|
Post by Roger on Oct 10, 2020 8:19:54 GMT
Hi Roger........I'm told that there's no such thing as Subtraction in Boolean Mathmatics ??...... OK, that's me way outside my comfort zone !!.......LoL !! Seriously though, that's a very impressive looking 8X injector....fingers crossed that it works ok...but I'm sure it will........ Hi Alan, I think it's the removal of the shape that's called the 'Subtraction'. It's the effect of the Boolean operation rather than the operation itself that's being referred to. I somehow doubt it will work, at least in the first itereation, I've changed so many things from the conventional arrangement. Not all of it's new though, combined Delivery and Combining cones have been tried successfully. Adam uses that in his scale 10x design. As far as I'm aware, nobody has made three cones in one piece and used drill ways for the overflow. I think the drilled overflows have been tried but not successfully. I haven't seen any designs with a single overflow valve, so that's completely novel. Anyway, it's all interesting and worth trying.
|
|
|
Post by delaplume on Oct 10, 2020 17:57:29 GMT
Absolutely !!...........anyway, you wouldn't be you if you didn't---eh ??.......In an earlier life I wanted to join the R&D team at BR-Derby working on the APT train.........In the end I couldn't and count that as a lucky escape !!
|
|
|
Post by Roger on Oct 10, 2020 21:23:23 GMT
A bit more progress on the prototype 8x scale injector. Here I'm facing off the water inlet flange with a 3mm cutter... 20201010_111221 by Roger Froud, on Flickr ... and then tapping the holes M1.4 Yes, I know these should be clearance holes with a nut and bolt, but it's going to be easier to assemble with just a bolt. I can always add a nut to the back on the ones that can be seen. 20201010_142924 by Roger Froud, on Flickr The end caps are secured by M7 x 0.75 (fine) threads. This end only has 3.5 turns, so I need to make sure I don't go too deep. I count the turns to do that while using the flat ended tap to finish it. Take a look at the gap between the halves of the fixture... 20201010_144207 by Roger Froud, on Flickr ... because it's gone in this view. I had my doubts about this method for getting the bore true to the outside, and didn't pay attention to the fact that this was going to get hot. I could have easily stopped and cooled it but I didn't think about it until it was too late. So the part softened the print and settled more firmly into it, releasing the pressure on the part and letting it float. I trued it up as well as I could and pressed on since it was either that or starting again. 20201010_150007 by Roger Froud, on Flickr The result is that the end of the hole isn't in the middle at the outlet end. I clocked it up to the bore and finished it slightly offset, but with hindsight it probably wouldn't have mattered if it was made true to the outside. 20201010_181509 by Roger Froud, on Flickr Anyway, this is just a proof of concept, so I've trimmed off the lugs and given it a quick tidy up. 20201010_220417 by Roger Froud, on Flickr 20201010_220502 by Roger Froud, on Flickr I've still got the top and bottom to machine, but I'm printing out a modified fixture which will hopefully allow it to clamp down better on the body and support the flanges more. I'll machine the rest of it and make the other bits and pieces shortly. So it's a mixed bag, but overall it's not a disaster even though I consider it to be scrap at this point. The main thing is to learn the best way to machine it and to check the dimensions. Hopefully it will prove that the design of the cones is right and that I'm not going to have to think again. I'm sure I can set it up better on the lathe to drill and ream the bore. I'm reducing the size of the roughed hole so there's a bit more metal to remove and I can use a rod through that when clamping the halves together and when setting up for machining the bore on the lathe.
|
|
|
Post by Roger on Oct 11, 2020 21:31:17 GMT
These are the final machining operations on the Steam inlet flange... 20201011_100443 by Roger Froud, on Flickr ... the output clack... 20201011_115028 by Roger Froud, on Flickr ... and the overflow flange and ball seat. 20201011_155341 by Roger Froud, on Flickr I realised that I'd got the location of the injector inside the fixture slightly offset, so I've fixed that, made a couple of other changes and that's printing out now. So now I've got a set of notes on how to go about this and all of the operations defined and tested. That's 31 separate rough and finish operations just to finish the various flange, 'O' ring, thread and bore details. Fortunately, on subsequent ones I can group many of these together in one Machining Operation Set so it can go from one to the other without intervention. It's all machined with a 3mm cutter so I can do that. Obviously the drilling and tapping is done by hand. Anyway, so far, so good. I need to get on and make the other ancillary bits and pieces.
|
|
oldnorton
Statesman
5" gauge LMS enthusiast
Posts: 696
|
Post by oldnorton on Oct 12, 2020 8:38:32 GMT
Lovely to see the process. I had thought that you were going to use the four small flanges as mounting points to a fixture for subsequent machining, but you have cut them off.
