Assembly and Alignment

This section is for anything directly related to the assembly of the locomotive, or any sections that needed reworking to make them fit.

1. Bogie Assembly    
All the various parts have been assembled and set into place followed by lowering the axle onto the springs and bolting the keeper plates into position. I have used a 2-part epoxy adhesive to fix the pressure pad cups to the axleboxes but they should be captive now. I have set MDF packing pieces between the axleboxes and the top of the horns to set the simulated ride height and I now need to adjust the guard irons for height because they are fouling the rail at present.
2. Wheel Assembly    
I made the keys for the main wheels from 1/8" square keysteel, just cutting them off at 5/8" long with a junior hacksaw. I then carefully dressed them with a needle file and emery cloth until they were a snug fit in the slot in the axle. I also had to make some brass spacers to put between the bearing face and the back of the wheels because the shoulder on the axle is not large enough to get a firm fit against when pressing the wheels on.
Next I pressed in all the crank pins, made sure they were all seated properly and finally pressed the wheelsets together. Everything was about half a thou interference fit and is unlikely to come apart in use. The effort needed to press them together was greater than my small driller could handle so I took all the parts to my local car workshop and got them to press everything together with their twenty-ton press in exchange for some universal beer tokens. The only modifications I had to make was to thin the axlebox covers down a little to ensure that there would be no fouling on the back of the wheels.
3. Finding Dead Centres    
I found an article by an American gentleman, Jeffrey G. Hook of the Deerfield and Roundabout Railway describing finding dead centres and I have followed his procedure on Britannia. If anyone wishes to read his original article I have created a link in the "Links" page. Packing pieces are set on top of the axle boxes to simulate the correct ride-height in normal operation. I started by making a wheel tram from a piece of tough 3mm dia steel and choosing reference points on the loco to locate the tram. I chose to use the expansion link brackets and set a small dimple in each one accordingly. I worked on the driver's side first, rolling the loco along the bench until the crosshead was at the far right in the slidebar brackets and blued up the top quarter of the driving wheel. Then I rolled the loco forward about an eighth of a turn of the wheel and zeroed the DTI on the front of the crosshead.
Using the wheel tram a line was scribed across the rim of the wheel. The loco was then rolled backwards until the crosshead "turned the corner" and came back to the same reading on the DTI. At this point, a second line was scribed on the wheel rim. I then repeated the operation for both lines to check that nothing had moved whilst rolling the loco back and forth. Using a digital caliper, I scribed a horizontal line across both perpendicular lines and set a tiny dimple at the two intersections. Using a pair of dividers, I scribed two more arcs to nearly meet in the middle and marked the horizontal to these arcs also.
An indent was made between the arcs on the horizontal line and when the wheel tram is placed in the indent and rested in the reference point, the wheel is exactly on back dead centre. To find the front dead centre, the procedure was followed again but with the crosshead at the front of the slidebars and the clock measuring from behind the crosshead. The loco was then turned round and the exercise repeated for the other side. Now that I can accurately find the dead centres I can move on with setting the return cranks to their correct positions. After that, the wheel tram will be put in a safe place to ensure it can't be found if it's ever needed again.
4. Setting the Return Cranks    
Following on from the procedure of finding front and back dead centres, the next process was to find the correct position for the return cranks. Once again I followed the information provided by Jeffrey G. Hook and started with the driver's side by setting the loco in neutral, the position where the die block at the end of the radius rod is exactly in the middle of the expansion link. Rocking the expansion link back and forth creates no movement of the radius rod in this position. A while back I made an adjustable, sliding gauge for marking out the coupling rod centres and this was used to set an approximate length of the eccentric rod as shown on the drawing. The gauge has a pair of bushes with 1/8" diameter holes, and two pins were made to fit into the return crank and the tail of the expansion link respectively, with 1/8" diameter spigots to fit the gauge.
The expansion link moves back and forth from vertical at the front dead centre position to fully tilted in one direction at middle dead centre, then onwards to vertical at rear dead centre and finally to fully tilted in the opposite direction at the other middle dead centre position. By setting a DTI on the expansion link, it is possible to measure the positions of both front and back dead centres quite accurately. When the two readings are the same, the return crank is exactly 90 degress to the centreline of the motion. Using the trammel from earlier, the loco was positioned at front dead centre and the clock set up as low down as I could. The loco was then wheeled backwards to the back dead centre position, keeping an eye on the clock. In my case, it came nowhere near contact with the stylus so the return crank clamp was loosened and the crank rotated until it just touched the stylus. The clock was moved until a zero reading was obtained and the loco was then wheeled forwards until I reached the front dead centre position again and a note made of the clock reading, which had slightly over-ran the FSD of the clock.
