Ian_C's workbench - P4 and S7 allsorts

[Ian_C]

...The Bangles, 2011.

Yes, finally I've made a start on the tender. There are still a few tidy up jobs on the loco, but I'm not in the mood right now and want to press on. I'm going to put the electrical pick ups on the tender so it helps a lot if all of the wheels contact the rails at the same time. As with the loco, the compensation arrangement on the tender is goofy. It's what you'd call 'statically indeterminate' if you wanted to impress somebody with your engineering knowledge. Or, to use the popular furniture analogy, if compensation is meant to give you the equivalent of a three legged stool on the uneven kitchen floor, the kit gives you a four legged chair that rocks from corner to corner. There's no cross chassis compensation. I'd imagined that I'd simply (!) redesign the inner chassis with sprung axle boxes or CSB...or something. Plus I was going to use insulated split axles and avoid wiper contacts. Ambition is an admirable thing. When I looked at the chassis etches and the instructions it was hard to see how it all goes together and comes apart again with brake linkages and water scoop gear strewn all over. Overwhelmed. Where on earth do you start?

I've decided to build according to the instructions up to the point I can see how all the gubbins fits together, then work out what changes I want to make. The tender instructions are patchy (but Mr MOK made that clear at time of purchase, so no surprises). The David Coulshed (circulating somewhere on WT) annotated version adds some useful notes and is worth obtaining. The Wild Swan 8F Loco Profile has no drawings of the 4,000 gallon tender, but they are present in Loco Profile 5 - The Mixed Traffic Class 5s - Nos. 5000-5224.

The initial build offers a respite from the mental toil of design and scratchbuild, and substitutes the tedium of filing off etch cusps.


The MOK tab & slot fabrication makes life very easy. There are a couple of misplaced tabs on one of the chassis end sub assemblies but it's easy enough to file them off and position the parts by eye. Very quickly you get to this stage...


Naturally there's a puzzle to solve. The lower crossmembers really do have to be removable to allow the fitting and removal of the inner working chassis. It took me a while to realise that one crossmember is flanges up and the other is flanges down. I discovered this when examining the drawings in the Wild Swan book, but it was stated in the instructions if I'd cared to read them! How to secure them without fasteners showing is the challenge. I don't know what the kit design intention was, but this is my solution...

There didn't seem to be any point in using all of the fixing points and doubling up the amount of work.

The inner chassis is the next job, along with a look at the compensation problem.

This episode brought to you with help from Bedsofaland, 9 Lazy 9, 2010

 

[Ian_C]

On the journey towards a fully compensated tender chassis, there were a few things I wanted to change on the compensation beams supplied on the kit. So far as I can tell it's not possible to remove the beams once the inner chassis is assembled. At least I couldn't work out how to do it. Once the side control tabs on the chassis etch (see a photo below) are bent into the final position to compensating beams are locked in. Sure, you can unbend the tabs but the half etch bend line won't withstand many bends before it breaks. Here's a way of providing an equivalent level of control of the beams and be able to dismantle them without bending or breaking anything.


The beam pivot in the kit is a length of 2mm diameter rod that's soldered in position across the chassis once all is assembled. I replaced that with a length of 2mm rod with a head soldered on one end and a 12BA retaining screw and washer threaded into the other. The pivot holes in the beams were opened out to 4mm diameter and some large top hat bushes soldered in. The bushes are a close working fit over the pivot rod, and provide the necessary side control of the beams. The beams are a double thickness laminate of 0.6mm N/S and plenty stiff enough laterally themselves.


The length of the bushes was calculated have them fit end to end across the inside of the chassis with a small clearance to allow them to rotate freely. The length of the pivot rod has to be adjusted carefully so that when the screw is tightened it doesn't pull the chassis in and pinch the bushes. Now it all goes together and comes apart easily enough.


Here's the offending bendy tag wotsit. The original intent is that once bent in to 90 degrees it just rubs the inside of the compensating beam and provides side control of the beam and axle bearing, which the original small pivot bush wasn't able to do. Once bent, there's no way of removing the beams. The outer diameter of the bushes is 4mm and I'm hoping that they don't clash with something unforeseen in the brake linkage or the water scoop gear. Flying blind somewhat...

Some useful reading here regarding compensation and suspension in general....

LOCOMOTIVE COMPENSATION by Brian Clapperton, ABC Gears
http://www.abcgears.co.uk/abc_gears_-_loco_suspension.pdf

The principles of model locomotive suspension
by Russ Elliott
Scalefour Digest 41.0 'The principles of model locomotive suspension'
Part of the Scalefour Digest, but principles applicable to any scale.


In the background right now - We are but hunks of wood - Little People 2012

 

[Ian_C]

The front axle sits in an awkward place relative to the MOK kit chassis structure. It wasn't immediately obvious how to fit a cross compensation beam in without making changes to a lot of chassis parts. Another of those jobs that looks straightforward, but turns out not to be. The problem was modelled on CAD, and after many schemes I ended up with this...


Because of the chassis structure in front of the axle (not modelled) a clevis arrangement can't practically be fitted in, so a bearing block pivoting on a cantilevered pin was the solution. The mounting block (yellow) locates up against a chassis cross member and is fixed with a couple of 10BA screws.The design was broken down into a few simple to make parts.

All from brass bar apart from the steel pivot pin. The pin is drilled through and threaded 12BA,and is a press fit into the mounting block. The axle bearing tube and axle pivot block were designed to self locate for soldering together. The reduced diameter section of the axle tube is 6.0mm so that the corresponding concave profile in the mating block can be cut with a 6mm ball end mill.


The axle tube was soldered to the pivot block. There's plenty of joint area so soft solder is fine. The tube was reamed through 3/16" to match the Slaters axle, and some of the centre section was milled away to reduce a drag (depending on where you shop, every little helps). Two small holes were drilled in the ends of the bearing tube for oiling the bearing (can't see them in this photo). A custom washer was turned for the retaining screw.


Here it is sub-assembled, front and rear views. The axle oscillation is limited by the distance between the two blocks. Doesn't really need to be that much, but y'know, round numbers in CAD. Pretty obvious how it works. Ignore the small aluminium block, that's just so the wheels clear the deck.


