Gibson LMS Compound

User avatar
Julian Roberts
Posts: 1389
Joined: Sat Jan 09, 2010 11:33 pm

Gibson LMS Compound

Postby Julian Roberts » Sun Jan 12, 2014 2:22 pm

This may be the equivalent of a discussion of the sun's orbit around the earth. But here, from an era perhaps a very long time ago, before CSB, is a description of my LMS Compound compensated suspension. The model has progressed to the point where it definitely runs reliably, though much work remains to be done before it is visually complete. That will take me many more months...

In my previous contribution to the Forum "In Search of Smooth Running..." I described my lessons in making my first P4 loco, an 0-4-4 tank loco (ex CR Class 439), and particularly how I have found that more weight needs to be on the leading wheels (in either direction) than on the inner ones for good roadholding and reliability not to derail. So here I am assuming that anyone with the patience to read that far will have the motivation to refer back to the other post...

The model was acquired in a semi-built condition. The tender was almost complete, and the loco chassis was fairly near being a working proposition, mostly made as per the Gibson instructions and with the cylinders and slidebars installed. Just the cab of the loco was fixed to the semi-completed footplate framing.

With only the light springs in the hornblocks to positively keep the wheels all on the track, I was not satisfied with how I thought this arrangement would run, so the wheels, hornblocks, motor and gearbox were removed, but the chassis frames were not further dismantled. The rear wheels were arranged to run in a fixed bearing, and a new lower reduction 80:1 Hi Level gearbox was made up and a much bigger motor with flywheel attached. According to the HiLevel gearbox ratio planner this would give a top speed of 35mph, slightly higher than my normal aim of 30mph. The front hornblocks were replaced (although that was not so necessary) with MJT ones that I had already prepared for a currently abandoned project.

The boiler was now made up and fixed to the frames etc. The chassis with new motor arrangement was checked to fit into the boiler space and attention now turned to making a running locomotive.

The objective was to have 4-4-0 locomotive that ran as successfully as my 0-4-4T Class 439, but incorporating the opportunity a tender presented to get more weight where it mattered. As the tender could load plenty of weight onto the rear driving wheel, the bogie could load more weight onto the inner driving axle than on the tank loco, by making them compensated with each other. The tender would run as in Mike Sharman’s design with only the rear wheels taking weight, the other wheels lightly sprung to keep them on the track, the front of the vehicle resting on the loco but allowing swivel between one and the other.

The loco bogie would steer in the same way as the 439, and so the inner driving wheel would again not need much sideplay, just half a mm or so.

In addition I would aim to have the bogie wheels picking up current, and also the tender wheels: not by using wiper pickups but by making each axle live to one wheel. I realised that the easiest option would be to have the whole bogie live to one rail, and the tender chassis live to the other.

The bogie would be made as per Mike Sharman’s diagram for a 4-4-0 bogie, but with improvements – it would have a roller where the front of the loco rests on it rather than a rubbing plate, and the guidance fixing to the loco would not be a spring but an arm to give positive steering and directional control. Somehow this would have to be electrically isolating and yet also provide freedom for the bogie wheels to move up and down independently of the loco.

Additionally the bogie would have the weight from the loco arranged to slightly bias this weight to the front wheels.

The weight of the loco from a boiler filled with lead would rest on the bogie, which would be compensated to the front (inner) driving wheel. The aim would be for twice as much weight to be on the bogie as the driver. As there was not so much scope for weight at the rear of the loco the outer rear driving wheel would be lighter on the track than the inner front one, and this would not work on its own in keeping the loco on the track (i.e. it would derail) but with sufficient weight from the tender the rear pair of wheels would have more weight than the inner front pair.

The steering arm would be pivoted a little in front of the inner driving wheels.

The options were considered with the drawing and photos to see where the least visible places could be for the bogie steering arm pivot and the compensation fulcrum. I made two fixing points for the steering arm in case the outer one needed more sideplay than I installed in the front driver. The motion plate that ran through the model chassis between the ends of the slide bars was about half way between the front driving wheel and the centre of the bogie, so that looked like a good area for the compensation fulcrum. This was rather nearer the driving wheels than I had intended though still less than half way from the bogie centre, so more weight is on the bogie than front driver. The actual distances are 21+a half mm front bogie asymmetrical resting point to fulcrum, and 26+a half mm fulcrum to front driver, centres. The slide bars of the centre connecting rod plus crosshead and piston looked from photos as though they were in much the same place as the compensation beam would or could be.
Midland Compound 006.jpg


So I made up a compensation beam to at least begin to resemble these elements (much improvement possible here!), and how I imagine the inside connecting rod to look – although this is hidden on this and any other photos I had, and so largely will be on the model as well, by the weighshafts and other gear. At the front a plate was arranged on which the bogie would actually rest. After this photo was taken a layer of Copperclad pcb for insulation was added. At the back I arranged a screw to give adjustment of the ride height of the loco.

