log book entry you only needed to cast a glance to know that all was OK....or immediately see when a reading was out.....
|
|
|
|
Post by Roger on Feb 9, 2021 22:57:11 GMT
I've spent much of the day experimenting with the more exotic designs of the kind I showed last time, and I've concluded that they all suffer from one important issue, stability. Some are more stable than others, you can hear it in the noise the flow makes, and you can see it in the movements of the pressure gauge and whether it wants to deliver cleanly into the Balanced Clack Valve. So, I've been looking back at other more simple designs that more closely emulate the separate cones, trying to understand why the ones I've tried have performed so badly. One of the thoughts is that the holes actually need to be a little larger than the ones I'd tried because some of the gap width is lost to the curved edges of the overlapping holes. I've just tried this version, and the results are a lot more encouraging. I think it ought to pick up a little more easily, and that might involve either moving the holes slightly upstream, or making them a little larger. One of the things I've changed is the way I'm drilling the 8 holes compared to before. I'm now drilling four holes at 90 degrees and then the other four starting at 45 degrees. That way the drill is always either in solid material or it's breaking into the existing holes equally on either side. I think this is creating a better edge. This version has the small parallel section where the throat is drilled to 0.66mm after the tapers have been reamed. This gives a more consistent size and it's easier to see the results of the changes are. Although I've pocketed the Delivery overflow holes, I'm not convinced that this is necessary. However, there's no harm in doing that for the time being.  Cone41 by Georgia Montgomery, on Flickr I need to run a few more tests on this with slightly different regulation gaps to see what difference that makes. However, this ran from 90-40psi with it being near silent from 70psi onwards. That's a good indication as to how stable the flow is. Usually it also seems to herald the fact that it will soon drop out, but that didn't happen. Having reached 40psi, regulating the water got it down to 30psi dry until it dropped out again. I then managed to get it to pick up again and go down to 25psi, so this is looking much more promising. Note that I've only put a very small radius on the inlet at the moment. So there are a lot of things to subtly change. Hopefully I'm finally on the right track and I can put this to bed. |
|
jma1009
Elder Statesman
Posts: 5,922
|
Post by jma1009 on Feb 9, 2021 23:44:11 GMT
Hi Roger,
Your last post and pic looks very promising, so fingers crossed!
You are doing far better than with some commercial jobs bought by Ron Hancock, Brian Leach, and Chris Vine!
|
|
Gary L
Elder Statesman
Posts: 1,208
|
Post by Gary L on Feb 9, 2021 23:47:52 GMT
I've spent much of the day experimenting with the more exotic designs of the kind I showed last time, and I've concluded that they all suffer from one important issue, stability. Some are more stable than others, you can hear it in the noise the flow makes, and you can see it in the movements of the pressure gauge and whether it wants to deliver cleanly into the Balanced Clack Valve. So, I've been looking back at other more simple designs that more closely emulate the separate cones, trying to understand why the ones I've tried have performed so badly. One of the thoughts is that the holes actually need to be a little larger than the ones I'd tried because some of the gap width is lost to the curved edges of the overlapping holes. I've just tried this version, and the results are a lot more encouraging. I think it ought to pick up a little more easily, and that might involve either moving the holes slightly upstream, or making them a little larger. One of the things I've changed is the way I'm drilling the 8 holes compared to before. I'm now drilling four holes at 90 degrees and then the other four starting at 45 degrees. That way the drill is always either in solid material or it's breaking into the existing holes equally on either side. I think this is creating a better edge. This version has the small parallel section where the throat is drilled to 0.66mm after the tapers have been reamed. This gives a more consistent size and it's easier to see the results of the changes are. Although I've pocketed the Delivery overflow holes, I'm not convinced that this is necessary. However, there's no harm in doing that for the time being. Cone41 by Georgia Montgomery, on Flickr I need to run a few more tests on this with slightly different regulation gaps to see what difference that makes. However, this ran from 90-40psi with it being near silent from 70psi onwards. That's a good indication as to how stable the flow is. Usually it also seems to herald the fact that it will soon drop out, but that didn't happen. Having reached 40psi, regulating the water got it down to 30psi dry until it dropped out again. I then managed to get it to pick up again and go down to 25psi, so this is looking much more promising. Note that I've only put a very small radius on the inlet at the moment. So there are a lot of things to subtly change. Hopefully I'm finally on the right track and I can put this to bed. Hi Roger It’s looking good. Drilling after reaming, how do you get rid of any burrs protruding into the cone? Do you put the reamer through again? Gary |
|
|
|
Post by Roger on Feb 10, 2021 0:01:11 GMT
I've spent much of the day experimenting with the more exotic designs of the kind I showed last time, and I've concluded that they all suffer from one important issue, stability. Some are more stable than others, you can hear it in the noise the flow makes, and you can see it in the movements of the pressure gauge and whether it wants to deliver cleanly into the Balanced Clack Valve. So, I've been looking back at other more simple designs that more closely emulate the separate cones, trying to understand why the ones I've tried have performed so badly. One of the thoughts is that the holes actually need to be a little larger than the ones I'd tried because some of the gap width is lost to the curved edges of the overlapping holes. I've just tried this version, and the results are a lot more encouraging. I think it ought to pick up a little more easily, and that might involve either moving the holes slightly upstream, or making them a little larger. One of the things I've changed is the way I'm drilling the 8 holes compared to before. I'm now drilling four holes at 90 degrees and then the other four starting at 45 degrees. That way the drill is always either in solid material or it's breaking into the existing holes equally on either side. I think this is creating a better edge. This