Shelf Marshes (first attempt at a cameo layout)

[Ian Simpson]

I may have missed this earlier in your writing, but what glue are you using? Some plastic cements seem to work better than others when the surface area is very small, e.g. on the base of railings. (My personal favourite for very fiddly plastic parts is Humbrol Liquid Poly.)

[47137]

23 hours ago, Ian Simpson said:

I may have missed this earlier in your writing, but what glue are you using? Some plastic cements seem to work better than others when the surface area is very small, e.g. on the base of railings. (My personal favourite for very fiddly plastic parts is Humbrol Liquid Poly.)

 

I haven't mentioned glues yet, and I don't think I've finished deciding. So far I've used three, none of them ideal:

1. EMA Plastic Weld (my favourite nowadays)

2. Carr's Butanone (because the Plastic Weld has no effect on the railings)

3. Revell Contacta in an old-fashioned tube (the best glue for the railings so far). Contacta also dissolves Tamiya paint, which is handy.

 

I used the Contacta for large parts like the tanks too, where there is a plug and socket style assembly.

 

I've never used Humbrol Liquid Poly. At the moment I have a liquid version of Contacta on order, but it would do no harm to order up some Liquid Poly in the meantime.

 

Edit: I'll wait now until one of the new glues turns up, in case this makes a quicker bond on the railings.

 

- Richard.

Edited May 8, 2020 by 47137

[47137]

I am beginning week 4 of the chemical plant with a new solvent - the liquid version of Revell Contacta. This has the consistency of weak wallpaper paste and goes on with a brush. And above all it works well with the yellow plastic Faller used for the railings. I have fixed back all the railings which fell off, and I have pulled off and reattached a few more. 

 

I have now made inroads into all but one of the fifty-odd sprues, gathered up some extra railings left over from other Faller kits, and filled a plastic bag with parts I reckon I will never need:

 

The fourth deck will go on today, and more railings. After these, construction reduces to adding the final ladders and pipes, and maybe adding some lighting.

 

50 useful hours logged now, not counting the five I spent faffing around with the base.

 

- Richard.

[steve1]

Just found this thread. It's a fascinating and original project which I will be following with interest.

 

Good luck with Faller kits!

 

steve

[47137]

21 hours ago, steve1 said:

Just found this thread. It's a fascinating and original project which I will be following with interest.

 

Good luck with Faller kits!

 

steve

 

Hi Steve,

 

Thanks for passing by. Sometimes this topic seems like a monologue but it still seems to  attract views.

 

I have one overwhelming fear for this layout - it is too small. I think it would be brilliant laid out on a 8 x 4 and very good indeed on 7 x 3. I want a layout of two square metres not one square metre. A larger size would let me have a 30-inch minimum radius, technically still underscale for a class 66 (4 chain minimum) but it would look so much better.

 

As it stands, it would still be good for an earlier period with 0-6-0 industrials. The clue is probably in the title of the topic! 

 

I want to redraw this plan for a 5 ft x 5 ft corner, but so far I always end up with an 18-inch curve.

 

- Richard.

[47137]

The main structure of the chemical plant has its lighting, photos on the SI Modelling topic so I could reply there to some people who helped me:

https://www.rmweb.co.uk/community/index.php?/topic/152919-si-modelling/&do=findComment&comment=3965538

 

Before this I spent three hours attaching the outstanding railings and then while threading in the LEDs I only knocked three of them off so Revell Liquid Contacta is pretty useful stuff.

 

With the LEDs done I can look at adding the final pipes and other details, and have a think about a better base large enough to hold the whole model.

 

I've settled on 'Shelf Marshes' as the name of the layout. The idea is to have most of the modelling towards the middle of the baseboard and an indication of empty marshland around it. This probably means, an empty seascape sort of a backscene.

 

- Richard.

Edited May 14, 2020 by 47137
Added the name of the layout

[47137]

The main structure of the chemical plant has become completely unwieldy since I attached its lighting so I have turned my attention to a base for the whole model.

