Tuesday, January 8, 2008

Tender Body

The original drawings showed what I think is the Hawksworth tender - I've previously made one in 5 in gauge, and thought the Collett tender would make an interesting change. It has a couple of major challenges - the top of the sides are flared, and also the rear corners of the tand are very rounded, so a tricky compound curve is created where they meet.

I took the basic dimentions from the drawings, and scaled the Collet specifics from a couple of photographs. The fact that the major lines of the tender matched up with the profile of the cab gave me some useful reference points.
Then there's a raised extension which fits on top. I left a tab which I curved to match the lower part of the tender, and riveted and soldered the two together.

The other tricky bit was the rivets -there are several hunders of them, and they're very prominent.

My solution was to make the tender tank out of thinner brass than usual -about 22 gauge. (Another reason for this was that I had a couple of sheets of this size availible!)
It made it easier to form the flared tops - I'd intended to form them around a 1 in dia steel bar, but eventually realised that the front edge of my workbench (originally intended as a kitchen work surface) was about the right radius. So I gently tapped tthe sheet to shape using a rubber mallet. I cut away the flared top in the region of the rear corners, and bent them to shape around a steel bar.

I took a lenght of 1 in dia copper tube and flared it (just like making the petticoat pipe.) It needed frequent annealing to get it to shape. I cut two 90 degree sectors out of this, and soft soldered them into the gaps in flares at the corners. A bit of excess solder and a file blended the curves together and it looked ok.

I worked out the rivet patterns from the photographs. To 'create" them, I made up a jig, to allow me to punch through from the inside. On the anvil of the jig, I formed recesses for the heads, using a ball nosed end mill. The upper leg of the jig had matching holes for the punch, and the two were kept in alignment by two pins in reamed holes. Finally, the punch had a step to fix the depth of the depressions.After a few trials on scraps of the same brass, I was surprised by how well this worked, creating convincing and even rivet lines, and very quickly.

A few real rivets held 1/4 x 1/4 / 1/16 brass angle to the bottom edges and to fix the coal plate. Once this was complete, it formed a very rigid structure, which was attached by 6 BA screws to the soleplate.

I made up the water filters from very fine mesh stainless woven cloth. The drawing showed these to be 1/4 in diameter; I chose to make them a good inch diameter to greatly increase the surface area and reduce the chance of them getting choked or reducing the flow rate. I soft soldered these together! I know stainless doesn't take solder, but it penetrated the mesh easily, and held the filters together, although they probably wern't soldered in the true sense of the word.

I painted the wheels, frames and body before final assembly.

Tender 1

The frames were already cut out, saving a bit of sawing. I checked that all dimensions were correct, and cut out the buffer beam and drag beam, from angle section. It never seems to be truely square, which shows up on the insides, when the angle fixings are mounted. So I lightly milled them all over, which got rid of the mill scale at the same time. Also the angle sections to lock it all together. I clamped the angle iron in place, and drilled through from the fames. Only then, I marked of the length of these pieces, taking the dimension from the frames rather than from the drawings - at least this way, they all fit accurately. I worked on a surface table, checking that all was square as I went along.

For the horn blocks, I milled the mounting surfaces to size, and riveted them to the frames, slightly overlapping the horn openings. then milled the horn gaps to size. This made sure that the horn slots were accurate and truly square. If I'd milled them completely to size before fitting, the chances are that riveting would have introduced at least slight distortion.

I then had to make up the spring hanger brackets. Twelve of them amounts to quite a lot of work. With small components, holding them during the work can be tricky. It took me a long time to realise that, by far the easiest way is to make up the part as far as possible before cutting them off the stock. So I made them as a 'production run', setting up only once, and drilling all at once - with a cutting allowance between components. Rather than trying to mark out and centre punch the fixing holes (always difficult on such a small area, I set them up the milling vice a fixed length from the end, and drilled them in turn - I still needed to start the holes with a centre drill - without a centre pop, drills, especially small ones, will wander quite a lot.
The photos show progress. I took less than two hours from starting to having them all fitted to the frames.