2013-11-11 15.59.42.jpg




The bogie as constructed by the original maker as per the Gibson instructions was dismantled. The side frames were used to construct an H shape arrangement of tubes.

2013-11-05 19.47.29.jpg


Both axles can tilt. A roller was arranged asymetrically to bias the weight of the loco to the front wheels.

2013-11-05 19.49.24.jpg


The bearings on the side frames were filed to just the thickness of the frame material so that they could allow the axles freedom of movement to tilt, as they now had only cosmetic purpose beyond restricting the tilt to a necessary amount of one or two mm.

Two bogie wheels were painted with electroconductive paint, but I found this did not conduct any current. Anticipating future unreliability, even if I could make the paint work, I abandoned this idea. Shorting wires from the EMGS (which are not a very satisfactory etch for the 2mm ones) were fixed to them instead. For this operation I removed the plastic wheel centre by immersing in hot water, and tinned the rim in two places opposite each other as lightly as possible. The shorting etch was cut till it only just reached the rim at each end, and soldered on after the rims had been reattached with Loctite - soldering as quickly/deftly as possible to avoid heating up the plastic. Care was taken to then remove any stray solder so that the flange thickness was undisturbed for running through pointwork. The following photo shows a modification to one axle which will be explained later.

2013-11-28 11.45.30.jpg
You do not have the required permissions to view the files attached to this post.
Last edited by Julian Roberts on Tue Dec 10, 2019 5:49 pm, edited 2 times in total.

User avatar
Julian Roberts
Posts: 1389
Joined: Sat Jan 09, 2010 11:33 pm

Re: Gone Retro: compensation for a Gibson LMS Compound

Postby Julian Roberts » Sun Jan 12, 2014 4:19 pm

The compensation beam was installed with the fulcrum pretty well invisible below the centre slidebar representation, fixed to the chassis just ahead of the motion plate. This was rather nearer the driving wheels than I had intended though still less than half way from the bogie centre. The fulcrum could always be removed and another one installed further forward if running was unsatisfactory, but it would have to be above the ‘slidebars’ because of the shape of the frame – it would be in the area of the rear bogie wheel frame cutout – so less satisfactory from a visual point of view.


2014-01-12 14.53.30.jpg
[atta Thchment=0]2014-01-05 22.54.47.jpg[/attachment

Much thought was needed before I came up with a means of connecting this bogie to the loco. I wanted something that:

1 gave as little slop as possible, so as to steer the loco precisely
2 allowed the bogie to adopt a different horizontal to the loco
3 electrically isolating
4 was invisible from side on viewpoint

Eventually I realised that a kind of hornblock arrangement was needed to attach the steering arm to the bogie in such a way that the rear axle could move up and down within the steering arm.

The steering arm Mk 1 started like this:

2013-11-16 22.39.10.jpg


The bogie rear axle was sawn off each side of the centre. In its place 2mm-1/8” reducing bearings soldered within square tube internal 1/8” became the new rear axle of the bogie. The next size up of sliding fit square tubing (with one side removed) was arranged to be the attachment from the steering arm.

2013-11-25 11.43.55.jpg


Additionally, the bogie rear axle needed to be able to twist relative to the steering arm yet be centred by it, so an additional sliding arrangement (again made from square tube with one side removed) was attached to the steering arm square tubing to fit over the end of the main bogie H centre piece.

2013-11-28 10.36.37.jpg


This square tubing fixing point of the steering arm was then glued (with Araldite and cigarette paper insulation) to the main part of the steering arm that fixes to the loco. The locomotive was placed on the track with the completed bogie in situ making sure all was sitting straight for this operation, and left for 24 hours to allow the glue to cure. It was necessary at the same time to ensure that the sliding hornblock arrangement was not at the top of its travel, so some paper was put between the axle square tubing top and the attachment piece for this operation. The completed steering arm looks like this:

2014-01-05 22.54.47.jpg



and on the loco the whole arrangement like this:

2014-01-12 14.51.14.jpg


and this...the pivot screw is temporarily long to ease the constant removal and refitting needed at this stage...any evidence will be hidden behind the footstep and assorted sanding etc gubbins

At present the bogie is free to drop out of the steering arm upon removal from the track.
2014-01-05 22.39.59.jpg
You do not have the required permissions to view the files attached to this post.