The point of no return now! The existing bearing lugs are cut off (alongside on the cutting mat) and the holes for fixing the mounting block are drilled in the chassis. The assembly is fixed to the chassis using two 10BA screws.


There's more etched gubbins to be soldered to this end of the chassis later. The tabs on the mating etched parts that locate in the slots in the crossmember will need to be shortened so that they don't project through and touch the mounting block behind.


Blimey! What a performance! But I now have a rolling chassis that keeps all six wheels on the railhead. In passing, there's a huge amount of clearance between the back of the wheels and the outside of the bearings on the two trailing axles. Some spacers needed next time I visit the lathe.

 

[Ian_C]

Yes, the water scoop gear turned out to be quite a challenge. Considering that most of the scoop gear and brake gear lurks in the dark void under the chassis it's represented in the MOK kit in considerable detail. The theory of authentic clutter dictates that if it's visible in a low angle photo it should be present on the model. The tricky bit is reconciling the assembly of the scoop gear with the need to assemble and dismantle the tender chassis for construction and maintenance. There may be a few different ways of doing this, but this is how I worked it out.

To start with, there are a lot of parts...


The instructions are a little foggy, but combined with the decent GA drawing of a 4,000 gal tender in the Wild Swan LMS Locomotive Profiles No.1 The Rebuilt Royal Scots it can all be worked out.

I turned up a couple of parts for the gear counterbalance weights and soldered them to the etched representation to make things a bit more 3D. The 'gear up' position of the deflector and the scoop were established from the Wild Swan drawing and the components were soldered up solid in that position. I'm sure there are WT'ers who could make the whole linkage work, and DCC it. The long shallow 'S' shaped link between the deflector gear and the scoop gear seems to be slightly too long to allow both to sit in the 'gear up' position. I pinned it at one end and let it sit where it wanted at the other, where it is soldered solid and out of sight. There's a long, cranked link from the main pivot shaft that runs forward to the operating gear at the front of the tender. Since most of it can't be seen, and it has a lot of getting in the way potential around wheels and brake gear, I chose to cut it short where it disappears from view behind the tender outer chassis frames. I'm not sure I got the detail right for that. I now think that the crank arm off the shaft is 'solid' and the end of the linkage rod is forked. I can live with it. I had to sort out the scoop assembly before I could establish the position of the two drop links that you can see waving in empty space. Much fiddling with lengths of wire and brass lacemaker's pins. About 3,000 words worth of photos below.

The scoop is supplied as two brass castings. In its retracted position the lower section of the scoop wraps round the outer chassis cross members and would make dismantling difficult, if not impossible. Notwithstanding the fact that the outer chassis cross members are removable, I just couldn't find a way to wriggle them in and out if the scoop assembly was fixed permanently to the inner chassis. Therefore the top of the big casting was drilled and tapped 10BA so it could be screwed in place on the inner chassis. Maybe not worth the effort, but I milled a little material out of the front face of the scoop so it shows some edge rather than a solid, flat face. The bracing rods are not supplied in the kit but are visible. They are easy enough to make from brass wire, a couple of small turned cylinders and some 16BA nuts. The rods terminate at one of the cross members but are not soldered to it, so that the scoop sub-assembly can be removed independently. You can see the scoop castings are soldered together solid in the 'gear up' position. Quite big lumps of brass, and needed the 100W iron to solder efficiently.


Here's how it all goes together. The scoop gear is fixed permanently to the inner chassis, as above. That inner chassis can be dropped into the outer chassis from above, the cross members having been fixed in place beforehand (did they really need to be made removable in the first place I wonder?). The long, vaguely 'Y' shaped, linkage for the deflector that is pivoted on the front cross member is just clipped in place by some short pins soldered to the cross member lugs. Just don't forget to unclip them before separating the inner and outer chassis! With inner and outer chassis fixed together, the scoop sub-assembly can be houdinied into place around the rear cross member and fixed to the inner chassis with the 10BA screw. Pictures...



The high water mark of the tender assembly looks like this at the moment...


There's a lot going on in a small space. It's arguably more complex than the loco chassis. I should mention here that there's not much clearance between the bosses on the Slaters wheels and the inside of the outer chassis frames. The distance between inside of frames on the prototype measures 5' -9 1/2 ", or 40.5mm in 7mm scale. On the model it measures between 39.8mm and 40.0mm, so a little undersize, probably down to etch material thickness being slightly over scale. And of course S7 pushes the wheels out where they should be, so the hair shirt of authenticity itches somewhat. The wheels need the bosses facing off a little to provide enough clearance for the wheels to move as the compensation works. I simply mounted them on an axle, held them in a collet on the lathe, and skimmed off 0.51mm from each boss, leaving a wheel width from back of tyre to front of boss of 3.71mm. That leaves the wheel sets at 38.6mm over bosses, which is OK. Not wishing to make life complicated, I reduced the Slaters countersunk screw at the same time. I could have taken more off the boss, but at that point the hex socket in the screw head was getting rather shallow. Pro Tip (eh?) - take a close look at the ends of your wheel screw allen keys. They're often a little undefined and rounded at the ends. Stone or grind them back a little so that the end is flat and perpendicular to the axis of the key, and all the corners are clean and sharp. Make sure the key is fully home in the socket before leaning on the allen key.

In the end this is what's visible...


It's clear that if you didn't fancy the full scoop gear monty you could leave out a lot of the fiddly internal gubbins. Brake gear next...

 

[Ian_C]

I haven't posted anything for a while. I haven't gone away, I'm just struggling with the whole tender chassis + water scoop gear + brake gear assembly. I've just got to the point where I've solved most of the problems (the ones I can see, anyway) and I've discovered that the brake hanger castings are about 1.5mm shorter than they should be. Since the upper pivots are at the right height relative to the axle centres, all the brake linkage and pull rods are 1.5mm higher than they should be. And that explains why some of the parts simply won't assemble. Clearances are small on the prototype, and a missing 1.5mm just can't be worked around. There's nothing for it but to scratch up some proper brake hangers and make up the brake gear again.