Before fitting the wheels, I tried the axles and axle boxes in the frames - any small mis-alignment would cause the axles and / or the boxes to bind. A small amount of filing of the axlebox / horn surfaces was needed on one axle to get them to run freely. I also filed a slight curve on the side faces of the axleboxes to allow a rocking movement, to allow for any irregularities in the track.
I then mounted the wheels. I'd been aiming for a slight press fit - some were ok, but others were more of a running fit! A spot of Loctite sorted that. Also, I trimmed a 1/32 off the end faces of the wheels on the centre axle to provide a bit of end float -this will help the 6 wheels to negotiate tighter curves.


I recently read the book 'How (not) to paint a locomotive' by Christopher Vine (ISBN: 9780955335907). It's full of useful information, and a council of perfection. I read a lot of the things I'd tried when painting the A3, and although I stopped short of what he'd done, I came to many of the same conclusions.

I sprayed using a Pasche internal mixing air brush, using automotive type paints (NOT two-pack cyano acrylics, which absolutely need an external air fed mask. I used water based paints (sounds odd, but I'm told are the replacement for cellulose paints) , thinned with cellulose thinners.
I started a very light coat of etch primer, just a token rub down, then grey primer. Finally, a few light coats of gloss.

Once its cured - at least a week - I flatted down the surface with 1200 grit wet and dry paper. I then used cutting paste in two grades, and was amazed at the quality of finish I achieved.

I've heard aerosol spray paints being recommended. I have tried them, and been totally disappointed with the results - so much so that I stripped it all off and started again. One of the problems was that the later coats would attack the earlier coats especially in areas that had been well sanded. Now, if you really wanted to create a crackle finish, that would be the way to go.

Its surprising how long it takes, and how little there is to show for the work. Also, there's very little worth photographing. I tried photos of the surface texture as work progressed, but they didn't show anything useful.

I left the completed painting for a couple of weeks to harden before doing the lining. Longer might have been better, but I had no problems. If the paint's well hardened, its easy to wipe off any lining errors and try again.
For lining, I use a draughtsman's ruling pen and Humbrol enamel. I take a few drops of the paint onto the lid of another tin, and add a drop (only) of Ronson lighter fluid - this totally changes the viscosity of the paint, but doesn't significantly dilute it. I made up templates for the cab lining, holding them off the paintwork with bluetak to prevent capillary action from spoiling the lining.
For lining that runs parallel with an edge, I use a set of bow-spring ink compasses, with the point set a long way out to act as a guide. With a little practice, and the knowledge that errors are easily removed, its quite easy. As soon as the paint stops flowing, I clean the pen and re-fill. Occasionally, I get a slight blob of extra paint where I start or stop - If they're only little, I tend to leave them, and trim away the excess with a modeling knife (scalpel !) once the paint's dry.

As you see, I've painted and lined all the loco, but still have the tender to complete building. Not very clever -it's much more efficient to paint it all at once, but I just wanted the loco complete and out of my workshop.

Monday, July 23, 2007

Work has progressed - the boiler is now complete. The assembly sequence I used was as described in Alex Farmers book on boiler making - and, as it happens, his was also a GWR boiler, although bigger.

As the assembly gets bigger, it always surprises me just how much heat it needs. I work outside, so can only work on wind-free ( and preferably dry) days. Also, I find it essential to use refractory bricks- without them, I'd never get up to temperature.
Even so, it needed two good sized gas torches, and occasional oxy-propane, to complete the assembly. I couldn't do this alone, and gratefully thank George and Russell for their help.

Where ever possible (and that's most of the time), we worked with pre-placed solder - its much easier than applying a stick, as it tends to melt - or at least sag - if you're not quick. We put rings of silver solder under the heads of all the rivet - stays in the firebox. As the solder flowed, each was grasped on the outside with pliers, rotated and pulled into place. This (I hope) helped the solder to penetrate. Also - and very important - old texts always advised reamed holes and minimum fit. Alex Farmer advises precisely drilled and chamfered holes to allow clearance for the solder to flow. Easyflow 2 type solder will fill a gap of a few thou, but not more, and does need a gap for capillary action to work. The pre-placed solder resulted in an excess of solder around the boiler, which is waste, and gives away the amateur status of its maker! - a small price to pay.