User avatar
Julian Roberts
Posts: 1389
Joined: Sat Jan 09, 2010 11:33 pm

Re: Gone Retro: compensation for a Gibson LMS Compound

Postby Julian Roberts » Sun Jan 12, 2014 5:08 pm

I had realised that compensating the bogie was not entirely straightforward (!!), in that upward movement of the bogie was restricted by the very small gap between bogie sideframe and loco frames, and there was no easy way round that which would allow the bogie as much swivel as I intended unless the gap was increased by a very unsightly amount. So I concluded that the bogie would just have a very small upward amount of travel (thus the front driver downward), but that for further driving wheel downward travel it would rely on the springing from the pick ups to keep it on the track. (Thus the original Gibson hornblocks might have been better) Downward movement of the bogie (and upward movement of the driving wheel) would be proper compensation as originally intended. So extra bits were soldered onto the end of the compensation beam, this amount was the now redundant amount travel of the beam. I was constantly pleased to have soldered on that adjusting screw at the other end.

The underneath looks like this without the bogie etc (the wire for the current from the bogie goes to some copperclad yet to be connected to the motor):

2014-01-05 22.50.52.jpg


The fulcrum of the compensation beam is only fixed to the frames with solder, maybe I'll have to do something to make that a stronger join.

Cut outs, as undiscernible as possible from normal viewing angles, were filed into the back and front of the inner halves of the cylinders to allow the much greater than prototypical bogie swing.

On the day I eventually put the working chassis to the test on my piece of test track with a 4’ reverse curve the loco behaved exactly as I expected, except that with a little to and fro running (on my other piece of test track with auto shuttle reversing), and proper lubrication of the front driving wheel hornblocks, the loco was in fact capable of holding the track without the tender.

2013-12-08 09.54.25.jpg


To make problem solving simple I had only put pickups onto the rear fixed driving wheels. However I had a problem with some shorting on the curves, so the cutouts in the cylinders were lined with a layer of Araldite soaked cigarette paper. (There was another elementary problem that made the loco live to one side that I had to sort as well). Sorting both cured the shorting problem.

The chassis of the tender will retrospectively have to be made insulating, but I foresee no big problem with this, mostly using the Araldite and cigarette paper fix. A coupling bar was attached to the tender,

2014-01-12 14.55.25.jpg

and the EMGS loco tender coupling assembly attached to the loco.

2014-01-05 23.26.44.jpg


This allows the tender to swivel relative to the loco, though there isn’t an equivalent arrangement to allow for any different horizontals, which just have to be accommodated by play in the whole lash up, of which there is sufficient. The tender wheels are fixed with the Alan Gibson sprung hornblocks, so all that was necessary was to adjust the screws that that fix the ride height of the vehicle, so that just the rear ones hold a correct horizontal, and the others are unscrewed as far as they can go while still holding the springs in place.

As an aside, I filled the boiler with liquid lead (having fixed on the smokebox door) and plugged it with a molten mix of 70deg solder and whitemetal offcuts. It is quite possible to then flush out all the flux (I use phosphoric acid 12% from London Road Models), as the boiler is unprototypically far from water tight, especially with holes drilled for the smokebox door handrail knobs. These holes are not big enough to allow the liquid lead an exit point. So the weight is firmly fixed in but without issues for the eventual paint job, as I have already proven to my satisfaction on my CR Class 812 loco where I did the same thing.

2014-01-05 23.11.23.jpg


2014-01-05 22.42.38.jpg
You do not have the required permissions to view the files attached to this post.
Last edited by Julian Roberts on Mon Jan 13, 2014 9:52 pm, edited 1 time in total.

DougN
Posts: 1252
Joined: Wed Sep 01, 2010 9:57 am

Re: Gone Retro: compensation for a Gibson LMS Compound

Postby DougN » Sun Jan 12, 2014 9:01 pm

Looking really good there Julian. I like the way you have set up the steering to the bogie/ locomotive. I will read through more once I continue battle with the LNER D17 I am building you have already thought through some of the issues I have come up with. I have a further challenge in that the foot plate effectively cuts off the access to get a compensating beam through to the front axle. Any how once I have some time (now being back at work! :evil: ) and enthusiasm I will get into finishing it.
Doug
Still not doing enough modelling

User avatar
Julian Roberts
Posts: 1389
Joined: Sat Jan 09, 2010 11:33 pm

Re: Gone Retro: compensation for a Gibson LMS Compound

Postby Julian Roberts » Mon Jan 13, 2014 9:30 am

Hi Doug - thanks, pleased that this is of some interest to someone out there, in fact the other side of the world I see! - despite being so "old hat" being in the compensation era still.

Just had a quick skim through your D17 thread, what a lovely looking model... just to say the compensation beam surely can be any shape whatever to follow the necessities of what may be in its way - to a degree at least.....? Also it could twist round any central fixing screw if that is a problem.

