Planning a new train room
The big picture: Reader's input.

After telling a few interested people about the useful emails I have got, It was suggested to me that I really should publish a few. So here they are, minus email addresses, and irrevelentant parts.
Quick index : This index is ordered from the most recent backto the oldest.
The text below is ordered the other way, starting from the oldest to the recent...
FROM : Robert Stanley 24/3/2001

Thanks for sharing. I remember being impressed sometime last year when I first encountered your site, and it is fascinating to see your ideas evolving and growing.

I'm in exactly the same situation as you (apart from living in Canada and working in N-scale!), in that I have real space, and I have ideas going back years, and I have a strong sense of what I want, and the practical experience to know what I can achieve. I'm also a computer geek, and feel that having the computer able to control many if not all of the layout functions is an absolute key feature of the whole system.

I'm a little further along than you are, but I don't have a web site. I must admit that I've been meaning to do a web site, but haven't been entirely sure of a design that would reflect both my short and long term goals for information sharing. I'm also hideously aware that a web site is capable of sucking all the time there is if you aren't careful.

Your web site is an inspiration as to how to present the design information; I can only hope that it acts as a spur to make me actually get on and create the pages. For looks at the layout itself, I guess there are several really well-done sites out on the web that I'll be happy to crib from. The Niagara and Pearl Creek springs to mind: ; I really like the notion of touching an area on a diagram and seeing what's there in fact. The Edmonton Model Railroaders are a club who also have an excellent virtual tour:

Anyway, I just wanted to say thanks for sharing, and seeing where you and I have adopted similar solutions to a common problem, and where we have diverged. I, too have both a mine branch and a narrow gauge (Nn3) branch: mine is logging. My logging branch is basically on the main (lower) level, but has a vast switchback across the back and sides of the C-shaped foot-print, up to the top level. Yup, really short trains!


Back to the top

FROM : John Dennis, 30/3/2001

Hi David,

Good to see you moving into planning the layout in its space. You certainly have ambitious plans, and I would recommend trying if at all possible to build your layout in phases, so you don't have to spend years before even seeing anything run.

I do like the way that generally an operator can follow his train pretty well. That is important on a large railway designed for operation.

I wish you luck in this venture ... it makes my Dutton Bay plans look so very simple and straightforward.


John Dennis. Melbourne,Australia
Home of the HOn30 Dutton Bay Tramway
and the Australian Narrow Gauge Web-Exhibition Gallery
Dutton Bay URL:

Back to the top

FROM : RobertStanley 31/3/2001

Comments : This was a good email, one that has looked at what I have shown, and provided valuable input, which has helped me chamge my plans, and the site as well....

Well, I have to say that the new plans make it much, much easier to understand your thoughts and ideas for your railway. So much so that I'm now going to complain about stuff ... :-(

Actually, no complaints, but I do have a few suggestions, and one or two questions about things that aren't clear.

(1) It took me a long time to figure out that you are working in HO-scale. The association with the Eltham Model Railway Club and the fact that it is an HO-scale club was the primary clue. As far as I can tell, you never anywhere explicitly state that it is an HO scale layout that you are planning; the Part 2 curve specification of Mainline 34inch radius minimum, 36inch normal is a strongish hint. Oh yes, the term is used in your 26th of November, 2000 update, talking about DCC decoders.

(2) Your original room diagram is the other way up from the orientation of all the new track plan drawings. I'm still having some trouble matching the track plans to the space and dimensions of the room diagram. I hand draw all my diagrams, too, and I have ended up putting a one-foot rule (alternating black and white bars) on both the X and Y axes of my drawings. It really helps when you are inspecting the drawing on screen, as opposed to working off a print-out. When I went back to the dimensioned room plan in Part 1, and tried to use the dimensions, they turned out either to be inconsistent (or incomprehensible to me). For instance, if I try to figure out the total vertical dimension, on the left I have a 22.4 and a 4.6.5, which gives me 26.10.5 or 27 feet approximately. However, working from the figures on the right hand side, I see 21.5 and 7.1, which give me 28.5. This is a big discrepancy; certainly, it's one that could really ruin your day when trying to convert a plan to real construction.

(3) The first thing that I examined in detail was the helix, in part because yours sounded cool in the earlier discussions, and in part because the three that I ended up with had all been such bears to get right.

My first question is to heights - in Part 2 all that you have to say is "at least 16 inches apart". The next is the radius of the helix track, and the number of turns needed to make the climb. This is where I had a hard time trying to dimension the helix from the diagrams. The floor plan shows 2.9 + 2.11 as the width of the stairs and the space between the stairs and the wall, for a total of 5.8 feet. Your lower diagram has a clear representation of the helix entire fitting within this width. Assuming one inch clearance outside the outer rail of the outer track, this gives us a radius for the outer track centreline of 64 inches, or a radius of 32 inches. The inner track will therefore be no more than 30 inches and maybe less (I use 2-inch centreline spacings in N-scale). Postulating a 4-inch rise per turn of the helix, we get a grade of 1.99% on the outer track, and 2.12% on the inner, which are really not bad at all.