That apart, I hope everybody's well. The world's a very different place from when I last posted at the beginning of March. Work's been a nightmare, and now we're 'furloughed' until 1st May. I ought to be able to find time for those brake hangers then...

 

[Ian_C]

Those tender brake hangers. Starting with some parts and tools to make the job easier.

a - Brass bosses that will form the boss cast (or forged?) into the lower end of the hanger.
b - A small 'button' to help with marking out and filing the profile of the upper end of the hanger. It's made from silver steel axle rod and hardened.
c - A radius template for marking out the hanger profile. In this case R39.4 (I know...the label's wrong) for the concave side the profile.
d - Another template at R=28.1 for the convex side of the profile.
e - A small piece of 0.97mm NS plate (nominally 1.0mm maybe, but who's arguing about 0.03?), pre-drilled with holes for upper pivot, lower boss, and brake block pin. Six plus a spare (plus a couple of misplaced centre drill dimples!)



Here we go with some actual making...


1 - The boss holes in the plate are drilled slightly undersize and carefully opened up until the brass bosses are a press fit. The 'button' is fixed to the upper pivot hole with a 14BA screw and nut. With the boss clamped gently in the vice, the radius templates are placed tangent to the boss and the button and the profile is scribed on the plate, taking care to angle the scriber to get the line right up to the edge of the profile. You could do this by marking out and scribing all the curves by compass, but you'd need to start with a bigger piece of plate, and this seems a lot easier when there are several to do.

2 - Profile scribed. The boss is pressed out, and boss and button moved on to the next set of holes. Repeat until done. Worth noting that you need to make a decent depth of scribe mark if you're silver soldering. The soldering and cleaning up process can erode the lines a bit.

3 - The bosses are all pressed into the plate so that they project equally either side. They're made longer than required and will be filed back to the correct length after soldering, so accuracy isn't important. The bosses are silver soldered in place with enough solder to form a fillet around the boss. The idea is that it approximates to a cast or forged part. I formed up small rings of silver solder wire for this to get an even fillet all around. The three left hand hangers have the fillet on one side, and the three right hand hanger have the fillet on the other side. Since they're a press fit, no solder creeps through to the other side. There's no problem having a fillet both sides, but since you can't see one side it didn't seem worth the effort.

4 - The bosses are all filed down the the correct projection of about 0.5mm either side before the hangers are cut from the plate. The hanger profiles are cut roughly from the plate by piercing saw ad nauseam...

5 - The silver steel button is fitted to the upper pivot hole as a filing guide, and the rough hanger is held in the toolmaker's clamp/ instrument maker's vice contraption for filing & finishing.

6 - Time passes...and eventually they're all cleaned up, and approximately the same. They're seen here with some of the 3D printed brake blocks (previous post - from some time back in the Neolithic - it's there if you fancy working back through the thread). An original cast hanger is shown for comparison, and you can see that it's about 1.5mm shorter than the 4,000 gallon Stanier tender GA drawings say it should be. By way of bonus, since the brake block pins are a press fit through the hole in the hanger, the blocks are free to pivot to align with the wheel tread. I'll leave them to do that , just like the prototype, knowing that paint will fix them in position eventually (that's assuming I ever get to the stage of painting it!).

I'm getting ahead of things here. I've not tried to assemble the brake gear with these longer hangers yet - but based on what happened last time, it SHOULD work.

 

[Ian_C]

Eventually it all goes together. The process of putting it together, finding why it wouldn't assemble, taking it apart, making small modifications, changing the assembly sequence and so on, was very patience sapping yesterday. It's as close as I've come to dumping it all back in the box and finding another project to do. Slept on it. Clear head this morning, and it's all sorted.


Here are the sub-assemblies at the start of the epic. There were some mods before I'd finished.


The eventual sequence was -
1 - Compensation beams and front axle carrier fitted to inner chassis.
2 - Wheels & axles fitted.
3 - Inner chassis fitted to outer chassis. Easier to insert inner chassis from above.
4 - Water scoop assembly inserted, but not secured.
5 - Rear cross member wriggled into position around the mounting brackets and through the scoop assembly.
6 - Scoop and cross member shuffled into final positions and screwed in place (harder than it sounds!)
7 - Front crossmember positioned and secured.
8 - Scoop linkage persuaded to clip over the pins on the front cross member.
9 - Brake hangers and linkage inserted from below, and with tweezers the tops of the hangers are sprung into place over the ends of the pivot wires. There's just enough space between inner and outer chassis to do this with the slim Mr Hobby tweezers.

Here's how it ends up.




It all goes together and I have full compensation movement, and the wheel treads don't foul anything. I chose to kick the question of electrical pick ups down the road for a long time, deciding to fit them to the tender before I properly understood the implications. So fitting some pick ups is the next job.

This has been the most challenging part of the project so far. Who'd have though the tender running gear would be that difficult? Hoping the tender body is relatively straightforward.

 

[Ian_C]

Having thought about how to fit wiper pickups, the next obvious consideration was where the wires would need to end up. The DCC decoder of course, that I'd confidently said would fit in the tender after I discovered that it wasn't going to fit in the locomotive! Now I have the tender chassis together I can see that it's not going to be so easy. The 'hump' in the compensation beams mean the decoder cannot be mounted at tender floor level. If I was to make a bracket to raise it above the compensation beams it'll come very close the the rear tender deck and the sloping coal space. The only way to assess that properly is to build the tender body up to a point where I can try and fit it all together. So...on with the tender body.

First challenge on the body is to make the bend at the top of the sides. That's not easy, it's quite a tight bend (scales out to about 5.5mm radius) and the top most section of the body is flat above the curve. Also the curve has to be accurately positioned to match the profile of the front and rear bulkheads. The NS etching is relatively thick and, I guess, is in a half hard state. Given the unpredictable amount of spring back I decided not to bother trying to bend it around a rod close to the curve radius. From photographs it appears that the prototype curve (on some tenders at least, and probably all) was made by multiple small bends in a press brake. I did it a similar way, by scribing the bend start and finish lines on the inside surface, along with some equally spaced parallel lines between them. The body was clamped in the big vice between two old steel parallels and gradually formed with a series of small bends. The force required make the bends at the edge of the parallels was considerable. Lots of bending. comparing with bulkhead profile, tweaking. Of course the material work hardens as you bend it so it doesn't get any easier! Eventually it came good. Some half etch lines on the outside would probably have helped. There's going to be an overlay on top of it, so they wouldn't show.


