A test on producing track with concrete ties
If you do model Belgian railways, you may also want to model concrete track such as in real use in Belgium of course. There do exist commercial track systems with concrete ties of but these are all of undefined origin, however none of these are suited for code 40 rail with 9 mm track gauge. This does also hold for track using the German B70 sleeper which is widely used, not only in Germany, since 1970. 50 year old sleeper designs (and even well over) can not be called modern track, this track form dates from before the introduction of N gauge itself, thus it is strange that there are no good presentations.
Development
A first thing for development is to find more information and particurlarly drawings. Even for the well known German B70 sleeper this turns out to be a task as the original drawings are hidden in some official standard that can only be bought for a lot of money at some standardisation institute. Lucky I found a useful drawing in a master thesis available on internet. For the NMBS track I didn't find such. Studying photo's it looks like every photo in my entire book collection that shows some form of concrete track has a different sleeper! Obviously the NMBS was experimenting a lot to find the right approach. However recent photo's in current magazines clearly show less variation. The main differences are a division in two forms with either single concrete ties or so-called bi-bloc ties with 2 blocks connected by some means, generally a steel profile. In my own photo collection there are some examples of bi-bloc ties, but next time I need to remember to take also ruler and notebook with me for penning down some dimensions. The more recent concrete track seems to have preference for single block ties but of a more straight design than on the photo below not having the waist in the center in horizontal plane.
A photo showing both forms of concrete track taken in Ciney 2014.
Close-up of the bi-bloc form with pandrol clips.
This is another form of bi-bloc showing a relative short concrete block having a more rounded shoulder, it also shows a bit of the connecting steel T-profile normally hidden by ballast.
This drawing by U. Lamalle from 1951 shows a bi-bloc called Franki-Bagon with some useful dimensions although not complete.
In the book of Ulysse Lamalle (see reference at the end) we find some drawings of several older systems. It also shows a drawing of the French Sonneville system, for which I also found photographic evidence in my photo books. I choose the Franki-Bagon because it will be slighty quicker to produce. Using the dimensions from above source we can draw up something that looks like this type of sleeper.
A 3D drawing of the Franki-Bagon sleeper, actually only 1/4 is drawn and the rest is simply mirrored and after that can be copied into multiple sets.
Typical is that these concrete ties with 2400 mm are much shorter and at least 2x heavier than wooden ties, but the distance between sleepers is generally much closer. For the SNCF I found a reference of 1722 sleepers per km (for welded track) which comes to 58 cm. For the drawing I used 60 cm giving 3.75 mm distance between centerlines. Although the drawing of Lamalle shows an Athus-Meuse chair plate on top of the sleeper, nearly all photos that I studied show pandrol clips such as those on the two photo's on top. The main problem in model is of course the fixing of the rail as these pandrol clips cannot be replicated in scale with sufficient strength to hold the rail. Thus here we are forced to follow a different approach, therefore the 3D drawing shows a small block on top which should somehow get an opening to hold the rail. The maximum height for these clips will be 0.4 mm in order not to get problems with wheel flanges touching the tops. This will automatically lead to a relative weak construction in comparison with rail soldered to copperclad sleepers, but only time can show if this will become a real problem.
This drawing gives the dimensions and shape of a DB B70 drawing.
The same can be done for the B70 sleeper. This 3D drawing needs an update as the resulting clips for holding the rail are too small.
Production
Initial approach
For testing purposes small scale production can be done by milling and casting or 3d printing. I choose for the first option as it is available in my studio and the 3D printing means a more costly outsourcing. When feasibility is shown the 3D printing route can be explored for larger scale reproduction when several meters of track are on top of the wish list. Currently those ideas are only slumbering.
Thus the above drawings were fed as STL files to Deskproto, which nicely generates CNC files per appropiate milling cutter. In this case 0.5 mm conical for roughing and 0.5 mm straight for finishing were used. The control program of the milling machine takes care of replicating the pattern into a concatenated series of 5 sleepers. (5 is just my choice, it can do any other number within reach). This pattern was milled from 2 mm thick polystyrene, a time investment of about 30 minutes for 5 sleepers (partly optimised thus not much more to gain here). A total of 3 of these patterns were used for producing 2-component Silicone rubber moulding blocks. A set of these rubber blocks allow to cast a small series of patterns in one go. The amounts of resin needed are minimal, but the fast settling time of the resin is limiting. In order to fit the rail, the only thing left to do is, to cut the gap for the rail itself. Several different methods were tried here but until now the most reliable one is using the milling machine again, now mounted with a small Woodruff cutter with reduced height to fit the rail cross section better.
Current State of the art for the Franki-Bagon bi-bloc track showing moulds and resulting track. The insert shows the holding clips on the inside rail in ideal form. Various anomalies can be observed, resulting from enclosure of small airbubbles to ripping off during the milling process, both are not an immediate problem and can be solved by appropiate measures. The main goal, that of feasibility, is provided by this exercise. However replicating a small series of track pieces made it clear that this was not the ideal approach. It concerns tiny amounts of PU that are difficult to weigh and mix correctly. Filling the moulds needs to be done quickly leaving not enough time the get a nice clean and flat back. This meant that the pieces need to be cleaned up to be able to fit them in the holder for cutting the groove for the rail. Even when this is done with quick methods it involves too much time and effort. Also the track pieces need at least 24 h to harden improving their strength considerably before making the groove. With only about 50% succes rate in casting track pieces for a bit of demo track the need for a quicker approach became obvious. Making better double sided rubber moulds and using vacuum may solve the problem of the back and air enclosures but I choose different.
