Harlequin

The key thing to avoid is the building’s ridge line getting chopped by the back scene because it will look artificial. The ridge line is the horizon and any chopping that is required behind it should be OK because it will be hidden. Grandmother? Eggs?

Balance lever plate, balance lever and weight with connection to down rod: Some of these bits will need to be improved later. (And I should really fix everything together with bolts and screws...)

If you don't want to go into business and the costs don't stack up for whatever reasons, how about making 3D models of the pipes publicly available for people to print on their own 3D printer?

I realise it's biased to return to danger but it's set up so that not all the weight is on the danger side of the balance (if you see what I mean). That's what I meant by not affecting one side or the other "excessively".

Yes, absolutely! Now that we have DCC Sound and locos with operable whistles, this is another small aspect of working practice that we could "get right". Are they publicly documented anywhere? I don't think I've ever encountered a list in any of the books I've read.

The spring ought to be fairly efficient at reflecting the energy that the returning arm/boss/rod/balance-lever system puts into it. That system has obviously been designed to move reasonably freely and balanced so that gravity isn't affecting one side or the other excessively. The chain/cable at the bottom would offer little resistance to the recoil, just slackening off. So, to the naive imagination, a bit of bounce would seem quite natural if the signal man releases the tension in the cable faster than the system naturally returns to Danger - unless there's a damper in the Arm Stop itself. Hmmm...

<Takes a deep breathe before questioning The Stationmaster because my knowledge is miniscule and mainly from books...> The "Arm stop" in old pattern signals contained a sprung buffer to absorb the impact when the boss plate swung down onto it. That could explain a certain amount of bounce when arms return to danger, couldn't it? (Ref: GWR Signalling Practice page 82.)

Ladder, ladder connection to stage, ladder stays, rod guides and down rod:

4mm, AnyCubic Photon S. (Build volume 115 by 65 by 165mm. Z resolution 10 micrometres.) I realise that I will need to adjust the thickness of some parts but for now I'm creating the model using real-world dimensions as an idealised starting point.

Hi Chris, Small correction to what I said above: I looked up the reference and, apparently, the bottom 4ft of the ladders was painted white during both the First and Second World Wars (GWR Signalling Practice page 89). So, you could have white in your period if the ladders had not been repainted since 1918 but black would be safest...

Lieutenant Seagoon, you say the walls of Sebastopol are 20ft thick? Why do you say they are 20ft thick? Have you ever measured the walls of Sebastopol? They might, in fact, be only 10ft 6in thick. What happened to the other 9ft 6in? Lieutenant, are you blaming me for the war?

Hi John, It’s a lovely finish - nice to see a bit of shine on the loco like you see in the pre-war photos. Matt grime seems to be more a post-war thing. Has the nearside topfeed pipe popped out of its locating hole at the cab end?

Hi Chris, The bottom of the ladder was painted white during the second world war (to improve visibility during the blackout). Before that GWR practice was just black all the way up.

Hang on! You need to be able to fit your largest loco and at least one wagon into each spur to make it work, operationally.

Imaginary exchange trials?

Does anyone know how Dave is getting on in the current horrible situation? I'm sure none of us want The Model Shop to go under. I tried to send a supportive message on his web page but the email system seems to be broken. @Dave: I know you don't like RMWeb much but this would be a great way to keep your customers up to date and maybe we can help, if you need it.

That is wonderful, Paul! One small point (erm, ha ha): I don't think the trap points are needed because the crossovers would do that job themselves.

Two good books that go into a lot of depth are: "GWR Goods Train Working" Volumes 1 and 2 by Tony Atkins, published by Crecy.

I think the shed off the running line is OK but I agree that P1 should ideally come off the P2 loop because it would then unambiguously be a trailing connection. Whether that's feasible in the space is another matter! P.S. If we take away P1 as a passenger platform then there is nothing left of the former "terminus"...

Idea:

I have to say, it seems to be a bit confused! Why not think of it as a through station with a junction and some bays for terminating specific traffic? Then you would be on more familiar, prototypical territory. It also feels like there's too much track in every area to me. Maybe you could abandon platform 4 as a platform and use it for the goods loop/headshunt instead - to both simplify the plan and make the headshunt longer. And could you curve the goods yard around inside the end curve - i.e. make the station L shaped to gain some extra length? Use the P2/P3 loop to run round your goods, perhaps. (So you wouldn't be able to leave a train running on the top circuit while you run round a goods train but having done that the ex-P4 headshunt should be long enough to shunt the yard. And maybe make P2 be the through line and P1 just be a bay for terminating traffic? Edit: P.S. The shed area would have just one junction with the main line and a splay out from there. (Two roads would be plenty. Three if you really need it.)

Not specifically about signalling but I found some images of Moretonhampstead I hadn't seen before here: https://thetransportlibrary.co.uk/?route=product/search&search=Moretonhampstead&category_id=148&page=1 The photos were taken in the 50s when it was still a working station, although you can see that the engine shed was being used as a coal store by then.

Added the "Blinker" and a basic top stage in the Sketchup model. I have finally got Fusion 360 installed (the problem was with my Autodesk account) so I will have to move across to that but Sketchup is like an old friend...

Thanks. The signal arms I'm modelling at the moment were wooden. 12 inches wide tapering down to 11 at the boss and half an inch thick tapering out to 1 inch thick at the boss. (Opposing tapers in the two dimensions...) Half an inch in 4mm scale is 167 microns (if my maths is right) and that should be within the capability of my printer.