The front edge of the tender body is also curved in. This small curve goes right to the edge and has to match a half etched guide in the tender floor. After the battle with the tender top I could see the same approach wasn't going to be work well here. I chose to anneal the ends and form them round a radius edge former. Annealing the first 10mm or so with the small torch makes the sheet very pliable, and it can be formed very closely round a tool with little spring back. That way you can get the curve right to the edge of the material. In passing - to anneal NS, heat to just dull red and let it air cool. A former was made simply by filing a 2.5mm radius on the edge of a piece of scrap steel plate.


Same approach again. Clamp the side between the former and a steel parallel, and then it's very easy to just push the annealed edge round the former with a piece wood (and a much smaller hammer). In retrospect I'd probably have done better to anneal the top curves too.

Once out of that turbulence we're into clear air, and the making up of the front bulkhead sub assembly was straightforward. It al goes together with uncanny accuracy using the MOK tab & slot approach. There's some very clever fine detail on the coal space doors. There's more detail to add, but some photos to show progress...



More or less a full day's uninterrupted modelling, apart from entertaining Titus the dog. Furlough dividend!

 

[Ian_C]

I might just get away with fitting the decoder above the compensating beams...or I might not. Need to find a way to reduce the comp beam height without reducing the stiffness too much. Finally there's use for a chunk of 3mm NS plate I acquired yonks ago.

A strip 4mm wide was sawn off the plate and blacksmithed to about the right shape. Big vice and hammer stuff, no finesse here! A lot of filing follows, to make it less offensive - there's a lot of NS dust on the floor now.


Soldered onto the comp beams with the 100W iron...


The beam above the stiffener is removed, and the thing cleaned up...


That's saved about 4mm in height. I'm pretty sure I can make it all fit in now. Famous last words?

 

[Ian_C]

With a bit more room over the compensating beams it started to look promising. Oh, yes, there's the stay alive capacitor to fit in there as well. Not much room to manoeuvre though, so the problem was schemed on CAD...

The blue translucent sheets represent the significant constraints of the tender body: rear body, rear deck (where the water filler sits), and the slope of the coal space floor. It was quickly apparent that the capacitor wouldn't fit alongside the decoder and fit through the aperture in the tender body floor. It was moved in front and dropped down to clear the coal space. There are few places where a fastener can be placed to fix the decoder mount to the running chassis. There's one right at the back near the body fixing screw, and the only other place clear of clutter where you can get access from below is well forward. There needs to be enough space in front and behind the decoder to get the wires out, so it can't be pushed up to the the rear of the body. The thickness of the mount has to clear the compensating beams below, and the combined height of decoder and mount must sit below the coal space slope. Given those constraints, it pretty much locates itself. Since the decoder mount was turning into quite a thing, I thought I may as well incorporate the pick up mountings as well, they're the little yellow cubes hanging off the sides. Making it out of non-conducting material allows the pickup assembly to be fixed directly to it without faffing with insulating bits & bobs. There's a big sheet/plate of 1/2" thick Tufnol in the workshop that was looted from a skip years ago - ideal! Fairly straightforward milling machine job, apart from the seat for the capacitor, which was scooped out with a boring head and the mounting block held edgeways up. In passing, where the Tufnol thickness was reduced to 2.5mm there was a little bit of curl. Nothing tragic though. Clearly ancient Tufnol still has some residual stress left from manufacturing.

I needed to remove the cross piece in the floor to make this arrangement work. It doesn't do anything, so no great sacrifice.


Had a couple of goes at the pickups - in a hurry to make progress, and not thinking everything through! The first attempt was with phosphor bronze strips bearing on the top of the wheel flanges. I reasoned that the strip, being vertically flexible and laterally stiff, would stay where it was wanted. Sure it does, but the unanticipated side effect was that the strips acted as little speaker diaphragms radiating all the mechanical contact noise as the tender rolled. Kind of sounded like leaking steam, but I've spent money on DCC sound, so didn't fancy drowning it all out with a hissing tender. After experimenting with different thicknesses of PB wire, I ended up at 0.5mm. Sufficiently robust to survive handling and assembly, and easy to set a balance between contact pressure and drag on the wheel.

Here are the PB strip pickups on their mounting blocks before being unsoldered and sent to the odd ends box...


The naming of parts...

The big brown thing is the decoder and pickup mounting block, machined from Tufnol. There are two holes along the centreline tapped M2 x 0.4 for the fixing screws to the chassis. An M2 thread tapped directly into Tufnol works OK, but its about as small as I'd go. With anything smaller the thread gets a bit 'statistical' in this composite material. Four through holes in a rectangular pattern for the decoder mounting, matching the holes in the decoder PCB. Four holes drilled through and counterbored for 12BA brass threaded inserts for the pickup mountings. Four M2 socket head screws (because that's what's to hand) with the heads reduced in height to fit into the decoder. The white doodads are spacers for decoder mounting. They're made from an offcut of Acetal rod, which is lovely stuff to machine with a dead sharp tool. Bottom left are the pickup mounting blocks machined from brass. The small lip on one edge orients the pickups against the edge of the mounting block. Once adjusted, the pickups always go back in the same position, no constant re-tweaking of pickup wires. One block already has a length of 0.5mm PB wire soldered to it. Incidentally, my PB wire has always been supplied as a coil, trying to straighten the stuff is a nightmare. The natural curve on the 0.5mm wire seemed about right, so that's another reason for using it! The little tag things centre bottom are the threaded inserts for the pickup mounts. Seen here, they have some wire solder tags silver soldered to them, which only turned out to be a half good idea.

Here's how the wire pickups work out...


You'll notice that I changed the arrangement of wire tags for the rear pickup mount. I'd forgotten that there was no space beneath the decoder for a top mounted wire tag.