The next question is how you are planning to do the break-out onto the middle level. If (as your route stick diagram suggests) it is possible to bypass the middle level (F), then I think you have to have a double crossing on the curve in the helix. This is going to be a very interesting piece of trackwork to construct. If you can find any way to do it, you just might want to make the helix footprint oval, separating two semicircles with tangents parallel to the wall between E and B. Putting the crossings on a stretch of tangent track will be way easier on both construction and operation. This is what I have done on all three of my helices, although mine are triple tracked, and have Y junctions between the centre and the two outer tracks on each circuit. These are all stacked vertically above each other in an easy reach from an aisle for maintenance.

If there is no bypass, which I think is what the middle layer plan shows, then the trains pass around the middle level plan in transit between the top and bottom decks in both directions. Now you have the double junction as shown in the bottom right-hand corner of the middle layer. Again, as I read the scale of the diagrams, that complex trackwork is going to be a screaming bastard to reach, and it looks like something that you're going to have to lay by hand. Speaking from expensively-learned personal experience, this is worth laying out on paper to full size and making absolutely certain about. I hand-lay my own pointwork, and use a program called TemPlot to do it: . You can also simply photocopy some commercial points and crossings, if that's what you are planning to do. Either way, I mock up the actual benchwork by cutting the plan out of heavy cardboard (from boxes), using the rest of the box to make spacers to get it at the right height, then shuffle the full-sized track diagrams around on the benchwork mock-up. Anything even slightly dodgy shows up instantly.

My last question on the helix is about the interchange from the middle level to the upper level. From lower to middle is clearly clockwise running, and the track leaves the middle for the upper maintaining the clockwise running. However, the double track seems to emerge on the upper level already split into two tracks in the centre of the helix footprint, at an angle that cannot conceivably be reached from the rim of the helix. Your stick diagram suggests that there is a helix between the middle and the upper levels, in which case the plans for the middle and upper levels just can't be made to work together as drawn. If there isn't a helix, but the track simply climbs around the middle level to emerge on the upper, then I think you have a significant grade problem.

I think you have a real problem with which way round trains are travelling on the helix as they climb or descend, and I half feel that the sense changes on either side of the middle level. Been there, and it took me ages to spot it. Given what you are trying to do on the upper level, it's probably the other two that will have to be rethought.

(4) Staging tracks. Your stick diagrams show staging at both end, and I finally found the "exit to upper level staging" in the top left corner of the upper level track plan. This appears to correspond with the staging beyond K shown on the stick diagram. I guess that this is going to be on the other side of the sliding doors.

What I can't find anywhere on the lower level plan is the staging shown beyond the B-A fork after C. I do see the single track to the right of B, paralleling the B-C route, and apparently doing a runner into the helix area. Is that doing a long hidden run out to the same place as the other staging, or is there going to be staging in and under the helix? Either way, I think you may start to have access problems to the helix. Sod's law states that if there is one place on the layout that is just out of reach, that's where trains will derail, probably tearing up the track in the process.

Access to trains on helices is really tricky. I made sure that my N-scale ones had a four inch rise per turn, simply so that it is physically possible to stick an arm into the loops and fish trains out. Mine are triple-tracked, on two-inch centres, so the roadbed is roughly seven inches wide. One has access from all four sides (runs from 36 inches above the floor to 51 inches); it's twin has access on three sides, but also has a wriggle space into its centre to reach the far side from the inside. The other is a monster, starting at 53 inches above the floor and climbing to 67 inches. This one is fully accessible from the inside, and has an operating station inside it, accessible via a duck-under at a height of 4 feet, which is a bastard, but one that people can manage.

Bottom line, nothing that you need to reach should ever be more than a couple of feet from a comfortable position.

Heck, you know all of this from the club, but I'm having trouble seeing it from the plans.

Did I say "nice layout"? Well, it sure looks it.

(5) Where is the operator at D going to be located? You have a very restrictive choke point right at D, which looks like 18 inches or so. Basically, people look like 2-foot diameter circles when plotted onto a track plan. Eighteen inches is turning sideways and sucking your gut in when you have one. Worse, it looks as if there will be an operator for that great engine facility of A right about where the operator for D is going to be. Then you have trains on the B-C and C-D-E runs passing on either side of the choke point. It's going to be a bad day when two trains with accompanying operators try to dance through during a busy shunting session at both A and D...