 

I found an offcut of 1/8 inch three-ply under the bed and I have cut this to a shape to hold the whole model and still fit into the space available with a bit of leeway all round:

 

I reinforced it underneath with bits of stripwood and strips of the same plywood:

 

Really this is like a miniature baseboard, and it took me nearly as long to build! It is not especially strong but it is very rigid - I think fixing some supports diagonally helps this.

 

There are five different sizes of stripwood in here but I got some satisfaction from using up what I had at home without a trip to B&Q. The framework is fairly shallow along the middle because I am expecting this base to straddle a baseboard support when I build the baseboard.

 

My plan now is to add a battery box and a switch underneath so the building lighting works as a stand-alone model. Add some card to represent concrete onto the top. Then assemble the various parts of the model onto this.

 

"The end is in sight".

 

- Richard.

Edited May 18, 2020 by 47137
Edited for continuity

[47137]

At the start of week 5, I have cleared the untouched parts away and the build has largely reduced to fixing two dozen or so subassemblies onto the base and onto each other:

 

Only two of my original labels have survived this far. The scrap of purple post-it note is there to remind me of a weak joint I can break apart to adjust the positions of the large tanks. 

 

I cannot imagine what it would be like to tackle this kit for a few hours a week as 'hobby evening', or even alongside a full-time job - I know I would run out of momentum. I've now logged 71 hours on the build, including painting, making the base and installing its lighting switch and battery box.

 

- Richard.

[47137]

The model is going together so well, I thought I'd best post some photos of my progress yesterday. Otherwise the model will seem to go from a pile of bits to completion all in one go.

 

The detailing on the ground below the main structure is fiddly because you want to fix down the main structure before you add the details, but they have to go into the build at the very beginning. So here is one of the assemblies of details lifted out of the way while I glue the base of the "small retort"(?) onto the base:

 

The base is card and the glue soaked away and hid itself.

 

Fortunately things then got a lot easier. Everything here is now fixed down:

 

The rectangle at the right is the base of Kibri portable office, this model has been looking for a home for months.

 

This leaves me to add a few more major items; complete the lighting; and touch in the paint damaged by repeatedly re-gluing the bits of railings which kept falling off.

 

Incidentally, the jar of Humbrol Liquid Poly suggested by Ian Simpson finally arrived yesterday. Too late for the build but it looks like a runny version of Revell Liquid Contacta. I'm sure it will be useful.

 

- Richard.

[47137]

On 20/04/2020 at 16:57, 47137 said:

I shared this photo on the charity shop vollie support group early today:

 

And the ensuing discussion went like this:

20/04/2020, 09:28 - Anna (shop): Richard, there will have to be a grand unveiling!
20/04/2020, 10:00 - Richard: I think I'm bonkers to tackle such a thing, but I will certainly try to arrange a suitably socially distanced unveiling! Have a great day everyone [blue heart emoji]
20/04/2020, 10:08 - Anna (shop): I think the bonkers bit came at the time of purchase, too late now. You are in for the long haul.

 

:-)

 

- Richard.

 

 

 

One month on, it is "built". I have written out a short snag list for the model, and I am now pondering what it needs in the way of extras and details:

 

I've put another photo from this morning on the SI Modelling topic, I am hoping people will reply but maybe of course they are as innocent of the subject matter as me :-)

 

- Richard.

 

Edit: link is https://www.rmweb.co.uk/community/index.php?/topic/152919-si-modelling/page/10/&tab=comments#comment-3974313

Edited May 20, 2020 by 47137

[Allegheny1600]

Good Morning, Richard!

It is looking brilliant and I can't believe it's a month ago that you posted that 'exchange'!

You have certainly filled your time well and have a very industrial looking model to show for it.

Well done,

John.

[47137]

45 minutes ago, Allegheny1600 said:

Good Morning, Richard!

It is looking brilliant and I can't believe it's a month ago that you posted that 'exchange'!

You have certainly filled your time well and have a very industrial looking model to show for it.

Well done,

John.

 

It's been fun to take some photos in the early morning sunlight.