Another 'by the way'. I find tapping small threads in small parts in the lathe to be more trouble than it's worth. But once you've parted them off it's not easy to hold them properly to tap them...unless you remove the collet holder from the lathe and put it in a vice. Tap 12BA through, holding the tap in a pin vice. Easy...


All put together, and hallelujah the Zimo Godzilla decoder fits in! The capacitor will be secured with a tiny blob of contact adhesive when the time comes for final assembly.



So that's how to get yourself backed into an expensive corner, and design your way back out of it. Just like being at work! With that sorted, I can get on with the rest of the body. DCC newbie and all, so still wiring and other adventures to come. All newfangled to us here...

 

[Ian_C]

This morning's work, the water filler and lid for the tender. There's a brass casting for this in the kit, it's not bad but it lacks a bit of crispness, and the tender top is very much on show. As usual, scaled from one of the tender GA drawings in Wild Swan, and broken down into makeable parts.


I'd thought about making the filler with the lid left open, not unusual in BR days it seems. Couldn't figure out how to make a lid with fine enough edges, so stepped back from that challenge and turned the neck and lid as one piece, being careful to leave a tiny lip at the edge of the lid, per prototype. The hinge and handle parts are all teeny weeny hand work. The main hinge part was silver soldered to the neck first. The rest was positioned with solder paint and the big iron was left in contact with the body until they just sweated in place. The handle is 0.4mm brass wire and the nuts are tiny slivers of microbore brass tube. The way to cut off tiny sections of tube like this is to open out the tube bore until the wire just fits in, then thread a short piece of tube over a short length of wire. The tube is then rolled back and forth on a cutting mat under the edge of a scalpel blade until it cuts through to the wire. Similar in principle to a plumbing pipe cutter. The little slivers of pipe don't ping all over , and since they were cut on the wire they don't close up and can easily be slid off. Getting the length right is a bit statistical, so you have to make a few and choose the ones you like best.

 

[Ian_C]

Here's where I've got to. Something emerging from the chaos on the workbench. Looking a bit Barry scrap line in this state.


The tender body presented a few challenges. For the most part the sheet metal goes together with commendable accuracy. The front section of the fire iron tunnel is a bit of a curse, having to bend on two lines to match the tunnel profile and the front bulkhead. Putting the main assembly of end & sides, front & rear bulkheads, rear deck and the main coal space profiles took some doing, not because they don't fit nicely together, but this is one instance where all the slots and tabs fight each other, some being at mutually antagonistic angles! Eventually the last tab is manipulated into its slot with a click and the whole structure relaxes into position. Soldering up was all done from the inside and reasonably straightforward. I didn't bother soldering all the seams because the fit is good, and most of it will be hidden under a few scale tons of late sixties slack.

The tender overlays were next - and a whole load of trouble. I've never got on well with overlays, and these ones scared me! The kit, rather generously, provides side and rear overlays for both the riveted and the welded 4,000 gallon tender. The riveted overlays look rather splendid, but 48142 had a welded tender, fortunately as it turned out. I found forming the overlays to match the curves previously made on the tender body to be really difficult. I figured it would be easier if the overlays were annealed, so I experimented on a riveted side overlay first. I'd wondered if heating the overlay to anneal it would cause it to pringle up, but mercifully it stayed flat. I formed the curves using the same approach as described for the body, but as the overlay bends more easily you have to take it carefully to avoid kinks. That's the downside of annealing. I got reasonably close to the right shape, but accepted that I'd have to use clips to hold the edges down when I soldered them in place. I started by fitting the rear overlay because that's easy to do, and it dictates the longitudinal position of the side overlays.

You'll have noticed in previous photos that the main body etches have a load of small rectangular holes in them for soldering on the overlay from the inside. With the main body assembled the holes higher up the tender sides, just below the curve, are virtually inaccessible by soldering iron, or RSU probe for that matter. My first attempt was to mark the positions of the holes on the overlay and pre-tin round them on the tender body. I used 145 degree solder for this to reduce the heat input, not that it made much difference in the end!

The overlay was clipped in position, adjusted to match the rear overlay, tack soldered at the rear, and then the 40W iron was applied to the outside of the overlay at the marked positions. In most positions the overlay expanded up where it was heated and needed some pressure with a lollipop stick to keep it in contact with the body. The result wasn't great...

You can see there are dimples all over in this cruel photo, where the marker pen's not rubbed off. Some prototype welded tenders do show ripples in the side plating, either from welding distortion during manufacture, or from use & abuse over the years. Some welded tenders however appear to have very flat sides. The ripples show up more on a clean, shiny tender than they do under a coat of matt BR grey. This 8F will be appropriately filthy, even so, there will need to be some filling & sanding on this side before painting. Think how much of a pain that would have been with a riveted overlay!

I had a rethink, and tackled the other side differently. The overlay was positioned and tacked at the rear, as before. The overlay was soldered from the outside in the uppermost hole positions below the curve, simply because there's no access from the inside. This time I chose to use the RSU instead of the iron. Seemed to get on better, but there are still small dimples discernible but they'll disappear under a rear of filler and some primer. This time most of the soldering was done from the inside. A blob of flux and a small chip of 145 solder was put into each hole. The RSU probe was applied to the main body etch alongside the hole until the solder melted and flashed into the joint. The advantage is that the overlay is held against the flat underlying surface by the RSU probe and body etch, and the probe can be held there until the solder's frozen. That worked out much better, very little distortion, and was probably closer to how Mr MOK had intended it to be done.


All around the edges of the body, where the overlay edges sit over the main body edges, I built up a decent bead of solder and then filed it back to produce a clean edge, free from cusp marks or divots. Time consuming, messy and it clogs up your files, but makes for a tidy job. The edge is slightly thicker than prototype, and the edge beading etched on the overlay isn't half round, but it's not practical to do it any other way with the kit. I've convinced myself that it'll be OK.

There are some things to note about building the welded tender from the MOK kit. There are some etched rivets that should be removed, and some riveted detail that doesn't need to be added. Considering how many welded tenders the were, it's been hard to find good detail photos of some areas to establish where they did and didn't have rivets. The GA drawing for the welded body suggests that everything was welded, although the drawing doesn't use any clear convention for showing weld detail, so it's a bit fuzzy. Photos suggest that welded tenders did still have some rivets. Some details, starting at the front...