This back and forth process was repeated a number of times, adjusting the crank and the clock until I attained the same reading on the clock at both dead centre positions, probably about nine or ten cycles in all. When I got to the point where I was getting less than one thou difference on the clock for both front and back dead centres I called it a day and locked the return crank clamp up tight. I also took the opportunity to pin the lifting arms to the weighshaft with 3/32" taper pins, pinning the driver's side first. The link was set exactly in mid-gear and the lifting arm on the fireman's side adjusted until this link was mid-gear also. After marking the position, the firman's side was also drilled, reamed and pinned. The crank on the fireman's side was then set, leaving (I thought) only the eccentric rods to be made to finish the motion. Then it all went pear-shaped. I also realised that I would not be able to fit the eccentric rod pins after assembly so these were made and fitted before continuing.
5. Eccentric Rod    
The eccentric rod needs to be made to suit the rest of the motion and differs by a small amount on each side of the loco due to machining tolerances. The distance between the bushes in the eccentric rod are correctly spaced when the loco is placed on either dead-centre and there is no movement of the valve when the radius rod is raised and lowered in the link. The distance gauge was set up as before, the locking screws released and the link pin removed from the bell crank to allow the radius rod to be moved freely. The gauge was adjusted until the condition above was obtained and the gauge locked up. This distance between the bushes was accurately measured on the mill with the DRO and this reading used to make the first eccentric rod. The procedure was repeated for the other side, with a slightly different reading obtained for the second bush spacings.
6. Return Crank Pin    
Whilst setting up the second eccentric rod the link hit the expansion bracket (see item 8) and managed to move the return crank which meant I had to reset the position. Upon closer inspection it turned out that the driving crank pin was not as tight a fit in the wheel hub as I thought. From all that I've read, this appears to be one of the most important parts of the motion and I didn't fancy it moving again in service. I decided, therefore, to disassemble the driving axle and fix the two crank pins to ensure they stay exactly where they are but can be removed for replacement if ever neccessary. To this end, I have drilled and tapped an M3 hole at the interface, counterboring to take a cap-screw head purely because I have loads of them. I would have just used a grub screw otherwise.
7. Cross-pinning the Return Crank  
Because I could see no way of pre-determining the position of the return crank, I knew there would be a problem pinning them after final adjustment. I didn't want to use a hand-drill to drill through and I couldn't see how the assembly could be dismantled without losing the position. I decided to pre-drill part way through the return crank and, after setting, take the whole axle to the mill to complete. I started by setting up a piece of 3/8" dia material in the milling vice and using this as a reverse drill jig to drill a 3/32" dia hole through the first side only of the return crank. The cranks were then set as described previously and locked up tight with the clamp screws. All three axles were then released from the frames and the coupling rods separated from the front and rear wheels.
The centre axle, complete with the front coupler and the con rod, was then taken to the mill and set up on parallel risers to enable rotation of the wheelset. A 3/32" dia drill was used to get the pre-drilled hole in the crank vertical and the assembly clamped to prevent movement. The drill was then passed right through, followed by a 3/32" taper-pin drill and finished with the matching reamer. Finally, taper pins were trimmed to size and fitted to both cranks. The axle was then returned to the frames and the wheeltrain reassembled.
8. Expansion Link Brackets    
The driver's side wheel fouls the expansion link bracket if the suspension is bottomed out. The drawing shows a slope to the offending area but this is nowhere near enough to clear the wheel. No dimensions are given, either. I have used a 12mm endmill to remove a section of the bracket to ensure the wheel has clearance at all times. I later found that the fireman's side requires a similar modification. It also needs a section cleared away where the link fouls the bracket on the return stroke.