|
|
|
Post by Roger on Oct 12, 2020 11:31:57 GMT
Lovely to see the process. I had thought that you were going to use the four small flanges as mounting points to a fixture for subsequent machining, but you have cut them off. I did consider doing that, but decided that I could support the faces being machined better using a 3D printed nest. You certainly could do it that way. I've made sure that the part sits in the middle of the fixture so that I can use the wobbler on the fixture to find the middle in both directions. Even if the clam shells don't quite meet, the part sits equally, at least close enough to use that at the reference. In reality, you probably need +/-0.1mm accuracy and I'm sure it's better than that. Whichever way you hold it, the challenge is to reference the CAM outputs to the rather imperfect Silver Soldered assembly. It's pretty much the same situation you face when machining a casting.
|
|
|
Post by andyhigham on Oct 12, 2020 11:53:11 GMT
Am I over estimating the capabilities of your mill here? Could the tapers and gaps between the cones be machined "as one" with the body halves then a final ream after soldering the halves together?
|
|
|
Post by Roger on Oct 12, 2020 13:43:45 GMT
Am I over estimating the capabilities of your mill here? Could the tapers and gaps between the cones be machined "as one" with the body halves then a final ream after soldering the halves together? That's an interesting thought, I guess you could do that in principle, and it would be simpler to make. I think you would have to leave the cones solid and then drill and ream them after Silver Soldering. There would be no opportunity to clean up any burrs on the exit side of the holes, but that might not matter. I guess you could 3D print the body from Bronze complete with cones too, then machine the holes. The most compelling reason not to do it this way is that you only get one shot at it, and the whole thing is scrap if it doesn't work. Lining up the body precisely so that you can ream the holes from either end is a tricky problem. Making the body as a container means you can change the working parts, and that's really useful. Making three cones in one is much easier to align the reamed holes too. The arrangement I've opted for means I ought to be able to swap out the cones on the locomotive without removing the injector. That could prove to be really handy.
|
|
mbrown
Elder Statesman
Posts: 1,725
|
Post by mbrown on Oct 12, 2020 16:02:51 GMT
Surely one reason for not making it in the solid is that full size injectors have removable cones and that they do get removed from time to time for cleaning or replacement. For example, even on large injectors, the flimsy edge of the steam cone can erode and need replacement long before the rest of the injector is life-expired.
Surely the same applies in our sizes?
Malcolm
|
|
|
Post by Roger on Oct 12, 2020 20:02:54 GMT
Surely one reason for not making it in the solid is that full size injectors have removable cones and that they do get removed from time to time for cleaning or replacement. For example, even on large injectors, the flimsy edge of the steam cone can erode and need replacement long before the rest of the injector is life-expired. Surely the same applies in our sizes? Malcolm Hi Malcolm, You only have that wafer thin edge for the Steam Cone if you use Annular Regulation. If you use End Regulation you have a flat end to the cone so you don't have that issue. Full size injectors appear to only use End Regulation. I'm not sure where Annular Regulation came from, it seems to be the source of many of the issues we have with miniature injectors. I agree with the general point that it's handy to be able to take it apart. However, in practice you probably won't take an injector apart since getting it back together exactly in the right place with the separate cones is not easy. That's the reason I've opted for one single cone for three of them if that can be made to work. Then there's only two cones and one gap to deal with.
|
|
|
Post by chris vine on Oct 12, 2020 21:29:43 GMT
Hi Roger,
When you make the cones, I wonder if you could make them from bronze rather than brass? I say this because brass ones do wear or dezincify. I realise that for these tiny reamers and drills that bronze is more difficult to machine than brass, but I am thinking to the future maintainability of your masterpiece!
Chris.
|
|
|
Post by Roger on Oct 12, 2020 22:16:19 GMT
Hi Roger, When you make the cones, I wonder if you could make them from bronze rather than brass? I say this because brass ones do wear or dezincify. I realise that for these tiny reamers and drills that bronze is more difficult to machine than brass, but I am thinking to the future maintainability of your masterpiece! Chris. Hi Chris, Yes, I don't think Brass is really suitable. I've got some Leaded Bronze, albeit in large diameters which I might use. I've actually got some Fluorosint that I thought I might try for a bit of fun! I'm not sure where the wear comes from. Maybe that's from dissolved solids scraping away, or maybe there are cavitation issues. The idea with Fluorosint is that it's dead easy to machine and dissolved solids probably won't stick to it when it dries. Whether it's feasible is somewhat doubtful, mainly because I thing it will be too fragile where the rows of drilled holes are. Still, it would be interesting to see if it would work.