There are rivet overlay strips provided to go around the coal space door and the locker. You don't need these on a welded tender. The lifting point doubler plates in the front bulkhead have etched dimples, but really need to be drilled through and have 0.6mm wire soldered in to represent rivets. The Front and rear bulkheads were topped off with a substantial half elliptical beading section, and photos show that was riveted, even on welded tenders. Possibly the riveter's union was pushing back against the rising tide of welding, who knows? So don't remove the rivets around the top curve of the F & R bulkheads. There are some etched parts (part 138) in the kit that I think are meant to be the bulkhead beading, but they're too wide and too thin. I cut some strips 1.4mm wide from some 0.6mm NS scrap, soldered then in position, and rounded them off to a vaguely elliptical profile with a file and wet & dry. They look the part. Also worth mentioning that the lamp guard is referred to as a step in the (David Coulshed) instructions, but it's actually a guard to stop lumps of coal falling onto a lamp stowed on the bracket below. If you thought it was a step you'd probably fit it the wrong way up!|


On the coal space side of the front bulkhead remove all of the etched rivets, but do apply the curved rivet strip to the top of the bulkhead.


The rear bulkhead keeps it's rivets and has a bead profile added, as per front bulkhead. There's an additional thin plate on the rear of the bulkhead shown on the GA drawing, and visible in photos. It's not very thick on the prototype, so I made a plate from some brass shim. Interestingly the GA drawing shows some stiffening angle sections between the tender side and the rear bulkhead, but they're not visible on any photos that I can find, and that'a good enough excuse to leave them off! On welded tenders the water scoop dome and the filler neck were simply welded to the rear deck plate, so there's no need to use the etched riveted flanges provided in the kit. The vents were still bolted to the deck, so don't go filing off the bolt heads! Where angle brackets were provided to fix the tender tank to the chassis, they were welded to the tank, so there's no need to add the etched rivet overlay parts here. In the photo it looks like I've added them, but that's just the reflection of the rivets on the tender floor!

Here's the best photo I could find of the rear deck of the 4,000 gallon welded tender on sister engine 48141 (from 'The Book of the Stanier 8Fs - Part 2 Wartime Engines 48126-48297', Irwell Press, and the photo's credited to Rail Online, so probably a Jim Carter snap). Another tiny detail - the the lifting lugs are squared off on this tender, but rounded on others. The GA drawings show rounded lugs, and the kit provides rounded lugs, so that's what 48142 has got! Oh yes, I've just realised that the tiny lug for the filler lid chain is missing from my dome, always another thing to do...


Next job is fitting the body to the floor, and a sprint to the finish on the remaining tender detail. I can see light at the end of the tunnel.

 

[Ian_C]

Had a phone call this morning calling me back to work from furlough tomorrow. So it'll be back to the usual geological timescales for progress I'm afraid. The end of the tunnel is further away than I'd thought. Made the best of it while it lasted, but you can't hold your breath forever.

Almost as postscript to the previous posts on the tender, I've just realised that there's a lot of useful information on these tenders in LMS Review No2 (Wild Swan, 2014, ISBN 978 1 908763 12 9 - LMS Tenders Part 2 - Stanier Standard Tenders - John Jennison). Worth looking at that series of articles if you have an LMS tender to build and, you're after the details.

So, before I try to find all my work stuff and remember how to iron a shirt, here's where I've got to.

The tender detail is nearly complete, but the finishing off list never seems to get much shorter. The tender front is complete apart from a peg in the fire iron tunnel entrance to stop the fire irons sliding forward. The small, semi-circular covers over the water valve linkages were not included in the kit, so were made from scratch. The handrail knobs look a bit lumpy in this photo, I'll replace them if I can find something better. Also worth noting that the hand rails on the rear of the body sides are a smaller diameter than the 'commode' handrails at the front (0.6mm diameter and 0.9mm diameter respectively.


Fitting the axle box castings is on the list, but I thought it worth making one improvement. I'd also thought about replacing the hanger rods with 0.9mm wire, but that's difficult to do without remaking the hanger brackets, and I can't see a way of doing that at the moment.

 

[Ian_C]

Springs and axleboxes. Quite a distinctive thing on the tender. The more I look at photos and drawings, the more I'm not in love with the whitemetal castings. They're not bad really, and from normal viewing distances they look OK. There's quite a lot to these items, and it must be near impossible to replicate them at this scale by traditional methods. There are a lot of features of the prototype spring and axle box assembly that don't show in the castings. I thought I'd see what 3D printing could make of it.

Took me 2 days to reconstruct them in CAD from drawings and photos, and I learned a huge amount about Stanier tenders in the process! Some simplification was made on detail that you're really not going to see at 7mm scale, but the resulting model looks the part I think. I chose to model the axle box and the spring and hanger assemblies as separate parts. Just in case one day I want to model the variations such as plain axlebox covers on early contractor built class 5 tenders, roller bearing boxes, or late type 'knock back' spring hangers.



I've remortgaged the house to get a set printed by Shapeways. Fine Detail Plastic 3D Printing Material Information - Shapeways

It costs way more than makes any sense, but I'm just curious to see what standard can be achieved. It takes a while, and I'll get them in June. Plenty left to do meantime.

 

[Ian_C]

Shapeways quoted an early June delivery, unless I paid a whole load more for a 'rush/priority' service. They wanted about $250 to reduce the delivery time to about a week. 250$? That's taking the P. I'm impatient, but not $250 impatient! As it turned out the 'early June delivery' dropped through the letterbox yesterday, with the bog standard priority and lowest cost shipment. Ironic? Cynical? Anyway...the parts.

They're printed in a translucent acrylic plastic and it's claimed that the process will resolve detail down to 0.1mm. I also opted to pay a premium for the smoothest surface finish option, just to see what was possible. This is what I got....