I had huge problems with aisles until I started making full-size mock-ups out of old cardboard boxes. Intellectually, I knew that anything less than 30 inches for aisles was going to be a problem, but you keep squeezing the aisles to get the track right. Don't. It took me a lot of restarts of the plan, but I ended up with a basic aisle width of 36 inches, and no chokes less than 30 inches. Major revision number 22 solved all the problems, and actually resulted in a far better plan, but I sure got sick of throwing wonderful plans away. I urge you to mock up the critical parts of the benchwork and trackwork full-sized. You don't need to do it in situ, you just need some 1-foot gridded butcher paper on the floor to provide accurate reference points.

(6) Where on the plan is the narrow gauge branch that you talked about in Part 2?

Anyway, looks like an amazingly cool plan. I wish you every success with it. And thanks for sharing all of this via the web.

Robert Stanley

Back to the top

FROM : Derick Cullen 12/5/2001

Thanks David

Looks like you have a lifetime project on your hands.

I have "glanced carefully" at the MK3 plans. Two things immediately struck me (apart from the huge scope of the thing) .....

(1) Without knowing how you plan to operate trains, I wonder about operator/train dynamics. On such a layout I would want to walk around with my train, not watch it from one or more vantage points. In this context one of the exits from the main station will disappear and re-emerge in a completely different direction after a substantial time, The situation reminds me of a tutourial on peninsular design by Allen McClelland in his RMC series on his Virginian and Ohio, reprinted in a book on the V&O. I'm sure this same issue comes up in other places on your plan.

(2) For a lifetime plan, stuffing things through that doorway, in a couple of layers already, seems a compromise which could be eliminated. Any way to knock down that wall before things get too far advanced?


Back to the top

FROM : Robert Stanley 9/11/2001

Luckily you seem to have been as busy as I, so I don't feel bad about being months adrift.

With your third plan (which is vastly more practical than plan 2, to my eyes) I see that you are starting to wander down a path which caught me, too. The problem is balancing all the features that one wants against the practical limits imposed by the physical space, as well as the probable operational constraints.

I'm delighted to see that you learned much earlier than I that entire features can be displaced and altered without in any way diminishing the spirit of the plan. That being said, let me comment on this plan as you have presented it. I have two sets of comments, the first addressing practicality and the second addressing operational usages.

As far as practicality goes, this looks hugely easier to work than did plan 2, from both construction and operating points of view. If I'm reading your ideas on height correctly, the upper staging will be sufficiently above head height that it will be possible to step under it. This will be important as you try to reach into the triangle at the root of the G peninsula. This I think remains one of the practical weak points of the design. First of all, you have some very complex trackwork on the bottom wall, which is going to be at the end of a practical reach from the toe of either of the two aisles. Secondly, in the left hand aisle you have the potential for collision between:

(1) Station operator at D
(2) Wharf operator at E
(3) Junction/yard operator at the toe of the same aisle
(I know, probably the wharf or D station operators)
(4) Through train driver on the F line
(5) Station operator at G above E
(6) Yard operator at the yard above D and F
(7) Through train driver on an I - triangle - H route

That's potentially seven bodies who have a legitimate need to be in the left hand aisle together. There are four who have a legitimate claim to be essentially permanently resident if all the operating features are simultaneously active.

My rule of thumb for aisles is that they are always too narrow.

For two people of average adult size, a 1 metre aisle width is enough to provide comfortable passage; any narrower, and people start turning sideways and sucking in their guts. For one person, 75 cm is adequate for operating, but not really if you have to step back or duck down to get at something well below waist height. Similarly, if you have to climb on a stool or something to reach way above normal eye level, more is needed. If two people are going to be working the opposite sides of the same aisle, and routinely passing behind each other, you need 2 x 75 cm, which is 1.5 m! Finally, if a third person needs to pass between such a back to back pair without disturbing them, you need another 50 cm. That's a 2 m aisle...

Yes, of course you can skimp, but you have to consider the operational consequences.

Note: in my final plan I have no aisles that require more than two people to be abreast, and I have essentially no back to back operational areas. Where there is a chance of people clustering I have achieved 1.5 m aisles, although most of mine are pegged at the 80 cm to 1 m mark.

My other mechanical issue concerns the helix. I think the new location is much better than the one on plan 2, and I think that the entries, exits, and branches off can all be made to work with much more ease. However, I note that some of the entries/exits/branches are exactly at the rear of the helix, which is only accessible from about 1/3rd of its circumference at best. I'm particularly uncertain as to how the route from H actually enters the helix. To make matters worse, the top of the helix is capped to support the scenery and track over its centre reached from G, on the upper level. This will make crawling into the centre of the helix from underneath a job for the slim and limber, bearing in mind that they won't have the headroom to stand upright once they have wriggle in...

All the rest appears to have easy reach from the edge of the layout, and to be entirely practical to work on and with. It took me way more than three plans to reach that point.