 

If I get too close I can find some awful bits, like this particular piece of railings which kept falling off and took far too much solvent to get it to hold:

 

(Also please excuse my use of horizontal H beam where it should be I beam but it is all I had at home and it is very small)

 

BUT I can also see how some film sets can come into being - if I can hide the holes in the steelwork this isn't far short of convincing:

 

There are 29 LEDs in the model, so there is scope for night-time shots too.

 

This project has taught me so much about comprehending the big picture and getting the right overall effect but I don't know how to get this across very well in words.

 

- Richard.

[47137]

I reckon, I have finished the chemical plant kit. It seemed to "become complete" as it were when I put in the last of the railings. I suppose, by then it looked complete; I had given up on the instructions and improvised for a few days:

 

My total time for the build was about 85 hours. Gulp.

 

This includes making the base (c. 7 hours!) and building-in and connecting the lighting (c. 5 hours). It also includes my efforts to make the decks (floors) stronger, to neaten up the edges of the floors, and re-attaching railings which kept falling off until I got some stronger solvent.  It also include many hours spent painting parts along the way. I haven't included the time spent backtracking from the first base, or building the portable office building (Kibri) ... I suppose these would take me over 90 hours.

 

I've put a few more photos on the SI Modelling topic, links here to avoid cross-posting:

Completed model

Night shot

Workers on level 4

 

I have missed out some of the internal pipework from the kit. I might put some of this in one day, but my efforts now are minor tidying-up tasks. Really, this means hiding things a more skilled person would have avoided in the first place :-)

 

- Richard.

[Ian Simpson]

Many congratulations, Richard: that's craftsmanship and perseverance! It looks fantastic.

[47137]

16 hours ago, Ian Simpson said:

Many congratulations, Richard: that's craftsmanship and perseverance! It looks fantastic.

 

Ian thank you for your kind comments. The kit was a bit of an eye-opener for me, I somehow remember the plastic kits of my youth as things I glued together in a few evenings!

 

It's a clever kit - it persuades you to think it looks like a real oil refinery, even though it's the size of a couple of good-size detached houses with their gardens, or perhaps a block of inner-urban flats. Footprint as built by me is 40 x 25 cm, this scales up to about 35 x 22 metres. The kit instructions take you to a squarer model.

 

If anyone fancies tackling the kit, I would buy two of them and unify them into a single plant. A refinery with duplicated sets of primary plant on the ground level and duplicated retorts would look quite business-like, and wouldn't be dwarfed by a rake of tankers. The build time would probably be half as long again not double, the hardest part will be interpreting the instructions.

 

However, I have my 5ft 4in alcove, and either I will build my industrial scene to fit into it, or reduce this section of my layout to something like a sweeping viaduct in open countryside. I would love to build "Shelf Marshes" about 7 ft long with a 30-inch minimum radius; but if I do this it won't slot into the scheme for "Shelf Island" in a topologically sensible way unless I build it as a peninsular coming into the middle of the room. This is physically possible but too challenging to live with ... life was bad enough having a picnic table of Faller bits in the middle of the room for a month :-)

 

- Richard.

[47137]

Something quite fortuitous is the way I have ended up with a gap through the middle of the model, between the tanks and the steelwork (same photo as I posted yesterday):

 

 

So on the layout, I will get a glimpse of a train making its way into the scene, but not the overhang of the buffers on the two-chain curve. 

 

- Richard. 

Edited May 24, 2020 by 47137
Replace png image with jpg

[47137]

The mock-up of the project now looks like this:

I have removed the loco spur to make room for a pair of large vertical storage tanks, these will hide the exit from the layout to the right.

 

I have also made the cardboard "wings" a bit taller. The viewing portal (thinking of this with a cameo style presentation) is going to be fairly tall, but I don't think this will do much harm. The baseboard will probably end up 27 not 24 inches deep, so a taller viewing window will complement the shape of the layout.

 

A non-modelling friend has a suggested a nature reserve. Supposing I represent this with a hide and a car park, I can put cars in the car park to show the period represented. My collection of stock means I can't be too fussy about period at the moment, and the rest of the "scenery" is pretty timeless for the last 50 years.