They looked pretty good out of the box. It's hard to show the detail in photos of translucent stuff, but even the smallest CAD details were present and well resolved. I'd chosen to keep the feature size to no smaller than about 0.2mm on CAD, but from the looks of this you could probably model detail right down to 0.1mm if you wished. There's still some visible layering with the 'smoothest' print option, and on parts like this there's no practical way of hand finishing to remove it. On larger, more accessible surfaces you could rub off the layering quite easily. The material files and sands nicely. I added some temporary diagonal bracing to support the hanger rods during manufacturing and handling. They're positioned so that they're easy to remove, and the scars don't show. The back of the box is hollowed out just to save material (the cost is partly determined by the volume of material) . There's a circular spigot on the base of the spring that pushes into the hole in the top of the axle box. Both modelled dead to size, and they're a light push fit. So the dimensional accuracy's good as well.


This is about as close as I can take a photo, and you can see the layering. The orientation of the layering depends on the orientation in which it was printed. I didn't specify that, it's done by the Shapeways folk when they prepare models for printing. This particular material and process uses a wax support matrix, so the design isn't so constrained as the 'build into thin air' process that a lot of 3D printers offer. Interestingly the springs were printed in a couple of different orientations, so the layering runs in different directions. You can tell if you look closely enough, but at model viewing differences I'm guessing it's invisible. We'll see.


It's recommended that you wash them in acetone before painting to remove any remaining trace of the support wax and uncured resin. Here's one with a coat of grey acrylic primer. Applied by brush in this instance, but I'll spray the rest along with the tender when it gets painted. You can still make out the layering, but it's not apparent at model viewing distances. Captures the look of the Stanier axle box very well, I think.


One of the cast white metal axle boxes with the same primer for comparison. They're not bad for overall size and proportion, but you can see that the crispness and fine detail gets lost at this size, and there are features of the prototype that I think would be impossible to replicate in a casting pattern by traditional means (somebody on WT is going to prove me wrong, you can guarantee it).

Was it worth it? In terms of what it brings to the model, yes, definitely. And a decent return for two days of research and CAD. In terms of money spent, not really. I'm not going to disclose what the exercise cost (I'm still working out the domestic forgiveness strategy). Technically it's viable and the results are brilliant, but I'd have to find a more cost effective way of getting them printed, and maybe experiment with materials and print resolution to get the cost down.

The cat's well and truly out of the bag after the stable door was bolted, and I can't help wondering how 3D printed number and works plates would turn out. Particularly since the Severn Mill chap is stood down at the moment. Hmmm...
------------
In other news, I've finally got around to making the DCC stuff work for the first time.

Ok, it's not in a loco, but after updating software and getting the bits to communicate, the motor responds as expected and the speaker makes noises. I've not set up the key press functions properly yet, so there was certain amount of pressing virtual buttons to see what happened. For a frustrating period there was motor but no sound, until the house was awoken by a surprisingly loud 8F whistle when I discovered the un-mute button. Next job is to work out how to wire it all up in the loco and tender. DCC duck broken, and I can see this has a huge amount of tinkering potential.
--------------

This is the dangerous distraction that I unearthed from a box when I was looking for something else.

A 4mm Impetus kit for the 15" Hunslet. I got it from a Scalefour Soc 'for sale' ad a couple of years ago. Original and untouched, it's a rare thing these days. There are Gibson wheel sets, what looks like a High Level gearbox, Mashima 1020 motor, Kean Maygib buffers, and a load of High Level CSB jigs and parts. The kit is very nicely done, but oh so retro! Typewritten and copied instructions and hand drawn assembly diagram. Life was simpler then. Looks like it'll make a nice little engine (somewhere between Percy and Thomas in the Fat Controller scheme of things). Can you get DCC and sound in something this size, I wonder? Like I said, dangerous distraction - I really ought to finish a project sometime, before starting another.

 

[Ian_C]

I wonder if this'll print OK? Exported from Adobe Illustrator with text as 'outlines' in either .dwg or .dxf format. Imported to CAD and tidied up a bit. Sprue and supports added. All deparameterised into a big dumb blob, exported as STL and checked over in Meshlab. The smallest letters are slightly below the claimed resolution of the print process. Probably doesn't matter much at this size as long as something's there. I've made the plates slightly thicker than scale for robust print and handling. I can thin them down a little on wet & dry if I feel the need.


48142 ended up with tender 9797 towards the end of its life. Still a 4,000 gallon welded type, but originally built in 1939 and put behind a Jubilee. Tender plates are guesswork on my part. So far as I can tell most tenders of this type had two plates: rectangular tender number plate above an oval water capacity plate, a few had a third builders plate. I've also assumed it had the sans serif style letters on the number plate. Some had a wacky turn of the century serif style for which Illustrator has no close match. Only the tender water and scoop plates and the shed plate had visible bolt heads, the rest having countersunk heads that don't show.

It'll be interesting if this comes out OK. It probably won't stand comparison with the best etched plates, but it'll be a more flexible alternative for sets of bespoke plates if an etcher doesn't have exactly what you want. No good for those splendid polished brass plates on Western locos, but probably OK for typical LMS plates; a bit rough out of the foundry and painted over several times, maybe highlighted in white if you're lucky, and subsequently lost under a layer of filth.

 

[Ian_C]

Here's how they turned out.


Again, it's difficult to capture the detail on a photo with this translucent plastic. It's far better than it looks in this photo. The thinnest lettering was 0.06mm across, and it's still printed OK. Not crisp and perfect, but more or less legible. The raised edges around some of the plates have thinned out a bit and look a little ragged. It seems to be related to the direction of the print head. I can compensate for that now that I know the rules.

Here are couple of plates coloured with a black marker pen and wiped off with a tissue moistened with IPA. They look a bit crude at electron microscope magnification, but they're absolutely fine at model viewing distance. I might do a bit better with careful dry brushing on the finished model.



Very happy with the result. It's a viable way of making plates, and I think I've found the limits of this process. Some day I'll have to have a go at etch artwork...another day.

 

[Ian_C]

Getting down the finishing off list slowly. Today's job was the drawbar between loco and tender. First question was 'how long?' I re-opened the CAD for loco axle side play, and added the tender. Allowing a bit of clearance on 6 ft radius curve gives a drawbar length of 38.3mm. That's part way between the dimples in the drawbar casting.


