LBSC Green Arrow
I looked in my selection of O rings and found some 1mm section rings that were just the right size to fit the exhaust outlets on the outside steam chest covers. All I had to do was machine a recess around the existing oval hole using a 5mm endmill. I made the recess 0.030" (0.75mm) deep so the rings would have plenty of compression on them when the flanges were tightened down.
The O rings fitted perfectly.
Flushed with success I thought why not fit them to the steam chest as well? I was always a bit worried about whether the joints around the oval holes would seal properly with the steam chest walls being so narrow. O rings would ensure that they did. The top of the steam chest wall was a bit too narrow for my liking so I machined the recess for that ring in the underside of the steam chest cover. The bottom of the steam chest wall was a bit wider so I did machine the recess into that.
All the steam and exhaust pipe joints will now have O rings so I won't have to faff around with gaskets or sealant.
Time to machine the valves to final width. I was a bit worried about clamping the valves in the mill vice because of the deep slot in them for the valve rod to pass through. If I was a bit heavy handed the valve would probably bend! To get around this I used a couple of gauge blocks that pushed into the slot with a nice tight fit. The valve was now solid and strong enough to be clamped across the vice.
It was an easy job now to machine the required amount off each side of the valve using the DRO whilst checking the width of the valve land with a micrometer (I had already decided to reduce the width of the lands from the LBSC dimension of 0.1563" which gives a negative lead of -0.004" to 0.1478" which gives a lead of about 0.005".
The valves were done so onto the piston rings. These are my pet design using two split rings side by side in the groove with the joints at 180° to each other. They've been working perfectly on Helen so that's good enough for me.
The PTFE bar was chucked in the lathe, faced off, drilled and bored and then the outside turned to diameter. The bore of the cylinders is 0.6875" so I made the outside of the bar 0.70" so the rings would be compressed when fitted into the bore. I bored the bar to 0.450" to give some clearance on the bottom of the piston slot which was 0.4375" diameter (I should have gone back and read my notes on Helen as I later found that I hadn't left enough clearance in the bore!)
I could have done with less stick out from the chuck but I don't like chopping bars into short lengths as you always seem to waste part of it when it becomes too short to hold.
Incidently, I've been meaning to try some of the polished carbide inserts that are meant for machining aluminium but can be used for brass and other non ferrous metals. Other model engineers swear by them. I've now got a couple and I'm quite impressed so far. They seem to cut brass very well and leave a good finish. I used one in the boring bar to bore the PTFE and one in the tool to machine the outside of the PTFE and the finish was excellent. Apparently, they cut mild steel quite well as well so I'll have to give them a try on that.
I parted off eight rings at about 0.065"to 0.066" thick (six for the three pistons and two spares incase I messed up!) The rings needed to fit into the 0.125" wide piston groove so this left some for finishing the width of the rings by rubbing them on emery paper.
To make it easier to hold the rings for finishing I found a scrap of brass bar that already had a spigot on it and turned the spigot so that the rings would just fit over it.
As well as holding the rings the 'handle' helped to maintain an even pressure over the whole surface of the rings as it was rubbed on the emery paper and helped to maintain an even thickness. Trying to hold the rings by just pressing on them with your fingers tends to make them uneven.
The rings were tried in the piston grooves until two of them were a nice fit side by side. They were then split by cutting a segment out with a sharp knife. The size of the split is not really critical so long as there is still a gap when the rings are fitted into the bore.
It was when I tried to fit the first piston and rings into the cylinder bore that I found out that it wouldn't go in! I realised that I had made the bore of the rings too small and the bore of the rings were touching the bottom of the piston slot before they could be compressed enough to fit into the bore. Oh x%&*! I couldn't enlarge the bores now as the rings were too thin to hold so rather than scrap them I did the next best thing and chucked the pistons in the lathe and took another 0.005" off the bottom of the piston groove!
It was a right pain trying to fit the pistons into the bores with the rings on them. You had to hold the piston in the end of the bore and then try and push the rings into the groove until they would slide into the bore. Very fiddly, especially with the recesses for the steam passages in the end of the cylinder. 'I need a piston ring compressor' I thought so I made one from another scrap of brass bar. It's just a ring with the bore the same size as the cylinder bore and a long taper machined in one end to compress the rings as you push the piston in.
You just push the piston into the tapered end, hold the other end against the end of the cylinder and then push the piston through into the bore. Dead easy.
In the photo below you can see the small gap left in the ring when it's compressed down to the size of the bore. The gap gives the ring a bit of room to expand when it gets hot.
It will be interesting to see what the rings are like in Myrtle Anne after all this time. I fitted this style of ring to her back in 2010 and she ran quite a few miles before getting stripped down for a rebuild by her new owner in 2012.
Back to Green Arrow and I'm now in the process of giving the cylinders a good clean to get rid of any dirt and swarf and then we can think about assembling them to try the chassis on air.