9. Cylinder Adjustments    
After bolting the cylinders on for what is, I hope, the final time I set about getting the length of the piston rods correct. I dismantled the end covers from the cylinders and removed the rings from the pistons to make it easier to work. I also found it easier to remove the slide bars and slide bar brackets leaving the crossheads floating. The driver's side piston rod was about seventy thou too long and causing fouling of the front cover so this was removed on the mill. I didn't want to break down the chuck that was on the table so used it as a vice instead. Another use for a 4-jaw chuck!
After checking that there was equal clearance at both ends of the cylinder, I removed the crosshead from the con rod and, with the piston rod located in the bore, the two parts were cross-drilled and reamed to accept a 3/32" dia taper pin. The fireman's side was much the same and it's probably where I made the crossheads to my own design rather than follow the drawing. On reassembly, I found that one of the slide bar brackets was still a smidgeon too low and putting undue pressure on the piston rod so the mounting holes in the frames were opened up a little more and the bracket clamped up tight where it wanted to naturally sit. It all slides as smooth as silk now.
10. Expansion Link Tailpin    
During assembly of the final part of the motion, it was obvious that the clearance between the eccentric rod and the con rod was going to foul the head of a clevis pin. Therefore, I decided to make a pair of short pins and hold them in place with tiny grub screws. A new hole was drilled and tapped M3 in the bottom of the expansion link, and the oil holes were put in at the same time - I had forgotten to do them previously.
11. Valve Guide Clearance  
Pushing the loco back and forth along the bench with the reverser in mid-gear, I noticed that the clearances of the die blocks at each end of the valve guide were different when on opposite dead-centres. On the driver's side the gap between the die block and the front endstop was 0.146" and at the other it was 0.122". To get these readings more equal, the combination link was removed and a small set put in it to compensate. On the fireman's side, it was the other way round, the front clearance was 0.101" and the back clearance 0.165". For this side, it was easier to move the whole valve guide nearer to the cylinder. This was achieved by taking the valve guide to the 4-jaw chuck and facing another thirty two thou off the rear face.
12. Setting the Valves    
Whilst making the cylinder valve liners and the bobbins, I deliberately made sure that the length dimensions were as accurate as I could make them. Because of this, I was able to make the preliminary valve settings quite easily. Working on one side at a time, I first screwed in the valve stem, without the valve bobbin, into the valve crosshead about half way and locked its position. Next, I set the loco on back dead centre using the wheel tram and accurately measured from the front of the valve liner to the front of the valve stem, noting the dimension. The loco was than rolled forward to the front dead centre position and the measurement repeated. Taking the average of these two measurements gives the exact mid-point of the valve travel and the loco was rolled backwards until exactly at this midpoint. Because I know the overall length of the valve liner and the length of the bobbin, I can calculate the position of the front of the bobbin relative to the front of the valve stem. Without moving the loco, the valve stem was then removed and the locknuts and bobbin loaded on.
After much fiddling around, I managed to get to the point where the bobbin could move radially about the valve stem but with only about a thou end-float and the distance from the front of the bobbin to the front of the valve stem within a couple of thou of the calculated dimension. The valve assembly was then reloaded to the cylinder and the valve stem screwed into the valve crosshead until the bobbin measured 1.437" from the front of the valve liner. The valve stem was then locked in this position and the valve should now be at the correct position. This procedure was then repeated on the other side. I don't know whether this is the correct way to set the valves but, as the short video on the "Overview" page shows, after lashing up a temporary air supply, the loco worked "straight out of the box" so to speak. It's all very tight and a bit jerky at the moment but I am going to set up a more substantial air supply and set it running for a few hours to try and run it in. After that, if the valves need tweaking, the motion should run a little more freely.
13. Smokebox Rivets  
The smokebox rivets have now been fitted with the exception of the top and bottom ones at the front. These will be round-headed screws and are to retain the smokebox door ring in position but allow removal at a later date if required. The door ring is a good fit but not tight enough to stay there on it's own. First, all of the rivets were fitted in the traditional manner - my fancy home-made rivet squeezer was not up to the job although it works well with softer copper rivets - banging into pre-formed countersinks on the inside and then dressed back to flush with sanding drums in the Dremell. Next, the door ring was pushed into position, lining up the scribed marks at the bottom, and the assembly mounted on the mill. The two holes were then spotted through followed by drilling for 10BA and tapping freehand. The first picture shows an offcut of a RSJ that I spotted in the skip at the local forge and this came in really handy for supporting the smokebox during drilling.
14. Next Item ...