Model Railways On-Line - Littlehempston - Part 7
Littlehempston Part 7 - Preparing for Track Laying By Paul Plowman The reliability of a model railway depends largely on the quality of the track. In this respect I have set myself a difficult challenge because I have chosen to work to P4 standards. There is a widespread belief among railway modellers that because P4 has a shallower wheel flange than OO and EM it is less reliable and more prone to derailments. This belief is based on the logic of simple geometry. However, there are other factors coming into play, which possibly make P4 more reliable than OO. The depth of a P4 flange is 0.38mm. In simple geometry a wagon of 40mm wheelbase could not derail unless the track twisted at a rate of 0.38mm in 40mm, i.e. approx 1 in 100. Now this is a severe twist by anyone’s standards equivalent to twisting a 4ft long baseboard by just over 35 degrees. Well baseboards just do not twist that much and good quality (12mm) plywood does not have local defects that bad. The problem comes from adopting Peco recommended practice of using wood fibre insulation board or other fibrous materials, which come under various trade names. These materials should not be used on any account for model railways not even as track sub-base supported by plywood or particle board. These materials will swell if dampened with water based paints or glues as I have in the past found to my cost. Localised swelling of these materials can easily cause a greater than 1 in 100 track twist. Unlike OO there is only one P4 standard for wheels and track (S4 standards excepted). Peco points are a compromise for all the different OO wheel profiles currently in use by modellers. With P4 no compromise is necessary and because of the much narrower flangeways in switches and crossings there is insignificant wheel drop in common crossings making for much improved reliability. With Peco track the rails are mounted vertically. Coned wheels only make contact at the gauge corner of the rails. I am using P4 flexi track from Exactoscale, which has the rails prototypically inclined inwards at 1 in 20 to match the coning of the wheels. This results in a wider area of contact between wheel and rail giving improved adhesion and electrical contact. Improved adhesion gives a better ride with less hunting and consequently improved reliability. I have not found the quality of ready to run models from Hornby or Bachmann to be a problem with respect to parallelism of axles in the vertical plane. If pin-point axles are running in plastic bearing any minuscule irregularity will be gradually eliminated by wear on the two bearings carrying the greater load. If the bearings are of brass or other bearing material some compensation can occur automatically provided the pinpoints are not a tight fit in the vee of the bearings. If the track twists then the two diagonally opposite bearings carrying the greater load will slip down over the pinpoints. If for example the axle has 0.5mm of side play then for a 45-degree coning of the bearing, the axle could move up by 0.5mm. Two axles thus allowing a 1mm defect or 1 in 40 twist to be negotiated. If we add the flange depth to this calculation a 1.38mm defect or 1 in 29 twist could be negotiated before derailment occurred. Surely no one lays track that badly! My fleet of rolling stock will for the most part be converted propriety r-t-r models and I have no plans to provide compensated suspension. It is therefore necessary for my track be to a high standard with respect to alignment, longitudinal level and cross level to achieve the best possible reliability of running. To achieve a high standard does not necessarily require a high degree of skill but it does require consideration of the methods of construction from the floor of the model railway room to the top of the rails. At every stage I have sought to minimise movement but at the same time I have adopted what I believe to be practical methods. Movement or warping at any stage in the structure will have a negative impact on the quality of the track. The floor My model railway room is built on a concrete slab for stability. Suspended timber floors move as people walk across them and they gradually deflect more and more as the load increases during construction of the layout. Ideally a model railway should be built standing on a solid concrete floor. Supporting structure All timber will shrink and warp to some degree. That is a fact of life. Using large, thick timber will not solve the problem. Ingenious solutions of using sandwiches of plywood and timber blocks are a step in the right direction but for anything more than small layouts these methods involve a significant amount of work. I have chosen to build a steel frame from Handy Angle™ and Dexion™, which will not be susceptible to shrinkage or warping. It is quick and easy to erect. It can be adjusted by slackening nuts and bolts and it is easily recycled for the next layout. It is also cheap. All of the material used for Littlehempston was second hand and the whole lot including nuts, bolts and gusset plates cost A$100 (£50). Above: Completed supporting structure built from Dexion and Handy Angle. Baseboards There are numerous methods of baseboard construction but it is my view that best results can be achieved if timber is avoided altogether and the layout is built in one permanent piece. I believe that the best method of construction is traditional open frame but using Handy Angle steel instead of timber. However, it was necessary for me to consider the possibility that we might relocate at sometime in the near future. With this in mind my layout needs to be ‘moveable’ rather than ‘portable’. Baseboards have been constructed to a larger size than would otherwise be required for a portable layout. Primarily the aim has been to simplify construction and reduce the number of joins the tracks have to cross. The baseboards