Or looking at it the other way, if it's spacers that are in the way, well, the motion bracket was in the way on this model, and I forgot to mention that I filed a big slot in it for the compensation beam (aka inside connecting rod etc)

DougN
Posts: 1252
Joined: Wed Sep 01, 2010 9:57 am

Re: Gone Retro: compensation for a Gibson LMS Compound

Postby DougN » Mon Jan 13, 2014 11:02 am

You are right I have yet to put the hole in the chassis spacer for the beam to go through. Though I have a few issues to solve with the bogie first. Back to work has the problem of taking al the spare time to do modelling though I do a lot more when it is hot... Though we are in for a hot spell, 31 today, 43, 39, 41, 40 before we get some relief... Watch the Aussie open for some ridiculous temperatures! :shock: .... Need an emoticon for sweltering.... I will not be doing any modelling til the week end with all this!
Doug
Still not doing enough modelling

User avatar
Julian Roberts
Posts: 1389
Joined: Sat Jan 09, 2010 11:33 pm

Gibson LMS Compound

Postby Julian Roberts » Tue Dec 10, 2019 2:09 pm

Six years have passed on what were supposed to be some quick projects. Now my focus is getting this finished. I had been rather set back with electrical problems, short circuits. I assumed I had not got full insulation of bogie or tender and spent much time trying to see what I'd done wrong. I overlooked the obvious - the motor tags were intermittently touching the boiler or part of the chassis. So the loco was sometimes live to one rail or another and it only took the tiniest touch of the bogie to loco to get a short. The bogie can easily just touch in several places but now that the motor tags are fully away from touching anything there is no problem.

One irritating constant issue was the front wheels of the tender. These tended to derail. Only the rear wheels serve any function, the front four just trundle along. The tender front rests on the loco coupling, adding to adhesion. The problem here is the front wheels need a lot of sideplay and more spring force than comes with the Gibson coil springs. (The centre pair seem to work OK with the coil springs, presumably because less sideplay is required). I find springs frustratingly difficult. I eventually came up with this design. It is long so the spring force doesn't change much over the up and down travel distance of the wheels in the hornblocks, which is what is needed - a constant force. It mustn't be too strong or the whole reason for the design is negated.

This all worked quite satisfactorily on both layouts on admittedly limited testing and I thought I was on the home straight. Then disaster - I found the driving wheels were loose on their axles. At least that explained why sometimes the slow running smoothness seemed slightly better or worse.

Nothing for it but to start all over again as the wheels had clearly had too much adjustment done to them in the past and lost their grip. They were hopelessly loose and pinning or various other bodges I couldn't contemplate on a new model.

I could have sent them back to Colin for new centres but on looking into it I found the crankpin throw on them was wrong so I ordered new ones.

I'll post correct picture later!!
You do not have the required permissions to view the files attached to this post.

User avatar
Julian Roberts
Posts: 1389
Joined: Sat Jan 09, 2010 11:33 pm

Re: Gibson LMS Compound

Postby Julian Roberts » Fri Dec 20, 2019 10:27 am

Hereon I'm working on both the chassis and body. I always find it a struggle deciding which actual loco I'm modelling. The first criteria is, one that was working in Scotland in the 1950s. I've fixed on extended frames and smoke box so that limits me. It's difficult finding photos of both sides of one loco at a roughly similar date. Jeremy Suter kindly sent me 4 photos of 40938 righthand side and I found Allan Goodwillie had in his tirelessly supportive way :thumb :) sent me in 2015 among about 20 photos one that showed the left hand side of the same loco. Thanks Allan and Jeremy!! So that's the one it's gong to be. Minor drawback - it was shredded in Perth, not quite appropriate for Calderside, but never mind. So here they are.
img20191208_20133455.jpg

img068.jpg

The washout plugs change from photo to photo of the four Jeremy sent. Current task is to make some representation for them. Very time consuming making 16 of these, 12 small and 4 large, and I am not absolutely sure how best to do the large ones. I've settled on a way of doing the small ones. Here the first three are complete. (Actually I'm not quite sure what is a washout plug and what is a mud hole door :o - anyone want to give this basic lesson?!))*
20191220_101031.jpg

Having picked up the kit semi built there are some problems I won't be able to sort out so I'm not aiming for absolute perfection but just something that satisfies me as being as good as "reasonably practical". So I'm not getting a beautiful turned chimney and sticking with the Gibson short Stanier lost brass casting. The backhead is too high but I'm afraid I'm just going to file a bit of it away where that is evident from these exterior viewpoints.
20191219_224238.jpg


*Edit....I now see which is which.
You do not have the required permissions to view the files attached to this post.
Last edited by Julian Roberts on Fri Dec 20, 2019 11:12 am, edited 1 time in total.

David Thorpe

Re: Gibson LMS Compound

Postby David Thorpe » Fri Dec 20, 2019 11:11 am

I now do my tenders with a combination of CSBs and split axles. All heve worked well with this combination.