|
|
|
Post by delaplume on Oct 13, 2020 11:32:29 GMT
Hello Roger, As always a great thread with lots of inspiration to be gained.....And you say you have a day job as well ??--LoL !!... One of the drawbacks with reproducing full size accurately in miniature is that faults/problems found on full size tend to replicate themselves in our models---only more so given our relative giant size compared to the model...........A classic example was a chap named Barry Hares who back in the 1980's made a 1/5 scale Rolls-Royce Merlin engine from works drawings... I went to one of his first public demonstrations and remember the story he told of a lubrication problem he had unexpectedly come across during early running trials....No matter what he devised the problem persisted...Eventually he approached Rolls themselves ( this is pre-sell off days ) and asked for help...... To their credit they searched their archives and found that they too had the same problem which had been cured by using modifications "a"...."b".....and "c"... He tried it on his model and it cured his problem too !! www.macsmotorcitygarage.com/video-15-scale-model-merlin-engine/If you are planning to remove an injector at the track site AND also attempt to change cones as well then I would advise against such action..As you already are aware they are small, flimsy and easily damaged....apart from which the location of the injectors makes just removing them a "less than pleasant" job on a full size 57xx or 15xx or even a 1366, never mind the smaller versions.... Why not assemble 3 or 4 different cone permutations in bodies then take them along on a track running day ??..........This way it's just a straight swap over... and you can keep the fire going nicely at the same time too !! PS}---- For goodness sake don't scrap any injector bodies -- I'm sure there's a market out there for dummy ones for visual use only !! Alan
|
|
|
Post by Roger on Oct 13, 2020 12:08:38 GMT
Hello Roger, As always a great thread with lots of inspiration to be gained.....And you say you have a day job as well ??--LoL !!... One of the drawbacks with reproducing full size accurately in miniature is that faults/problems found on full size tend to replicate themselves in our models---only more so given our relative giant size compared to the model...........A classic example was a chap named Barry Hares who back in the 1980's made a 1/5 scale Rolls-Royce Merlin engine from works drawings... I went to one of his first public demonstrations and remember the story he told of a lubrication problem he had unexpectedly come across during early running trials....No matter what he devised the problem persisted...Eventually he approached Rolls themselves ( this is pre-sell off days ) and asked for help...... To their credit they searched their archives and found that they too had the same problem which had been cured by using modifications "a"...."b".....and "c"... He tried it on his model and it cured his problem too !! www.macsmotorcitygarage.com/video-15-scale-model-merlin-engine/If you are planning to remove an injector at the track site AND also attempt to change cones as well then I would advise against such action..As you already are aware they are small, flimsy and easily damaged....apart from which the location of the injectors makes just removing them a "less than pleasant" job on a full size 57xx or 15xx or even a 1366, never mind the smaller versions.... Why not assemble 3 or 4 different cone permutations in bodies then take them along on a track running day ??..........This way it's just a straight swap over... and you can keep the fire going nicely at the same time too !! PS}---- For goodness sake don't scrap any injector bodies -- I'm sure there's a market out there for dummy ones for visual use only !! Alan Hi Alan, Yeah, I've got some work to do from time to time to keep the Wolf from the door. That's a jaw dropping piece of work in the form of the Merlin, what an achievement! Getting injectors on and off at the track is more daunting than geting the cones out. Remember there are only two of them and a spacer. To be honest, I don't intend to take them off or dismantle them at the track. They're supposed to work and be reliable. I've got the axle pump to fall back on, so it's not the end of the world if I ended up using that. No, I won't be scrapping the prototype body. It might serve as a backup at some point, we'll see.