It's alive !
I've spent the last couple of days cleaning and assembling the cylinders and then fitting them to the frames ready for an air test. The port faces and the valves were all lapped on a surface plate by rubbing them on 1200 grade carborundum paper using WD 40 as a lubricant. This gave them a nice matt finish that should seal well. I used the Loctite 574 on the steam chest to cylinder joints and the end cover to cylinder joints. I didn't realise that it sets quite hard so getting them apart again if needed could be fun!
The valves on the outside cylinders were roughly set by eye i.e. they were set so that the ports just cracked open when the pistons were at top and bottom dead centres and then the steam chest covers fitted using a paper gasket for the time being. I didn't bother fitting the O rings between the covers and the steam chests for this trial run as the paper gaskets should seal well enough for that.
I had to just guess the setting of the inside cylinder valve and play with it later whilst the chassis was running on air.
I didn't want to get the big noisy compressor out so had a go with my little air brush compressor to see if that would run it. The chassis was quite tight in places so I didn't hold out much hope of it coping. It did show some encouraging signs when the wheels were turned over by hand and eventually it did try to run if the pressure was allowed to build up in the air tank. I did notice air leaking from a few places around the steam chest covers so I gave the screws a bit more grunt to seal the gaskets better. I then realised that I hadn't sealed the screws that held the steam manifold onto the top of the steam chests so quite a bit of air was leaking from those. A dab of threadlock sorted that out. With a bit more air going to the cylinders the chassis turned over a lot better and after adjusting the inside valve by trial and error it eventually ran quite well. There was no reading on the pressure gauge on the airbrush compressor tank so it must have been running on only a few psi. I was well chuffed!
I took a bit of video of it running in forward gear:
It sounds a bit noisy in the video but the knocking is the axle boxes on the driving axle rattling about in the horns. Putting some oil on them quietened them down considerably.
It runs in full gear and the next notch back on the reverser but it won't run in reverse at all. It completes half a revolution and then stops. I suspect that it's the middle crank not being set correctly that is causing that. I've adjusted the inside valve so it runs in forward gear ok but it's probably miles out for reverse. The outside cylinders are going in reverse but the inside cylinder is probably trying to go forwards! The only thing to do is take the axle out and check the quartering and see what's what. I did think that it may not be set right.
My suspicions about the motion brackets not being strong enough was proved correct as they move about quite a bit under the load of the valve gear. You can see that in the video. I'm going to have to strengthen them somehow to stop them flexing. There's play in the weighshaft bearings as well causing that to move about so I'll replace the bearings. I think the weighshaft itself could do with shortening as it's too wide and forces the radius rods outwards against the expansion links.
So, still a few things to sort out but overall I'm pleased. It's a lot better than when I first tried it on air last year!
I've been for my first Covid jab today so wasn't sure if I would be in the workshop tonight due to any after effects. So far, I've just got a sore arm but we'll see how I feel tomorrow! My brother had his last week and he felt awful the day after.
Anyway, I managed a couple of hours and took the crank axle out of the chassis with a view to trying to measure the angles of the cranks. I drew up a CAD drawing showing how the cranks should be set to show me what I needed to look for.
With the loco going forwards, the righthand crank should lead, followed by the lefthand crank and then the middle crank. The angle between the middle crank and the righthand crank should be 114° and between the middle and the lefthand cranks it should be 126°.
The first thing I noticed was that the cranks had been set up with the lefthand crank leading followed by the righthand crank and then the middle crank. I don't think this makes any difference to the valve timing for the middle cylinder so long as the angles are correct. It definitely doesn't make any difference to a two cylinder loco which crank leads. It's just a convention for the righthand to lead.
I didn't take any photos of the set up I used to measure the crank angles but I'll try and remember to take some tomorrow and add them. Basically I set the wheels up between centres with the righthand crank at the bottom and in line with the vertical centreline of the lathe axis and then used gauge blocks to measure the height of the other two cranks from the cross slide of the lathe. A quick CAD drawing and I had the results although they probably were a bit innaccurate due to the method I used. The centres in the ends of the axle are not brilliant so the axle may not have been centred properly. However, it would give me some sort of indication of the position of the cranks.
To my amazement the angles came out virtually spot on! I wasn't expecting that at all. They are certainly within a degree I would think even allowing for a bit of error in my measurements.
So, the angles are correct but the righthand and lefthand cranks are reversed in the 'firing order'. That obviously doesn't affect the running in forward gear as it runs quite well. I would have thought that if there was a problem then it wouldn't run in forward gear either. Maybe the fact that it doesn't like running in reverse is just that the valve gear is not very good?
I have modelled the valve gear as drawn by LBSC in the Wallace simulator and after getting rid of the negative lead it actually is pretty good. However, I haven't checked things like the length of the eccentric rods which can badly affect the equality of the forward and reverse running. More investigation is needed.
First a few photos of how I measured the position of the cranks on the driving axle.