 

Both of the entrances to the layout are concealed by the wings, without overbridges. There are no civil engineering structures near the track either, so the project suits me to run British outline in H0, 1:82, 00 and even conceivably S-16.5 or 0-16.5 too. Although, if it ever makes it to a show, I'll try to moderate such excesses :-)

 

- Richard.

[47137]

Driving the white LEDs in the Chemical Plant

 

I have wired up the LEDs in the chemical plant so I can drive them from two AA batteries or from a regulated supply (driven from an external source):

 

This lets me take photos of the model with its lights on outdoors and without trailing wires, or show it to other people. The circuit board here is a DC to DC converter using the LM2596 power converter, these boards are incredibly cheap to buy and they include a display to show input and output voltages.

 

With no series resistor, the LEDs looked about right running on 2.5V and consuming 50mA from the circuit board, and rather too bright running from the batteries. Two fresh batteries giving about 3.1V together.

 

So I ordered up some 10 ohm resistors to try to lose about half a volt from the batteries (V = I x R) and I have connected one of these in series with the supply to all of the LEDs:

 

The actual voltage drop across the resistor measures 0.4V, and the appearance running on batteries has calmed down nicely. I’ve increased the output of the regulator board to 2.8V, this gives a nice appearance for use on the layout.

 

The LM2596 can drive up to three amps, so sixty chemical plants! Realistically, I think I’ll buy a few more of these boards for other parts of the layout so I can tweak the brightness for different installations.

 

“Footnote”!

The forward voltage of a white LED is about 3.4V. More than this will shorten the life of the device, or indeed destroy it straight away. Having 30 LEDs painstakingly glued into the model and no series resistor, I checked the output of the voltage regulator very carefully before I connected it up. It would have destroyed the lot as received, before I turned down its output voltage.

 

- Richard.

[47137]

Input and Output

I've returned to the chemical plant to give it some loading and unloading facilities. These are from a Kibri kit for an oil storage tank:

The input is on the left and has a pipe disappearing towards the primary refining plant on the ground floor of the plant. The output is on the right and has a pipe connecting it to the holding tank. This is all extremely sketchy and I doubt it has a prototype, but I think it makes the plant look more business-like. Also it provides the terminal for rail tankers. The lights are N gauge ones from Kytes Lights, extended with bits of brass tubing to make them taller.

 

I can add some hoses to the input and output after the whole model gets bedded onto the layout. The plain grey area can be "tarmac" but again this is best done with the model on a baseboard. Right now I can't work up much enthusiasm to build the layout itself, though I'm working my way through deciding how to control the points. I'm having far more fun doing the buildings for it.

 

- Richard.

[47137]

Facilites for Passenger Trains

 

My design for the layout has a flaw for the passenger services. I cannot have a passenger platform on the spur in the corner ('B' in this photo) because there is nowhere to shut in the train:

 

If I want to follow British practice to protect the passenger train, I need to move the passenger platform off the baseboard to the left of location A, and add a trap point between the platform and A. The stub branch / spur at B becomes a place to stable the passenger train between turns. There is a discussion here - it's quite fun to see how I managed to ask myself the right questions after some knowledgeable people tried to answer the first ones:

https://www.rmweb.co.uk/community/index.php?/topic/155383-connecting-a-freight-only-line-without-using-signals/

 

Fundamentally, there is no benefit to me in making a nice model of something which could never happen on a British railway. So the passenger platform goes, and the spur at B becomes just another siding.  The "Shelf Marshes" section of my layout becomes an entirely freight-only operation, and the only passenger train here has no passengers. The trap point at A marks the end of a block section connecting the scheme to the rest of my layout, and "Shelf Marshes" can be as I want it, without signals to get in the way of shunting operations. Trains leaving here to go onto the rest of the layout should pause at A to collect a token.

 

I can ponder whether to provide a goods platform at B (parcels?) or maybe some non-railway scenic feature. I've bought a kit for a low-relief block of flats, and this might tuck into this corner.

 

- Richard.

Edited June 9, 2020 by 47137

[47137]

A week on I have done no model-making worth writing about, just a lot of soul-searching about how I should be controlling the points.