Drawbar was drilled out at the required centres on the mill. Because of the way I'v only partly followed the instructions I've ended up with a M3 screw at the loco end and a 6BA screw at the tender end. Disciples of Mr Webb's Crewe standardisation, may tear their hair out. I turned up a couple of bushes so that the screws can be tightened without pinching the drawbar. There's enough clearance in the drawbar to allow some vertical articulation, but hopefully not so much that there's chatter betwteen loco and tender at low speeds. The bushes were soldered to the screws to make ersatz shoulder screws.


And, for the first time, the tender is coupled to the loco. Feels like progress. The gap between steps is larger than scale by a little, but it can't be helped, and it's not tragic. I briefly thought about a sprung close coupled drawbar, before I came to my senses. A slight peeve is that the drawbar is secured from above on the loco, and from below on the tender. Better if both ends could be secured from beneath. You wonder about lateral forces on the tender when reversing a train through curves. Hopefully the tender will be heavy enough to prevent the flanges climbing the rail.


The next job is to work out the electrical connections, and any dummy pipes between loco and tender.

 

[Ian_C]

I had a laptop PC apart the other day, and there were some very small connectors in it. Some puzzling and googling later, and this is what I found...



Just made for the job, innit? These are JST type SH connectors and headers. They're 1.0 mm pin pitch, and the smallest I can find that are readily available. They come in 2, 3, 4, 5,.....up to 20 pins. The 2 pin connectors in the photo are 4.0mm wide and 2.9 mm high. They'll carry 1 Amp, which is probably enough. Amazon and eBay will oblige; about £5 for a bag of 10 pre-wired connectors and 10 headers. I need to take 4 wires across the gap, two for the motor and two for the speaker. It makes sense to use 2 x 2 pin connectors rather than a single 4 pin. That way the speaker and the motor can be removed separately if need be. Also there are more places to fit a 2 pin connector than a wider 4 pin.

Here's the manufacturer's data sheet for the connectors.. http://www.jst-mfg.com/product/pdf/eng/eSH.pdf?5ecacf8e5317b

There's enough space under the cab either side of the brake cylinder to take a couple of connectors. The headers are intended to be flow soldered to a PCB to anchor them. I doubt that it would be practical to try and solder the headers to the loco chassis, so I made a couple of tiny clamps. They're machined from brass and threaded 12BA.



I didn't fancy measuring and marking the hole centres on the chassis, so I made a couple of small pre-drilled plates as drilling guides. They were soldered in position with the RSU, and the holes were drilled through the chassis. The clamps are fixed with 12BA screws.



The connectors are 'theoretically' visible when looking from the side beneath the cab, but in reality they'll be lost in the murk behind the injector pipework, especially when they're painted black. This view is with the chassis inverted looking forward from the back end. One connector header is clamped in place on the right , and you can see an empty clamp to the left of the brake cylinder.


Seen here from below with both connector headers clamped in position, and the pre-wired connectors plugged in. The wires will cross the gap amongst the water, vacuum and brake hoses, and hopefully not be too obtrusive.


It should be straightforward to solder wires to the the back of the headers and run them to the motor and speaker. The next objective is to complete the wiring and have the loco run on live track under DCC.

 

[Ian_C]

A couple of days of picking away at the job when I'm in the mood gets the wiring almost finished.

I'm sure the connector board that I bought for the decoder isn't actually designed for this decoder, because the labelling on the connector board doesn't correspond with that on the decoder. So it took a while to establish which decoder output ended up where on the connector board. Luckily (?) all of the outputs I'm using all map to a screw terminal on the connector board.

There's not much room at the back of the tender for wiring to exit the connector, so I made leadouts from lengths of solid copper wire and formed them to sit tight around the back and end up pointing forward along each side. In fact, I made solid wire 'pins' for each connector location. They're easier to insert than wire, and tolerate repeated clamping better than wire ends. They're covered with some small heatshrink tubing. Wires were soldered to the copper pins and covered with a bit of heatshrink tubing. The heatshrink insulates the joint and provides some flex relief for the end of the wire.


Tender wiring for stay alive capacitor, motor, speaker and pick ups added. Tiny bits of heat shrink are used to keep wires together and tidy(ish). A couple of loops were fitted to guide the wires down in front of the tender wheels and under the drag beam. Motor supply is on the loco's left, and speaker on the right. And, yes, Simon, they're both black/red, and it's entirely possible for me to cross them over in the tender and feed motor supply to the speaker! To be honest, I hadn't considered that. It's 50/50 - how unlucky can I be? Coloured paint blobs on the connectors should make it clearer - thanks for pointing that out.


The speaker's right up front at the noisy end, as previously related. The wiring from that just runs through an empty boiler. I added a length of silicone tubing to protect the wires from a multitude of edges and where they squeeze past the motor in the firebox. I find silicone tubing useful for all sorts of things. You can get short lengths of the smaller sizes here... https://www.hilltop-products.co.uk


The motor wires just loop round in the firebox and dive under the cab. While I was working on this I discovered why the body and chassis were always a bit reluctant to line up perfectly. The motor's offset about 1.5mm off centre to shift the gearbox across the axle enough to insert a taper pin through the axle halves. The original motor restraining bracket wrapped around the outside of the motor, and in so doing it just made contact with the inside of the firebox casting when the chassis was fitted to the upperworks. I made a different bracket that engages the spigot on the motor end piece. It allows virtually no fore/aft motion of the motor, but it can slide laterally as the compensation tilts the axle. Incidentally I've since found a data sheet for this Maxon motor. It'll draw 0.6A continuously, which is it's steady stay winding temperature limit at 25C , and up to 1.2A for intermittent overloads. So the 1A rating for the SH connectors is probably OK. It's already been pointed out that I could have used a smaller decoder


This is how it ends up. It's a bit of a fiddle, and not something you'd want to be doing often. Also snookered myself, in that it's impossible to tighten the clamp for the speaker connector when the body is on. I'll alter the clamps so they can be tightened from the underside. What I've learned from all this is that any loco electrics beyond the usual pickups and motor in loco need to be thought out in advance.


Still got some dummy hose connections to add. The finishing off list ain't finished yet.