DT

User avatar
Horsetan
Posts: 1371
Joined: Tue Jul 29, 2008 9:24 am

Re: Gibson LMS Compound

Postby Horsetan » Fri Dec 20, 2019 12:47 pm

Julian Roberts wrote:.....Having picked up the kit semi built there are some problems I won't be able to sort out so I'm not aiming for absolute perfection....


There seems to be an amazing amount of solder all over it, almost as if there had been a leak! :shock: I usually try to get by on the minimum necessary to hold parts together, but as my varifocal-assisted eyesight steadily worsens with age it becomes ever more difficult.

I am a little unclear on the tender that the preserved 1000 has. As I understand it, it is a full-length body on a Deeley underframe, but came from a withdrawn S&DJ 7F 2-8-0 and the shape of it is different at the front.
That would be an ecumenical matter.

User avatar
Julian Roberts
Posts: 1389
Joined: Sat Jan 09, 2010 11:33 pm

Re: Gibson LMS Compound

Postby Julian Roberts » Fri Dec 20, 2019 3:10 pm

Yes.....Solder or no solder here is an attempt at the mudhole door. I wonder if there's a better way of representing them.

20191220_150548-1.jpg
I took the type of door on this loco
img023-1.jpg

forgetting one of Jeremy's photos showed no door at all on the right side... :cry:
You do not have the required permissions to view the files attached to this post.

Phil O
Posts: 363
Joined: Sun May 05, 2013 5:23 pm

Re: Gibson LMS Compound

Postby Phil O » Sun Dec 29, 2019 3:25 pm

Horsetan wrote:
Julian Roberts wrote:.....Having picked up the kit semi built there are some problems I won't be able to sort out so I'm not aiming for absolute perfection....


There seems to be an amazing amount of solder all over it, almost as if there had been a leak! :shock: I usually try to get by on the minimum necessary to hold parts together, but as my varifocal-assisted eyesight steadily worsens with age it becomes ever more difficult.

I am a little unclear on the tender that the preserved 1000 has. As I understand it, it is a full-length body on a Deeley underframe, but came from a withdrawn S&DJ 7F 2-8-0 and the shape of it is different at the front.


Hi Horsetan

And others in the same prdicament sightwise, I was also struggling with varivocals whilst modelling and so the last time I saw the optician I mentioned it and I now have a x3 single focus pair for modelling and other close work. It was a BOGOF deal.

Cheers

Phil

User avatar
Julian Roberts
Posts: 1389
Joined: Sat Jan 09, 2010 11:33 pm

Re: Gibson LMS Compound

Postby Julian Roberts » Mon Dec 30, 2019 3:11 pm

Hi Horsetan - my pc was Out of Service but now I can upload the pictures I took of 1000, that show the tender. I don't know the answer to your question, but here they are if that helps.

The tender of this kit was more or less complete when I bought it.

Regarding all the solder, I hope my skills may have marginally developed over the last six years. Some of this loco around the splashers hangs together with separate pieces of metal that simply abut each other at right angles. However I never quite see what is so bad about a very thin surface solder layer that will be covered in paint, if there is no perceptible surface change of level after painting, and maybe erroneously understand paint will adhere better to it than brass. I see this may be bad practice but I'm not quite sure why.

Midland Compound 074.jpg
Midland Compound 073.jpg
Midland Compound 065.jpg
Midland Compound 066.jpg


Meanwhile here is my completed effort at the RHS washout plugs etc.
You do not have the required permissions to view the files attached to this post.

essdee
Posts: 554
Joined: Sat Nov 29, 2008 4:47 pm

Re: Gibson LMS Compound

Postby essdee » Mon Dec 30, 2019 5:31 pm

Julian, Horse,

Re the unusual tender..

It originated behind a SDJR 1914 series engine, and had a tender cab prior to ca 1920-22. I understand some mods were made for it to be attached to 1000, but the front side-sheet profile is the give-away. The curve does not begin immediately in front of the front coal plate, but some way forward of it. Imagine that extended upwards....and you can envisage the tender cab fitting.

I will try to dig out comparative pics, once New Year is behind us...

Cheers,

Steve

User avatar
Julian Roberts
Posts: 1389
Joined: Sat Jan 09, 2010 11:33 pm

Re: Gibson LMS Compound

Postby Julian Roberts » Mon Dec 30, 2019 8:55 pm

Found I had this too - any use?
You do not have the required permissions to view the files attached to this post.