I fitted the firebox doorplate and backhead after the stays had been fixed internally. This made it very easy to work on the stays which are pretty inaccessible if you follow the traditional sequence of construction), but I was worried that it might be more difficult to fit these plates at such a late stage. In face, it was perfectly straight foreword - my only slight problem was that the rear foundation ring needed to be nearer to 5/16 than to the 1/4 inch specified in the drawings, so a new piece of copper was machined up to fit.

I made up plugs to blank all the bushes - in some cases, my blanking pieces were in fact the part machined finished components - minus their through -holes.

The boiler was tested - at 160 psi - twice its working pressure - and certified by the club boiler inspectors. Up to this point, I'd left all the stay - rivets t their original length, as, If I'd needed to redo anything major, the extra length is important to get the heat in.

So now they've been cut to length with a Dremmel (and lots of cutting disks) but it doesn't distort the stays the way cutters would.
I plan to re-tset it at 1.5 x pressure, to prove that I havn't compromised the seal on any of the stays.

I've been working on the cladding and false backhead. Older practice was to leave a 3 1/2 in gauge boiler painted but unclad, as far as I can see, with small fittings screwed directly into the boiler. I didn't fancy this one bit, and all my fittings will be attached to the cladding only. Also, I'll put a bit of thermal insulation between the layers - if only to protect the cladding from the stay heads! I remember doing heat transfer calculations a long time ago, and finding that it's the existence of an air gap that is the major factor in preventing heat loss, not its size.
Having tried the boiler between the frames, its now starting to look like a loco. My list of things still to do is only half a page!

Tuesday, June 12, 2007

Boiler 3

Some of my boiler making tools.

I formed the taper barrel using a set of bending rolls. I decieded that the chances of cutting the sheet to exactly the right dimensions was pretty small - the copper stretches when rolled - so I left the sheet oversize, and planned to cut it after forming. I know this makes for less convenient cutting, but at least it wont be short.
Just as well; after annealing and rolling, the taper ended up at the opposite end to what I'd planned - it didn't matter one little bit. So once it was almost to shape, I could work out exactly where to cut it.
I rechecked the dimensions once I'd got rid of the overlap, and found there was still a little to file off.

I drilled the strap and riveted it to one side, pulled the barrel up tight with a strap, and drilled and riveted the other side. Before doing so, all the surfaces were finally cleaned and covered with flux.
I laid two strips of (high temperature) silver solder at the centre joint, and added some more flux. I heated it mainly from underneath - with the strap at the bottom and once the solder flowed,added some to the rivet heads. Then I rotated the boiler so the strap was at the top and added more solder, again including the rivets. A few minutes in the pickle bath left me with a clean barrel, which I marked off and trimmed to size - in typical GWR fashion, all the taper is at the top, with the base of the barrel horizontal.

I could now do a trial assembly on most of the boiler; I'll use a few copper rivets to keep the components in place during soldering - but for the time being, I use 8ba bolts in these holes.

The trickiest bit for me was to drill for the cross stays needed on a Belpaire firebox. The problems are that everything slopes, and the reference dimensions are on the firebox inner, so not visible. In the end, I marked and drilled pilot holes on the girder stays, and put in a rod ground to a point at each end. Then I assembled the firebox and used a wedge to drive the rod into the outer walls, giving me centre pops. I transfered these to the outside with large springbow calipers - and a lot of checking. I drilled through the end holes, then could mark out the intermediate holes more sensibly. After a lot of checking, I had all the stay holes in place - as pilot holes. A couple needed 'adjusting' as the drill wandered on the inner surfaces -probably because the surfaces are not all horizontal. I had to enlarge some of the holes in the girder stay to get everything in line. Once I'd got them all lined up, I made up a small 'scribe' which mounted on the pilot rods, so that I could get the holes to their final diameter in the correct position. I hope the photos show what I did. The final diameter of the girder stay holes was 3/8 inch - I couldn't use a drill as it would just have grabbed and damaged things, so it was a matter of grinding with the Dremmel and filing.