 

The proposed plan of the layout now looks like this:

 

In particular:

• The lengths of the two tracks marked x are important because they need to be long enough to let me shunt the exchange sidings;
• I've reinstated the loco spur bottom right, it will look quite pleasing with a Peco 83 line no.4 wye point  (it's such a shame Peco don't do a "medium radius wye" for their mainstream Streamline ranges);
• There is a maintenance facility for the tram, so I can contain the OLE to the extreme left of the layout - so this goes top left, where the passenger platform was.

 

I would like to provide manual controls for all of the points, but this has snags - the control knobs will clutter the fascia, and the linkages will have to cross over each other if I am to put the control knobs in a logical sequence. So after a lot of soul searching, I've ordered up some servos and a control board from Megapoints. This is due to arrive tomorrow.

 

I've tried solenoids, stall motors, bicycle spokes and wire in tube on previous layouts  ... servos are completely new to me, something fresh to try. The idea is to operate the servos from local switches on the right-hand wing of the fascia. The first tasks are probably to get the board up and running, and see how to best switch the frogs. 

 

- Richard.

[47137]

A top-level plan

 

This layout is going to be a complete pain throughout its build, because it will have to go into the middle of the hobby room. I will only be able to tuck it into its alcove after it is finished.

 

So I am going to try to make every subassembly on the bench - everything be it a control panel or a building - before I even buy the timber for the baseboard. The mock-up can stay on its flush door throughout this activity. I can rearrange the mock-up to my heart's desire, know what the finished layout will look like, and maybe even to manage to built the layout quite quickly.

 

I suppose, the obvious downside is the project stays a mock-up for ever ...

 

- Richard.

[47137]

Servo mechanisms

I have spent a week+ faffing about (experimenting?) with servo mounts. I made three prototypes, the first two with a microswitch for the frog and the third without:

 

All of these move point blades cleanly and reliably.

 

Unfortunately, my overwhelming conclusion is I while can arrange a mechanism where the microswitch works some of the time, even most of the time, I cannot get even 99% reliability and I want at least three 9's and probably four. So I settled on the third arrangement, a piece of 15 mm aluminium channel as suggested by Dave Fenton of Megapoints in one of his many videos.

 

This is the sweetest, quietest and neatest point actuator I've ever had. I am using a fairly long throw from the servo horn because I am thinking of 6 mm ply for my baseboard when I build it. Also the piano wire here is 0.4mm diameter, rather thinner than suggested so it flexes easily at the end stops:

 

The result doesn't need any setting up beyond using the controller board to centre the servo to find the mid point for the movement.

 

I've made another seven of these:

 

And so I now have the mechanisms for the eight points in the plan:

 

The ninth servo here is for a notional gate or barrier. I've abandoned the loco spur again.

 

This leaves me to focus on how to switch the polarity of the frogs. Really, there are only two ways with any appeal to me: double-pole switches on the control panel (one pole for the servo controller, one pole for the frog) or a frog juicer. I have a collection of old-fashioned switches - a panel from H&M and some Lucas toggle switches from 1960s cars. These would build up into a control panel with some character, but they are single-pole and would need the frog juicer.

 

Conversely, if I use double-pole switches (servo + frog) I deny myself the ability to use the serial interface of the control board in the future. I have a hankering to try for some semi-automated operation in which trains approaching a turnout from the trailing direction trip a sensor and the point blades move to the favourable direction. This would be a project for the Arduino. I've bought a couple of IR sensors for it. But somehow, I think I need to cut my teeth on something less critical for my first Arduino project. This is another way of saying, crib from someone else! I've not done any programming since Delphi 4, and I never achieved a great deal with that even though it was supposedly easy. Programming isn't really my thing unless I can find something mostly ready-made and then butcher it.

 

For Shelf Marshes it would be sensible to aim for a system able to work without some half-baked personal Arduino application. This rather rules out using the Arduino to drive relays to switch the frogs. I know there is also a route in which I add a relay driver board and then two relay boards, all available off the shelf from Megapoints, but if I end up with a row of toggle switches needing four circuit boards to drive eight frogs I will feel I have lost the plot.