version has the small parallel section where the throat is drilled to 0.66mm after the tapers have been reamed. This gives a more consistent size and it's easier to see the results of the changes are. Although I've pocketed the Delivery overflow holes, I'm not convinced that this is necessary. However, there's no harm in doing that for the time being. Cone41 by Georgia Montgomery, on Flickr I need to run a few more tests on this with slightly different regulation gaps to see what difference that makes. However, this ran from 90-40psi with it being near silent from 70psi onwards. That's a good indication as to how stable the flow is. Usually it also seems to herald the fact that it will soon drop out, but that didn't happen. Having reached 40psi, regulating the water got it down to 30psi dry until it dropped out again. I then managed to get it to pick up again and go down to 25psi, so this is looking much more promising. Note that I've only put a very small radius on the inlet at the moment. So there are a lot of things to subtly change. Hopefully I'm finally on the right track and I can put this to bed. Hi Roger It’s looking good. Drilling after reaming, how do you get rid of any burrs protruding into the cone? Do you put the reamer through again? Gary Hi Gary, I first drill the centre hole right through, then I drill all of the cross holes, and finally I ream the tapers. Any burrs are in the cross holes. I inspect those, and if necessary run the drill carefully by hand to clean them up. Usually that's not necessary with Brass though. If I do put a drill through, I spin the reamer by hand in there to make sure it hasn't pushed any sharp edges into the tapered hole. |
|
|
|
Post by Roger on Feb 10, 2021 20:21:05 GMT
I've been subtly changing this basic design to see if I can get it to pick up more readily at the highest pressure. The ejector stage wasn't creating enough suction to overcome the heat in a warm body. So I've tried moving the 0.8mm holes slightly upstream, then increased their size at that point through 0.85mm to 0.9mm and that does seem to have helped without destroying the low pressure performance. This is the next iteration with 0.9mm holes and a further 0.1mm upstream position. Although I've shown a small radius on the inlet, there's actually just the lightest breaking of the sharp edge at the moment. I want to optimise the setup before exploring a radius. What's interesting is that when it's working with restricted water below 40psi, it doen't just drop out. It seems to struggle and momentarily drop out, only to pick up straight away again. It does that several times before finally giving up and needing to adjust the water again. The last one ran down to about 25psi with a little fiddling.  Cone45 by Georgia Montgomery, on Flickr Anyway, I'm now on Cone 45 and have used up the last of the second batch of cone bodies. I've set up the Mill as a CNC Lathe again, and I'm just finishing off the 6th one of the third batch. I've only got enough material to make eleven this time, but I've got another couple of half Meter lengths on order that ought to be here soon. I reckon another 20-30 cones ought to be enough to finish the development. |
|
|
|
Post by delaplume on Feb 10, 2021 21:11:17 GMT
Hi Roger, Quote}---"What's interesting is that when it's working with restricted water below 40psi, it doen't just drop out. It seems to struggle and momentarily drop out, only to pick up straight away again." The GWR 8x and 10x live steam injectors have that hinged flap which gives a similar performance when air gets drawn into the feedwater supply.....looks like you're on the right track then ?? I've posted these next two just by way of a refresher and for those who would like to understand a bit more...I hope you don't mind ??}------- www.bluebell-railway.co.uk/bluebell/locos/vt/3217/injector.htmlNotice the quoted delivery rate for a 10x @ 40 Gall./ min  |
|
|
|
Post by Roger on Feb 10, 2021 21:31:15 GMT
Hi Roger, Quote}---"What's interesting is that when it's working with restricted water below 40psi, it doen't just drop out. It seems to struggle and momentarily drop out, only to pick up straight away again." The GWR 8x and 10x live steam injectors have that hinged flap which gives a similar performance when air gets drawn into the feedwater supply.....looks like you're on the right track then ?? I've posted these next two just by way of a refresher and for those who would like to understand a bit more...I hope you don't mind ??}------- www.bluebell-railway.co.uk/bluebell/locos/vt/3217/injector.htmlHi Alan, It's interesting that these designs use a mechanical arrangement for allowing the excess Steam and Water out for starting. My guess is that you can't do what I'm doing when you get to as big a size as this. Maybe nobody has tried it, it's simple enough to add a valve that there was no incentive to try anything else. Who knows. Does anyone know who designed the original model Injector arrangement that we all know so well? I feel sure it was inspired by the sort of design in the picture. I guess they discovered that the Overflow Valve wasn't strictly necessary for our purposes, so they left it off. Personally I think that was a mistake, you don't want air to be drawn into the boiler, it compromises the performance when the Water temperature is high. |
|
|
|
Post by John Baguley on Feb 10, 2021 23:31:16 GMT
Does anyone know who designed the original model Injector arrangement that we all know so well? I feel sure it was inspired by the sort of design in the picture. I guess they discovered that the Overflow Valve wasn't strictly necessary for our purposes, so they left it off. Personally I think that was a mistake, you don't want air to be drawn into the boiler, it compromises the performance when the Water temperature is high. There's an article in Model Engineer for September 1898 describing the 'Vic' type injector being sold by a W. Eaton which is the common type we know today. I suspect the actual design goes back before then.
John
|
|
Gary L
Elder Statesman
Posts: 1,208
|
Post by Gary L on Feb 10, 2021 23:51:03 GMT
Hi Roger It’s looking good. Drilling after reaming, how do you get rid of any burrs protruding into the cone? Do you put the reamer through again? Gary Hi Gary, I first drill the centre hole right through, then I drill all of the cross holes, and finally I ream the tapers. Any burrs are in the cross holes. I inspect those, and if necessary run the drill carefully by hand to clean them up. Usually that's not necessary with Brass though. If I do put a drill through, I spin the reamer by hand in there to make sure it hasn't pushed any sharp edges into the tapered hole. Sorry Roger, I didn’t read you closely enough. If I was paying attention I’d have seen “the throat is drilled...after reaming”. Get back to the good work! Gary |
|