|
|
|
Post by Roger on Oct 13, 2020 20:27:20 GMT
Moving on to the overflow valve and elbow at the bottom of this sectioned view. (Note:- this is an old section with different internals) Sectioned injector V1.0 by Roger Froud, on Flickr I don't have much 16mm Phosphor Bronze bar, just a 40mm length which is too short to hold in the chuck. Here I've machined a register in the end of a piece of Steel Bar, notched the end to allow Silver Solder to get into the joint, and set it up on the lathe so it's pressed in true with some flux in the gap. Note:- The notches also let the steam out so it doesn't explode when heated! Guess how I know about that one... Then I wrapped a piece of 24% Silver Solder round it and got it really hot so that melted. 20201013_114839 by Roger Froud, on Flickr That was clocked up dead centre and the wobbler was used to find the centre in Y and the end. I always use the end as the initial reference, then move another 0.1mm over so I know it will definitely clean up the end. Once I've got that set up, I use the wobbler again on the face of the chuck and note the reading in case I need to find the end again when it's been machined away. 20201013_130237 by Roger Froud, on Flickr The first operation cleans up the top face of the flange and machines the stepped bore to the full depth. That gets tapped from the other side later. 20201013_153544 by Roger Froud, on Flickr The flange is 3D machined from either side so I don't have to set it up with that vertical. I'm using a 1.5mm cutter to rough it out and a 1mm ball nose to finish it off. 20201013_160944 by Roger Froud, on Flickr Here's a wobbly video of the finishing pass which is at 150mm/min with a 50micron step over (2 thou) 20201013_160951 by Roger Froud, on Flickr That was then flipped 180 degrees and the underside roughed out with a 3mm cutter. 20201013_201957 by Roger Froud, on Flickr The other half of the rear flange was machined with the same programs as the top since the centre of the flange is Z0, ie at the centre of rotation. 20201013_203939 by Roger Froud, on Flickr
|
|
|
Post by David on Oct 13, 2020 22:02:03 GMT
Hi Roger........I'm told that there's no such thing as Subtraction in Boolean Mathmatics ??...... LOL, something has finally appeared in this thread where I feel on equal footing. I also laughed at Roger's comment about 'this is what I come back to when leaving a job' and thought what I come back to is broken cutters, smoke, and huge cuts through the whole width of the table.
|
|
|
Post by Roger on Oct 14, 2020 9:49:17 GMT
This was a bit painful to do with the large spindle, it really needed much higher RPM. It's a 1mm PCB 2 flute cutter which was needed to get into the tight rear corner of the body profile where it meets the rear flange. I could only take 0.1mm deeps cuts removing 0.2mm off the profile and 20mm/min so that took a while. 20201014_080809 by Roger Froud, on Flickr The bottom face and blend with the flange was 3D machined with a 1mm ball nosed cutter and the thread tapped M6 x 0.75 (fine) 20201014_090412 by Roger Froud, on Flickr Then it was just a matter of standing it up on end. Even though I've lost 50mm in the Z-axis due to leaving the High Speed Spindle in place, there was still enough room to reach this easily now I've put that stubby 4-jaw chuck on the 4th axis. 20201014_094624 by Roger Froud, on Flickr It was possible to reach the flange with a 1.6mm PCB drill for the fixing holes. You can just see that the bottom flange isn't fully machined the edges have been missed. That's because the Parallel Finishing operation stopped at the end of the radius. It does that because a vertical edge is problematical. I've changed the program to rough that out on the next one. 20201014_094619 by Roger Froud, on Flickr For the moment, I've just used the parting tool to remove the bulk of that, I'll finish this one by hand. So here it is having been parted off. I'll hold that in the vice so I can bore the hole in the back. That will have to wait until I've finished machining the next one to this stage. 20201014_100941 by Roger Froud, on Flickr
|
|
|
Post by delaplume on Oct 14, 2020 10:21:13 GMT
Hi Roger,
If we ever get back to "Normal"---- whatever that may be --- I can see the organisers of The Model Engineering Exhibition asking you to give "on a stick" demonstrations in the Machine shop section.....
Now I think you may have eluded to this next bit before but can you devote a small section as to the actual programming of a simple item........It looks all too easy from the photos...
|
|
timb
Statesman
Posts: 512
|
Post by timb on Oct 14, 2020 15:38:30 GMT
Hi Roger........I'm told that there's no such thing as Subtraction in Boolean Mathmatics ??...... Not strictly true..... If logic then yes, you are right, i.e. a boolean object is either true or false, 1 or 0 it canot be 'negative true' or 'anti true' as opposed to false.
HOWEVER
if counting in base 2 (binary) then -1 is acceptable as is -101010101 or any other combination of 0s and 1s with a - sign in front.
Just sayin!
;-)
|
|
|
Post by Roger on Oct 14, 2020 16:09:07 GMT
Hi Roger, If we ever get back to "Normal"---- whatever that may be --- I can see the organisers of The Model Engineering Exhibition asking you to give "on a stick" demonstrations in the Machine shop section..... Now I think you may have eluded to this next bit before but can you devote a small section as to the actual programming of a simple item........It looks all too easy from the photos... Hi Alan, Gosh, will there ever be 'normal' again as we knew it? Maybe, but that's probably years away. Sadly I can't demonstrate without the machine, so that's a bit of a problem. I suppose a presentation with some video might work. I'm not sure that enough people would be interested to be honest. Here's a link to another thread where I do explain how simple things are done with Alibre Design and the Mecsoft CAM package.
|
|