User avatar
Julian Roberts
Posts: 1389
Joined: Sat Jan 09, 2010 11:33 pm

Re: Gibson LMS Compound

Postby Julian Roberts » Mon Dec 30, 2019 10:29 pm

Anyway, now it's the bit that possibly explains why I found I had more urgent loco projects six years ago. :o

Midland Compound 015.jpg

Midland Compound 108.jpg

Midland Compound
107.jpg

Midland Compound 106.jpg

This is the right hand side of 1000. I believe all this plumbing is the large ejector (?). It is on the same side as the reversing lever. My chosen Compound 40938 was left hand drive at my chosen time, and this plumbing is discernible on the only LH side picture I have of that loco.

img068.jpg


My question is whether anyone has any hints on how to model this exuberance of pipework, whether anyone knows if there are castings that might save me a lot of time and invention :?: :!: I have several detail pictures of 1000 taken from ground level, but only this photo from the time of another Compound showing a good amount of clarity.

img023.jpg

I have the drawings in the Wild Swan book
20191230_221543.jpg

20191230_221513.jpg


All that comes with the kit is this casting which I so far can't relate exactly with the pictures.
20191230_213849-1.jpg

I realize I'm better off than I might be, being in possession of the drawings. :idea:
Very many thanks if anyone has any hints!
You do not have the required permissions to view the files attached to this post.

User avatar
Horsetan
Posts: 1371
Joined: Tue Jul 29, 2008 9:24 am

Re: Gibson LMS Compound

Postby Horsetan » Mon Dec 30, 2019 10:56 pm

Julian: thanks for the 1000 tender pics - all real-life preservation photos are useful.

essdee wrote:Julian, Horse,

Re the unusual tender..

It originated behind a SDJR 1914 series engine, and had a tender cab prior to ca 1920-22. I understand some mods were made for it to be attached to 1000, but the front side-sheet profile is the give-away. The curve does not begin immediately in front of the front coal plate, but some way forward of it. Imagine that extended upwards....and you can envisage the tender cab fitting.

I will try to dig out comparative pics, once New Year is behind us...


Thanks. I'm trying to work out which bits of the various AGW Fowler and Deeley tenders I can use to replicate this.
That would be an ecumenical matter.

John Palmer
Posts: 825
Joined: Fri Jul 15, 2011 11:09 pm

Re: Gibson LMS Compound

Postby John Palmer » Tue Dec 31, 2019 1:13 am

I'm sorry to have to say that making a reasonable representation of these large ejectors is an exacting and fiddly task, involving the manipulation of tiny pieces of brass that end up with more holes in them than one of those Dutch cheeses. The drawing is of limited help because it fails to capture the subtleties in shape of the main ejector body casting in the way that your photographs do.

The rear part of the casting is a relatively straightforward box shape with a stepped top. It is somewhat complicated by the 'cut out' in the bottom front corner, within which sits a recessed, curved section of casting. Ahead of the box-shaped rear section this curved section is extended upwards to the 'step' in the rear section, and in front of this is the lozenge-shaped jointing flange into which is inset the offset exhaust pipe - which has then to be cranked inwards so as to lie close alongside the smokebox until it reaches the elbow via which steam and air are vented within. Just to complicate matters further, the main steam feed is led into the ejector body casting between the jointing flange and the 'box-shaped' element.

My approach has been to make the box-shaped part of the casting as one component, into which I cut the step in the uppermost surface. Probably it's best first to bore the four holes in the top surfaces (two per step), into which stubs of wire can be fitted to represent the plugs, which are quite prominent. I make the curved front section of the casting and the jointing flange as a separate lathe-turned item, turning the flange oversize before filing it to its final lozenge shape. In theory it ought to be possible to offset this turning in a four-jaw chuck in order to turn the exhaust pipe integrally with the front half of the ejector body, but you are then faced with the awkward task of putting an accurate double crank into this length of pipe, and up to now I've chickened out of this approach and used a separate turning to represent the exhaust, having first drilled an offset bore in the front of the jointing flange to receive it.

Above the main casting you have a separate casting for the steam admission valve. My best attempt at this has, again, been a turning into which I have milled the flat surfaces of the body housing the valve (the bit painted black in your photograph). This has to be bored on the bottom surface to receive a length of wire representing the pipe feeding steam into the main ejector body. The lower end of this stub of wire has to be fixed in a holed bored immediately behind the jointing flange but ahead of the 'stepped box'.

Probably the biggest problem to be overcome is the number of 'pipes' (bits of wire/rod) that have to be terminated securely in the body of the ejector. With holes having to be bored front, back, top and bottom to receive these (not to mention the wire stubs representing the aforementioned plugs!) the area available to create a satisfactory joint is minimal, and there is a high probability that when securing one length of rod or wire, you will compromise the joint of another.

It certainly assists the fabrication of one these ejectors to be able to turn and mill its constituent components, but I hadn't acquired my Unimat at the time of making my first for Bulldog 43216, so it must have proved possible to make a reasonable representation without resort to machine tools. I've attached a picture of the ejector I fitted to 44417, to illustrate how the approach described pans out.
4mm ejector.jpg
You do not have the required permissions to view the files attached to this post.

