I've been dabbling in hand built track for many years, though never built a running layout. The points worked and that was that. My bibles for hand building track were two books by Iain Rice; An Approach to Building Finescale Track in 4mm Scale, by Iain Rice, Wild Swan Publications, 1991 and A Pragmatic Guide to Building, Wiring and Laying PCB Track, by Iain Rice, A Railmodel Handbook, 1997. Although somewhat long in the tooth they are nevertheless as relevant today as they were when written. I would whole heartedly recommend both, as I found that some parts that I did not understand in one book, I understood better in the other. More recently there was a general series of articles by Norman Solomon in Model Railway Journal, Nos 143, 144 and 146, 2003 followed by Making Tie Bars, Model Railway Journal, No 151, 2004

 
My interests are in light railway practice (Colonel Stephens) and I construct my track and points using PCS in strategic places and balsa wood elsewhere using code 100 flat bottom rail. My test points had stretcher bars made out of wide PCS strip with point blades soldered to the strip and installed as described by Iain Rice in his 1997 book. The drive mechanism (wire in tube or similar) was via omega loops soldered to the stretcher bar. This method I found satisfactory in the test mode, but I was concerned about the longevity of the soldered joint and I now wanted to use under-board motors such as the Tortoise.

 
In his 1991 book Iain Rice describes cosmetic stretcher bars and dropper wires reaching through the base board to the drive mechanism (his 'patented' turnout operating unit (TOU)). He also discusses surface stretcher bars as does Norman Solomon, and both advocate the use of glass fibre reinforced PCS strip, very narrow for cosmetic purposes but wide enough for a hole to be drilled in the middle to take a point motor drive wire. Norman Solomon also advocates the use of dropper wires (lill pins) that are free to revolve and so put no strain on the point blades or on any soldered joint.

Both of the above solutions are shown diagrammatically in drawings A and B.

In both cases I was worried about the point blades riding up over the stock rail (I found it difficult to prevent this during construction, though on final assembly the soldered stretcher bar of course prevented this problem.) I wanted to use lill pins as dropper wires so I came up with the solution shown in drawing C which shows an above ground stretcher bar. Note that the PCB is turned copper side down, and sleeves soldered to the till pin from underneath. Care must be taken not to solder the sleeve to the PCB, so some cladding needs to be removed, or you'll defeat the object of the exercise, and this also insulates the tie bar. The reason for having the PCB upside down is to avoid soldering the lill pin to the board whilst soldering it to the foot of the point blade. Any strong insulating material may be used in place of PCB. I would also advocate making the sleeves shorter than the protruding part of the pin, so that when you turn the point upside down to solder the sleeves on, you know when they are pushed fully home.

So far so good, but I still had two problems. I was unable to source glass-reinforced PCB locally, only the much weaker phenolic resin based type, and although I had achieved my objective of effective stretcher bars, I was really impressed by Iain Rice's cosmetic stretcher bars, and since rod was used in my prototypes that's what I really wanted, all at 0.5mm diameter. So I adapted the Rice method shown in drawing B, to give my final refinement shown in drawing D. The functional tie bar is taken down to below track level and buried in the underlay (cork, or in my case thick neoprene rubber). Below each point blade there is small rectangular cube of PCB built up from single and double sided material. The bottom full-length stretcher bar can now be quite wide to give it robustness, but believing in belt and braces I also soldered a small block of brass to the PCB strip to engage with the point motor drive wire. It is essential to make sure that as the stretcher bar moves from side to side the PCB cube supports remain fully under the stock rails, otherwise the blade is not restrained from rising above the stock rail and the block may jam on the side of the stock rail when the turnout is switched.

The rationale for using a combination of single and double sided PCB is for ease of manufacture with materials already on hand for point construction. This method allows the whole assembly to be soldered together, rather than using screws which may be difficult to hide. A solid block could also be filed up from a solid piece of insulating material if you are so inclined. I wasn't!

Views of the final assembly from above and below, complete with cosmetic 0.5mm rod stretcher bar are shown in photos 1 and 2. The cosmetic rod is glued in place (Araldite) with small plastic sleeves for insulation and to give some outward grip whilst the glue is setting.