Given that you can build it, if you build it, they will come.

You are certain to end up ultimately with an operating crew. This introduces the question of what the layout is for. As it stands, it looks like an enormous amount of fun for one or two people to potter around in, either driving a train from end to end on one of the various routes, or stopping off at a feature to do some shunting for a while. However, it is less clear that there is any real purpose for the railroad in terms of modelling a believable transportation system.

The staging at either end is well balanced, in that there seem to be six roughly equal length tracks at each end. I think that this is light for such a large layout, but that is based on my focus on North American railroading. It is entirely possible that this capacity is fine for what you are modelling.

A staging yard allows you to create exactly two operational scenarios: to introduce traffic onto the layout that has originated somewhere off stage, or to act as a destination for traffic that originates on stage on the modelled layout. When you have two or more staging yards connected to (and by) the modelled layout, you introduce a third scenario, namely through traffic that both originates and terminates off the stage of the modelled layout.

On the layout you have created with plan 3 you can therefore have the following types of traffic moving between distinct locations:

  • (1) Staging U to Staging L (and vice versa) through traffic. This can be of any type, and may take a variety of routes, with or without on layout stops en route. If without stops, it starts to provide a moving hazard around which other types of traffic must be routed.
  • (2) Traffic originating at Staging U and terminating on the layout. Different types of traffic will presumably find their various ways to different destinations.
  • (3) Traffic originating at Staging L and terminating on the layout. In the abstract, identical to the previous type, but in practice will depend on the nature of what the two staging yards actually represent.
  • (4) Traffic originating on the layout and terminating off it at Staging U. This is the inverse of point 2.
  • (5) Traffic originating on the layout and terminating off it at Staging L. This is the inverse of point 3.
  • (6) Traffic originating at one point on the modelled layout and terminating at another.

These are what one might describe as the multi-location operations. There are also single location operations, which are typically shunting or special purpose such as MOW or the operation of some form of large scale loading/unloading like a flood loader at a mine, or a train ferry at a dock.

What this introduces is the need to create a sense of purpose to justify the railway that you have designed. Where are the upper and lower staging yards on the larger world stage, and what traffic do they generate and receive? The same then needs to be done for each location on your layout. What does the location represent, and what traffic does it generate and receive.

Traffic needs to be expressed by type, quantity, and frequency. The type dictates the type of rolling stock needed to transport it. The quantity speaks to the quantity of rolling stock and the size of the typical train needed (although these may be constrained by the physical dictates of the railroad) and hence the motive power. The frequency dictates the number and size of trains and sometimes the speed at which they must travel.

There is a considerable sub-culture within the model railroad world that devotes much of its life to this, and there is some pretty good software out there to help you simulate and model operational traffic patterns.

You say in one of your web pages that this plan isn't the killer plan for which you are looking. Part of the problem may be that you don't have a well-define purpose for each of the features on your layout. Let me rephrase that, and say that you have as yet not published any such information, which leads me to believe (based on my experience) that this was probably not what gave rise to the plans so far created.

You really have to be able to ask why a train would originate at any possible point of origin on your layout (including one of the staging yards) and why it would travel to another. In answering that question you will start to have a pretty good idea of what sort of trains will be running that route, and how frequently.

As an example of how thinking about this influenced my plans, one of the major features that I'd been determined to have was a steel mill - a large steel mill. I didn't have any real issue with the finished product going out, because that can always be sent off stage; the real problem was the materials coming in. Why I had a problem with this is that I'd always wanted to have a huge ore dock as well, but that's for shipping ore mined nearby to a remote mill. If I have an ore source and a steel mill on my layout, then I don't have any reason for an ore dock. Equally, I really wanted to have a harbour with the potential to put a big ore carrier in it. It sounds stupid in retrospect, but I was about a dozen plans into the cycle when it finally struck me that I could have an ore unloading dock, and run the unit trains of ore hoppers between the dock and the steel mill.

I still want to build an ore dock, but now I have the challenge of building Hullett unloaders, and I have a reason for all of the dock, the unit trains of ore hoppers, and the steel mill to be on the layout. At the detail level, I can work out how much iron ore has to arrive daily to support the size of steel mill I am modelling, and break this down into the number of hoppers per day, and thence the number of trains of whatever number of hoppers per 24 hours. This in turn has to be matched to the size of bulk ore carrier at the dock, and how quickly one can be unloaded.

Working out how to justify a feature in the end helped me to put together a plan that felt more satisfactory. For every stretch of track, and every siding, I knew why my railroad had laid the track, and why they continue to maintain it in service. I also know why some of the track is now marginal, and why the management might like to see it gone.

Anyway, just some more thoughts inspired by spending time with your plan 3, for which many thanks.

Back to the top


Page written & maintained by David Head.