The axle was held between centres in the lathe. The centre held in the collet chuck is 0.25" diameter, the same as the diameter of the crankpins.
The square was clamped to the top of the topslide with it's edge against the centre in the collet chuck. If the crankpin is held against the edge of the square then it will be vertically in line with the lathe centre.
The height of the middle crank from the topslide was then measured using gauge blocks with the outside crankpin held firmly against the edge of the square.
This was then repeated for the other outside crank.
I must admit that I find using gauge blocks very fiddly, especially if you are trying to make up a specific dimension! I did do a rough measurement with a height gauge first so I had somewhere to start from with the blocks.
Finally, I measured the centre height of the lathe from the top of the cross slide and then I had all the dimensions I needed to make a CAD drawing and work out the angles between the crank pins. As mentioned above, they were pretty much spot on.
Before going any further, I thought it a good idea to get some dimensions of just where the expansion link pivots and the weighshaft were in relation to the driving axle and the top of the frame. I may have done this before but it was a long time ago and I've forgotten the results!
The easiest way to do this was set the frames up in the mill and use the DRO to plot the positions.
I used a small 60° centre in the drill chuck to find the centres of the driving axle, the weighshaft and the expansion link bearing. It wouldn't be 100% accurate but it would be close enough for what I wanted to do.
I did the same for both sides of the frames. The weighshaft was within 0.020" between the positions on both frames but the distance between the expansion link bearing and the driving axle was 0.060" shorter on the lefthand side than on the righthand side. That meant the motion bracket was still in the wrong position even though I thought I had sorted it ages ago! I'm going to have to move it again.
Whilst I was at it, I measured the throw of the return cranks and also the length of the eccentric rods. The return cranks were set to exactly the same radius which was good news.
The actual measurements that I got are quite different to what they should be according to the drawings but so long as I know what they are that's fine.
I transferred the measurements I had taken to a CAD drawing and then worked out the dimensions needed to go into the Wallace simulator. The results were not good and showed that the eccentric rods were too short. Taking the righthand side as an example, the eccentric rod was 2.412" long but it needs to be 2.490" long, a difference of about 0.080". That may not seem a huge difference but it's enough to completely mess the valve events up. You can set the valve so that it will run in full forward gear and maybe the next notch back but the valve events in reverse are totally wrong. That's exactly what I found when trying to run the chassis on air.
The lefthand side is different because of the different position of the expansion link. To get that to run correctly the eccentric rod would need to be 2.440" long.
So, I am going to have to 'adjust' the length of the eccentric rods somehow. I don't really fancy having to make new ones from scratch so I'll see if it's possible to extend the original ones somehow.
What I did today though was strip the lefthand frame off the chassis and move the motion bracket so that it's more or less in the same position as the righthand one. At least then I can make both eccentric rods the same instead of having two different lengths.
I decided to modify the existing eccentric rods, at least for the time being, but I may replace them at some point in the future. At the moment I just want to prove that the rods being the wrong length is indeed the problem with the reverse running.
The easiest way to lengthen them seemed to be to cut the very end of the return crank end off and solder a new piece on made from a 0.125" slice of 0.25" steel bar. So, the end of the existing rods were cut off roughly with a saw and then mounted in the mill and the end trued up with a 0.25" endmill. I used the DRO to measure the distance between the centre of the endmill and the expansion link fork on the rod (2.490" as calculated from the simulator) and then the end of the rod was trued up by plunging down with the endmill which left a scalloped end that fitted the piece of 0.25" rod nicely and gave a good area for silver solder.
The small vice is one I use clamped in the bench vice and is great for holding small components. I just clamped it to the bed of the mill with a couple of strap clamps.
I made another jig to hold the rod and the new endpiece in alignment whilst I silver soldered them together. The blue on the steel base was a mask in an attempt to stop me soldering everything together!
The mask wasn't entirely successful as I still managed to solder the little spacer that held the piece of bar to the top of the bar! It filed off easily enough though.
I didn't bother too much about cleaning the rods up after soldering as I didn't want to spend the time until I knew that the modification had worked. With that in mind I just put everything back together to do another air test. I had to take the steam chest covers off again to reset the valves as the timing would have changed because of the different length of the eccentric rods but that was only a 10 minute job. I connected the compressor and fiddled about with the timing of the middle valve as that needed resetting as well. I eventually got it to run forwards reasonably well and then tried it in reverse. I was very pleased to see that it now runs in reverse as well as it does in forwards so problem solved.
I'm going to have to lengthen the reach rod as the mid position of the reverser doesn't tie up with mid gear on the lifting arms but that's no big deal. The reverser itself needs fixing better as at the moment it moves about on it's two fixing bolts. It needs at least three to make it secure.
I think we are now approaching the point at long last where I can strip the chassis down again for cleaning and painting. Hopefully, everything chassis wise has been sorted.
To be continued