User avatar
Julian Roberts
Posts: 1389
Joined: Sat Jan 09, 2010 11:33 pm

Re: Gibson LMS Compound

Postby Julian Roberts » Tue Dec 31, 2019 11:04 am

Very many thanks for your comprehensive reply John. Your photo looks very convincing. I don't own any of these machine tools, though I can have access to a lathe at our club. I wonder how you fabricated the one done without machine tools and how it compares. Mulling this over, I have some double sided copperclad sleeper strip that might be a basis, in that it could be more readily worked on than a lump of brass, yet with a solderable surface. Once I've worked out the dimensions of this housing in 4mm scale it'll be clearer what the practicalities are, and whether it would be a useful learning project for lathe use following your methods. Interesting that you point out the limitation of the drawing. I thought it was my lack of know-how that stopped me seeing them particularly readily in three dimensions.

Edit -

Turns out the dimension is around 3.5mm, same as the sleeper strip!

John Palmer
Posts: 825
Joined: Fri Jul 15, 2011 11:09 pm

Re: Gibson LMS Compound

Postby John Palmer » Tue Dec 31, 2019 12:54 pm

OK, for comparison I have attached a shot of my first ejector, as fabricated without the assistance of the Unimat.
Ejector on 43216.jpg
You can see that I made fairly liberal use of 14 BA nuts to represent the big pipe unions, but the 'lever' at the top representing the control arm on the steam admission valve is a pretty crude bit of wire. Still looking for a better way to simulate that...

I would be disinclined to use PCB as the basis of one of these. Even with double-faced PCB you are still left with only two planes on which a solder joint can be made, and you need four. Moreover, the surface area of copper on the finished article will be minimal, with a high risk that it will de-laminate in the course of assembly. Also, if the body casting is represented by a solid piece of brass you have a better chance of securing the various attached bits of rod and wire as interference fits in what will prove to be bores having a very shallow depth - this will mean that the application of minute amounts of solder need be nothing more than a 'belt and braces' measure.

My other recommendation is to start with a piece of bar (whether circular or rectangular in cross section) that is long enough to be held securely in your work-holding device of choice until the entire assembly is almost complete - to me, the most obvious way of doing this is to have the back of the ejector casting closest to the chuck or whatever you are using. It's much easier to work on small components like this if they remain attached to a larger object on which you can maintain a firm grip.
You do not have the required permissions to view the files attached to this post.

User avatar
Julian Roberts
Posts: 1389
Joined: Sat Jan 09, 2010 11:33 pm

Re: Gibson LMS Compound

Postby Julian Roberts » Tue Dec 31, 2019 9:34 pm

Again very many thanks for your suggestions John, very wise in all aspects. And thanks for sharing your photos. I'll be pleased to get half as good an outcome.

Happy New Decade anyone reading this.

User avatar
Will L
Posts: 2516
Joined: Sun Jul 20, 2008 3:54 pm

Re: Gibson LMS Compound

Postby Will L » Wed Jan 01, 2020 1:47 pm

John Palmer wrote:...My other recommendation is to start with a piece of bar (whether circular or rectangular in cross section) that is long enough to be held securely in your work-holding device of choice until the entire assembly is almost complete - to me, the most obvious way of doing this is to have the back of the ejector casting closest to the chuck or whatever you are using. It's much easier to work on small components like this if they remain attached to a larger object on which you can maintain a firm grip.

I want to heartily endorse this bit of advice for making up small detailed components. I would like to add that when making up this sort of plumbing it is useful to have not just a vraiety of rod/wire sizes, but also stocks of fine tub in a variety of sizes which will help you fabricate a virality of plumbing details which can be safely threaded together on an underlying continuous bit of wire, (or a finer bit of tub, these things get recursive).

User avatar
Julian Roberts
Posts: 1389
Joined: Sat Jan 09, 2010 11:33 pm

Re: Gibson LMS Compound

Postby Julian Roberts » Wed Jan 01, 2020 9:18 pm

Absolutely Will. I haven't encountered the terms tub or virality but I think my activity today illustrates the point you are both making about leaving small items large enough to be able to hold firmly as long as possible.

I'm going from the biggest parts of the loco to the smallest - this exhaust steam injector came into the substantial category - as the smaller the detail the more easily damaged. Getting into practice before tackling the ejector. The injector pipe is made of 1.6 mm tube - easier to bend than rod the same diameter. Just a few of the Compounds had these but the advantage is that on these locos there's almost nothing else on this side after the handrail and lubricator.
20200101_171129-1.jpg

20200101_204858.jpg

20200101_211127.jpg

More cleaning up to do....!
You do not have the required permissions to view the files attached to this post.