Modifications in the milling process
The main reason for the long milling time is of course getting the shape of the sleeper correct which requires a small cutter and lots of small steps in the order of 0.02-0.03 mm. The milling process can be faster if we introduce a special cutter that can cut the sleeper cross section in one go. Compare this with cutting coach roofs where, using a dedicated fly cutter, you can do half a roof in a single cut. Thus this approach was tried by producing a simple fly cutter from a piece of 1.5 mm steel by filing and grinding with a Proxxon minidrill. This piece of steel was mounted in simple holder, the same as used for coach roofs, and, hey presto, we now create the sleeper profile with only 4 cuts, 1 for each shoulder for a line of blocks. This also allows the sleepers to be cut in a 2.5 D process using larger cutters and steps instead of complex 3D CNC'ing with a single small cutter. The only drawback is that we now need a series of different cutters requiring time for tool changes. In total 5 different milling cutters are used for producing the bi-bloc sleepers. The time for tool changes can be considerably optimised by doing a large series of sleepers in a single production run. The 4 pieces used for the demo track of 21cm took about an hour production time which means that for a single module the time investment will be in the order of about half a day, which I find acceptable.
Above we see 2 stages of the development process, on the left blocks that were still a bit too large and right the final block form but still needing some fine tuning. The sleepers now sit on a base making them considerably more rigid, in order to built curved track cuts can be made to flex the whole base but the final design is using only half this number as we certainly don't need anything below 800 mm radius. These track bases are cut from 2 mm PS and the height of track comes out as 2.6 mm thus to connect track with 2mm SA pcb sleepers we need 0.8 mm under layment below the wooden sleepers and/or shift the coupling holes in the module end to compensate this height difference. In theory about 0.5 mm height can be spared in the ballast height and using a thinner base but the current sleeper height conforms the drawing.
The cutting pattern using colours for different cutters and cutting depths.
From right to left 3 stages in painting and ballasting.
Above we see the production of some 20 cm demo track stuck to a base of wood with double sided tape. The right end shows the painted base in tan colour, the steel connections in rust and grey primer for the concrete blocks. The center is partly ballasted up to the height of the steel connections and on the left you see fully ballasted track.
Closeup of demo track.
This photo shows a good impression for concrete bi-bloc track which of course entirely differs from that of wooden sleepers. The demo track was hand painted when already stuck to the base and with the rail mounted. Obviously painting will be much quicker before mounting, certainly the base can be done by spraycan or airbrush. Using a dedicated template from thin film with cut outs for the concrete blocks will allow the block to be directly airbrushed over the tan base. In theory you can also first mount the strips on the base board using a 0.5 mm rulers in the rail groove to line up the various strips and than ballast before mounting the rail, but this might be one step too far. However it may make ballasting somewhat easier but certainly faster.
Doing what it is meant for.
Other concrete track types can be made in the same way, but will need production of a dedicated cutter as the profile will be different of course. The stand time of these home made fly cutters won't be too long as they are unhardened but as long as they are only used in soft plastic then several 100's of ties should be possible. You can see in the enlargement of the closeup a small difference between the CAD drawn sleepers and the current production method in that the top edges of the CAD drawn sleepers are more rounded, but I find that an acceptable compromise.
First production run on a short curved module with about 50 cm track to 1200 mm radius showing the latest home produced wagon, a NMBS Hbikks
A series of 14 pieces with concrete ties were cut for a small module for a practice test. Several 100 ties turned to be a far too optimistic expectation, several 10's was maximum. Clearly the flycutter toolbit need to be redone from hardened steel as I had to resharpen it 3 times, effectively it lasts about 1m cutting length. A dull cutter clearly influences the result as the surface that it leaves isn't smooth. I will try to make a new cutter from a thin piece of hard steel for instance from a steel saw cutting blade, a suitable tooling bit for the lathe or maybe a broken milling cutter.
For the rest the procedure is similar to the demo track but this time the lot was first painted before the rail was fitted. This fitting required some care as a length of 25 cm seemed about maximum needing quite a bit of force with the rest flexing away. Some form of jig to keep it straight while fitting the rail might be helpful here or alternatively widen the cut a tiny bit for the foot of the rail. Gluing down was done with 2 component epoxy using a set of wooden jigs from 0.8 mm birch ply to keep the 1200 mm radius. Some care need to be execised in soldering the dropper wires as the ps easily melts when dwelled too long with the soldering iron. But using one of my brass gauge jigs in combination with a drop of water on the nearby claws sufficed to keep the ps intact.
copyright: Henk Oversloot
date: 20 Aug 2018
updated: 13 Nov 2018
literature
With thanks to Rixke's for his rail archive. This publication from 1951 holds important information about track in Belgium and can be downloaded from 'Spoorse Zolder'
Also thanks to Heiko for supplying this link : Brochure with information over concrete ties used in Germany by DBAG