I've left the holes in the outer wrapper just as pilot holes for the moment - in case I burn them; I would rather not have drilled them at all at this stage, but I couldn't see how to do them later!
Now I cut both the inner and outer wrappers to size - except for where the backhead joins the outer wrapper.

Its now just a matter of soldering things up. Its often recommended to do this in very few heats. I find it more reassuring to do more smaller ones - a very real problem is that once the job is up to temperature, the flux doesn't last long, and at that stage all you can do is stop.

Sunday, May 27, 2007

Boiler 2

I decided that I didn't want to cut holes in any of the plates until I'd formed the firebox wrappers - so I started on those next. One way to form them is to make a wooden template and beat the annealed copper around them. Or you can form the sheet using bars of the appropriate diameter, which was the way I chose to go ( not being an enthusiastic woodworker !)

Once annealed, the copper is incredibly soft, and needs little more than hand pressure to start to form it. I used various formers, as the pictures show. This firebox has a combination of inside and outside curves (most do) which makes it arkward.
I just bent a bit at a time, checking against the flanged plates. A good fit is important, as silver solder has no gap filling properties.
When the copper stops bending easily, its time to re-anneal. It took me five or six heats to get the inner firebox wrapper to my liking - less for the outer.
Incidentally, I deliberately left the sheets oversize; putting lots of bends and folds in sheet is difficult enough without having to worry about having enough length - the first fold on my outer sheet ended up 1/4 inch from where I'd planned it, so I was glad of the excess. I won't trim the excess length for a long while yet.

Boiler 1

I know several model engineers who wouldn't touch boiler making - it's not for everyone. You need several sources of heat - a lot of it, somewhere safe to work, and you need to be aware that - unlike most aspects of model engineering, it is possible to get into a situation where all your work - and expensive materials - are irrecoverably scrapped.
I've got to say, I like the challenge, and satisfaction that its 'all my own work'. Then there's the lack of waiting time. Whilst its also true that there's a significant cost saving, this is definately not a good reason for doing it.

I bought a kit of materials from Reeves- this comprises all the copper sheet, bar and tubes. You really do need to have the correct materials, all in good condition, and, these days, of traceable quality.
Copper is interesting to work with. After annealing, its amazingly soft, and can easily be shaped - until it reaches a point where it work hardens. Its then time to stop and re-anneal before continuing. This is how the flanged plates are formed. It requires patience, and some very substantial formers . The formers are a lot of work to make unless you have a serious bandsaw, and are of course only required during the making of the components.

The firebox front and rear plates (Inner and outer), and the front tube plate were bought already flanged.
The boiler barrel is tapered - 3/8 inch difference in diameter over 10.5 inches, which works out as a 1 degree taper. As is typical of GWR practice, all the taper is all on the top of the boiler, with the underside lying horizontal. The drawings (and materials supplied ) require the taper section to be made up from flat plate. But there are alternatives - it could be made up from extruded tube, slightly streatched over most of its length. I've never done this, so can't comment on how difficult it is. I did seriously think about abandoning the taper altogether, and using seamless tubing of the right diameter. The small amount of taper could be built into the cladding, and no one would be any wiser - and I doubt if it would affect the operation or performance of the loco. (Can anyone comment?)

To make the tapered barrel, I needed some reference dimensions. I started with the front tube-plate, which I chucked on a 3 jaw chuck, from the inside. I also made use of a tailstock centre to stabilise it. I then took extremely light cuts - like 1 -2 thou. to create a smooth surface ready for silver soldering. Turning copper like this can be tricky - its material properties are such that it can grab, pull out of the chuck, and tear. This is mad much worse by the fact that it cannot be held very securely in the chuck without being damaged. My solution is to turn the chuck by hand. Also, to use a suitable cutting lubricant. Paraffin is suitable - and so is WD40! It makes a big difference to how freely the metal comes off. just take off enough material to produce a smooth, round surface.There was a significant hollow in the plate, so I removed it and tapped it to a better shape before continuing, as I wanted to remove a minimum of metal.