User avatar
Julian Roberts
Posts: 1389
Joined: Sat Jan 09, 2010 11:33 pm

Re: Gibson LMS Compound

Postby Julian Roberts » Sun Jan 05, 2020 7:51 pm

Meanwhile the chassis has been revisited, having found I needed renew the driving wheels.

Looking up the Sharman bible I found the crank throw should be greater than the ones that came with the semi-built model. The correct wheels are those also for one of the large GWR locos. I knew this might give piston problems but there was plenty of clearance below the splashers.

Another reason to start right over again would be to check the coupling rod to wheelbase correlation, as I had not been quite satisfied with the smoothness of slow speed movement. This meant removing the rear axle from the Araldite fixed final drive gearwheel of the High Level gearbox. This was much easier than I feared. The axle is quite a soft steel so the piercing saw went through quite easily. The gear came loose from the remaining axle stub within a few moments of holding it above one of the cooker jets flame.

As I expected the correlation was not perfect. In general when making a loco I think it's much easier to adjust rods to wheelbase than adjust hornblock positioning in the frame to rods. I couldn't unsolder the rods that came with the loco. Maybe they were just one piece. I made new ones from Gibson universals to correlate with jig axles.
20191202_152543.jpg

20191202_174903.jpg

At one end, where the double lamination is short of the full length, these are really quite weak, so I made templates in case they should ever break.
20191205_140022.jpg

The new wheels were fitted with crankpins as I described here. viewtopic.php?f=20&t=6689&start=25
And the front pair put together to test for fouling.

As expected, there were problems. One piston was now too short and came right out of the cylinder at the new rear position.
20191207_151709.jpg

I had been troubled whether the strength of the whitemetal piston tail pipes coming from the cylinder front was sufficient, so I cut them off. With no wheels in the way it was possible to drill a hole properly all the way through the cylinder,
20191207_150212.jpg

and I soldered brass rod replacements into the front 2mm or so, reckoning a tube might not align fully with the piston, and not be so strong. I soldered a new longer piston onto the crosshead, adjusting length to not foul at the forward position. The chassis was then painted properly.
20191213_152622.jpg

Wheels were then completed and here it is testing for fouling with the bigger crank throw, in push along mode. Somehow to me the bigger throw makes the engine look more real powerful and business like.
20200103_165825-2.jpg
You do not have the required permissions to view the files attached to this post.

User avatar
Julian Roberts
Posts: 1389
Joined: Sat Jan 09, 2010 11:33 pm

Re: Gibson LMS Compound

Postby Julian Roberts » Wed Jan 22, 2020 9:41 pm

With the new wheels the loco went like this on first powering up. This was minimum speed.

The front tender wheels prove to be damaged and the centre ones at this stage had a coil spring tangled up in the axle. That has since been sorted, and Colin will replace the centres of the front pair at no cost even though the damage was (I'm sure) my fault.
Much time was wasted as the loco was tending to derail at some track joints on my straight test track when coupled to the tender. Eventually I realised the very simple power connection (designed to connect power automatically when coupling tender to loco - fairly shamelessly lacking in realism I'm afraid) between loco and tender had been replaced too short since my work on the chassis, so it was just catching. I'll put a photo of that later here. Here it is, long enough so plenty of overlap now.
20200123_081913.jpg

The handrails were then the task. Both sides are 0.9 tube where full diameter. Not having a lathe there was no question of making it all in the same way as John Palmer did (see earlier). I thought there should be enough rigidity in all the pipework in itself, resulting from the fixings to the boiler, and thus the casting could be fabricated as a non-structural add-on last in the process. Photos in construction order.
20200113_122613.jpg

20200113_164635.jpg

20200115_220110.jpg


Short Gibson handrail knobs could be broached out to 0.9mm but the medium ones disintegrated so I had to use the split pins that came with the kit. On the RHS, having seen these are rather heavy, I used NS wire twists.
20200122_101411.jpg

20200119_215104.jpg

The boiler band strip that came with the kit was useful for making the small lever linking the main casting to the steam admission valve. A 7mm scale handrail knob was the base for this bit.
20200120_115649.jpg

20200120_121418.jpg

The actual casting I fabricated from two pieces of NS strip.
20200120_184447.jpg

I used the "bolts" to hold the thing when attaching to the almost complete assembly.
20200120_200156.jpg

20200120_214123.jpg

Last I drilled 3 more holes and fixed the wires with Loctite.
20200121_185027.jpg

Lots more fettling up to do e.g. filing down the heavy looking handrail knobs.


During this process I bought a set of 20 Dormer drills 0.3 - 1mm, which made the whole thing far easier than previously struggling with duff drills.
You do not have the required permissions to view the files attached to this post.


Return to “Chassis and Suspensions”

Who is online

Users browsing this forum: AhrefsBot, ClaudeBot and 0 guests