 

There is also the matter of keeping compatibility with analogue (DC) trains. I have a pet theory I can use one frog juicer to drive every frog on the layout because the turnouts are too far apart for any of my locos to bridge two frogs. The obvious constraint would be I must never stop a loco on a frog if I want to then move a second loco. In such a scheme I would simply disconnect the frog juicer and run the layout with dead frogs for analogue operation. This has some appeal because most of my locos are now DCC, and the exceptions are things like a 12-wheel Dynadrive loco which will hardly be troubled by a few inches of dead frog. Then again, something like my 00 gauge 2-EPB (and DCC) is definitely going to bridge 2+ frogs so this won't run on the layout.

 

For me, every new layout has got to stretch me with something I haven't really tried before, with a view to seeing it through in a successful way. The main board of Shelf Island has its compound curves and gradients, and Fairport has its visual balance, mirror and unconventional baseboard. The trouble is, controlling the points is really fundamental to success.  It might just possibly be best to start with something I know I can build and get to work: eight double-pole toggle switches.

 

- Richard.

Edited June 30, 2020 by 47137
typo

[47137]

The die is cast

I really do want to try something different for the baseboard construction. I have looked at the Combitech tubular baseboard system but I think I will be happier with something more generic.

 

I have found an aluminium wholesaler happy to cut sections to length, and they have my order for 1" square aluminium tubing with a cutting list. I have tweaked the dimensions of the layout so there will be next to no waste:

 

1 in x 1 in x 16 swg - Aluminium Square Tube

Length 2500 mm

Grade 6063T6

Custom Cut

Item 1 of 5 - Cut
Length: 1580 mm Quantity: 1
Length: 710 mm Quantity: 1
Item 2 of 5 - Cut
Length: 1580 mm Quantity: 1
Length: 710 mm Quantity: 1
Item 3 of 5 - Cut
Length: 1580 mm Quantity: 1
Length: 710 mm Quantity: 1
Item 4 of 5 - Cut
Length: 710 mm Quantity: 1
Length: 600 mm Quantity: 2
Item 5 of 5 - Cut
Length: 1580 mm Quantity: 1

Offcuts

send

 

Their cutter takes 5 mm so there are some 200 mm offcuts to support the two wings.

 

So - I've designed my own baseboard kit in about ten minutes flat. Assembly will be screws or pop rivets and Araldite. Cost is about £60 including cutting - not a negligible sum, but probably only about £30 more than the equivalent in pine strip wood. Spruce would be a lot cheaper but the stuff I find for sale is too unseasoned and too unstable to build a baseboard.

 

I suspect, 19 mm square section aluminium would be amply strong. I went for 1 inch square because two pieces will give me enough depth to hold my servo mechanisms.

 

Postscript:

I've ordered up the ply for the track bed and backscene boards too. This is on a 30-day lead time. The cost of the aluminium and ply for the basic baseboard will be about £60, another £40 for the backscenes and a dust cover, plus unfortunately most of £40 carriage because I am using two suppliers. Still, a total bill of £140 is bearable compared to the cost of a baseboard kit.

 

- Richard.

Edited July 3, 2020 by 47137
Added postscript.

[47137]

Route setting

 

The MegaPoints servo controller board has three memories to provide for simple route setting. You set the points (servos) to the route you want and then press and hold one of three buttons to store the route in memory. You can then continue to operate the layout as desired, changing the points from their levers (switches) to suit. Pushing the button again sets up the stored route and disables the point levers, and there is another button to release the memory and restore control from the point levers.

 

Anyway, to try it out I needed more wiring and switches than I want to assemble together so I improvised with a breadboard and a load of wire links:

 

The memory functions seem to work perfectly.

 

I think this will be useful on a small layout like this - one route for through movements to/from the fiddle yard, one route for arrivals, and one route for the run-round loop.

 

(The Arduino here has nothing to do with the circuit on test, it merely has a breadboard attached to its base)

 

- Richard.

Edited July 9, 2020 by 47137