Designing the Route
We have a long thin garden, depressingly huge by modern standards, but a bonus when it comes to hobbies like garden railways. However, having such a large canvass brings it's own dilemmas.
Starting points were,
Gardening practicalities are important - especially if the wife, gardener, and local planning authority are all one person. We settled on the mid section of the length of the garden being dedicated to the railway. In this way there was no conflict (yet) between the railway and the flower beds at one end or the veg plot at the other.
Existing garden beds, trees and structures need to be considered even if only to plan to [re]move them. Also consider how you are going to get the mower, or wheelbarrow, to the areas you need to in the future.
You can do a lot 'by eye', even better with a hosepipe or two marking out the approximate layout for days on end to see what effect it really has. However, you really need to get a reasonably accurate drawing down on paper - you will be surprised what your garden plan really looks like.
For this you need a long tape and either a lot of graph paper, or, in my case, a simple CAD program. I used a simple shareware one CadStd (http://www.cadstd.com) which is fairly basic (and it's User Interface has nothing to commend it) but then the needs of this job are quite simple. The main value of the CAD package is to maintain a readable document, and maintaining multiple variants, whilst you are bouncing the design options around.
Locate and draw up the immovable objects: fences, trees etc. and make sure you have all of them!
Survey the ground levels. Our garden is remarkably flat rising only a few inches over 50 yards (but then, this is Cambridgeshire). Work out what one point is going to determine the 'zero' height reference. For us, it was that that the circuit around the shrub bed was only just above lawn level, which also happened to be the level of the concrete base of the small metal shed I'd installed as a store for the railway.
The observant will have noticed that it would probably have been best to finalise the layout before laying large concrete shed bases in it's way. Wonderful stuff, hindsight.
Beware of trees. Trunks are reasonably easy to plan around but carefully consider the consequences of running the line under the canopy. There is the obvious 'leaves on the line' problem but a quick brush can solve that. More consideration needs to be given to fruit trees (we have a number of plum and apple trees in the target area). How are you going to harvest the fruit without risking standing on the track? Chances are you'll want to mow under those trees to make finding fallen fruit easier, so allow space for that. Even if you're no fruit lover, rotten fallen plums/apples don't half make a mess of track.
As do pigeons.
There are three basic options. 1. dead flat, 2. gradient for the sake of having a gradient, 3. gradient for a purpose - e.g. you want to have the route turn and cross itself.
Dead flat is boring. One of the garden railway owners we visited commented on his regret of not having any gradient at all, which would have made running his steam loco a more interesting challenge. That said - it's probably worth making sidings level unless you have working brakes on your wagons!
Beware of trying to cross the same line. You need a long long distance to be able to achieve this comfortably. Our original plan had such a loop, but even with an unrealistically thin bridge (i.e. thickness between the top of the lower aperture to the top of the upper track), and an 8ft radius curve, the resultant gradient was somewhat agressive. Aggressive gradients are feasible on straights with typical garden railway loco's but don't look good, but on a curve the tractive effort required is multiplied by the sideways friction on the train wheels.
This table is the result of a small spreadsheet I used to sanity-check ideas about radii and gradients.
Gradient (1 in N) for a circle of radius R requiring to reach a height H.
So, for a loop of radius 60 inches, to reach 9 inches height you need a gradient of 1:42.
In any case, I didn't really have the space to keep to both the desire to have a bridge crossing and maintain large minimum radii. At it happens, this was a stroke of good fortune as, when I did eventually run a scale 1:20 loco around the track, albeit a large one, it didn't fit under the 'tunnel' I'd mocked up and would never have fitted under the planned rail bridge! The tunnel was easily removed - the bridge and gradient track based around it would have been a very different matter.
Bridges make an attractive feature but you do have to plan carefully for them, especially if, like us, you have a flat garden. With a gradient of 1 in 60 you do need 20 yards to get a mere 12 inches into the air. Don't forget to take into account the vertical thickness of the bridge itself. Ironically, the location where we wanted a bridge for visual reasons was at the slightly higher end of the garden, so I was fighting the gradient all the way. I'd recommend do it the other way round if you can!
Look for operating interest. A single loop is 'OK' but that's about all. Plan in a passing loop so that running more than one train is fairly easy. Points/Turnouts are expensive so you may hesitate to have many sidings - even so, allow for them in the plan so that you're not stymied by the horse chestnut when you come to want more later. It's even worth at least considering what large scale expansions might be possible and, at least, try not to exclude them with your layout.