have been constructed using 12mm plywood for the tops, sides and cross members. The boards sit on the steel frame supported by packing pieces at their corners. Adjacent boards rest on common packing so that their vertical relationship can be restored should they need to be moved in the future. Boards are bolted together but I have not made any attempt at horizontal registration. I constructed the boards as carefully as possible but I found it impossible to achieve a level surface to the standard I was seeking. Fortunately track will only be fixed directly to the boards in the hidden sidings. Throughout the whole of the visible section the track will be 4.5 inches above the surface of the baseboard. I carefully checked the surface of the boards in the hidden sidings with a spirit level and straight edge to identify low spots. These low spots were raised with layers of cardboard from Cornflake packets, glued down with PVA. Preparing the track bed Throughout the visible section the track bed is supported on wooden blocks. These blocks are glued to the baseboard surface and screwed from underneath. They are spaced at a maximum of nine inches apart. I took a lot of time using a builder’s spirit level and a steel straight edge, taking great care to ensure that the tops of these blocks were as near to being perfectly level as possible. On curves there are two blocks at each location fixed to the baseboard surface with a third sandwiched between, which has been fixed in position for longitudinal level and cant (superelevation). Above: Completed support for raised track bed. The embankment will carry three tracks. The nearest will not be canted. The middle piece of wood is inclined to carry the two canted tracks. The rebate in the right hand piece has no significance and is due to the wood having been recycled. The blocks of wood are glued to the baseboard with PVA and screwed from below. Above: Blocks of wood stand ready to support the track bed. I have used 7mm plywood for the track bed. I considered using 12mm plywood but it would not have been flexible enough to be twisted through a transition as the cant increases or decreases. Also my layout will include three underline bridges. If I had used 12mm plywood topped with a track sub-base the depth of construction of the bridge decks would have become excessive and unrealistic. The track bed on each baseboard is cut from a single piece of plywood. Joins only occur where baseboards meet. This strengthens the structure and minimises irregularities in the track. The plywood track bed has been screwed and glued to the timber supporting blocks. This task must be completed for the whole layout before any track laying is even contemplated. It is not uncommon for modellers to construct one or two baseboards at a time and complete the track laying on each before commencing another baseboard. This approach is satisfactory when working with coarse OO standards but not when attempting to achieve the highest possible standard for P4. Once the track bed is installed one must crawl all over it with a straight edge and spirit level. Special attention must be given to the areas around the baseboard joins. Low spots were identified and again packed with thin cardboard. Only when I was satisfied with the entire layout did I move on to the next task. The Sub base For a sub base I am using a layer of cork. The cork has been obtained from the local hardware store in the form of 6mm thick floor tiles. There was a discussion in the BRMA e-group, which I initiated, regarding the most suitable glue for fixing the tiles down. Members were generally of the opinion that Bondcrete™ or PVA were the best option. It was however pointed out to me by another member that PVA takes a long time to dry and needs to remain weighted down during this period. It was suggested that I should use a contact adhesive such as Selley’s Kwik Grip™ (Identical to Evostic in the UK) as this would enable the work to be completed much quicker and with a better result. Again I crawled all over the layout with a straight edge. Even after all the previous effort there were still high and low spots! High spots were dealt with by sanding down the cork, while one low spot needed an additional intermediate support to be installed between the track bed and the surface of the baseboard. Again only when I was satisfied with the entire layout did I move on to the next task. Above: The plywood track bed has been attached to the wooden blocks and a cork sub-base has been added but not yet painted. The track is loosely laid Peco code 100, just to get something running. My previous layout was ruined when diluted PVA being used to fix ballast soaked into the fibreboard sub base causing it to swell. Having proceeded thus far with my new layout taking extreme care to get everything, as near perfect as possible I did not want to have this layout ruined the same way. There seems to be divided opinion as to whether cork will absorb water and swell. For the sake of a coat of paint it is not worth taking the risk. I have painted the cork with two coats of a brown acrylic paint. In addition to sealing the cork it provides two other useful advantages. Firstly it is easier to draw on with a pencil than the bare cork and secondly if any ballast should become detached the brown colour, which is not unlike track colour will be exposed. Laying the track Finally at long last I have reached the stage where I can lay the track. I have decided to use point kits and flexible track supplied by the P4 Track Co. In my next article I will describe my experiences laying the track at Littlehempston.
The reliability of a model railway depends largely on the quality of the track. In this respect I have set myself a difficult challenge because I have chosen to work to P4 standards.
There is a widespread belief among railway modellers that because P4 has a shallower wheel flange than OO and EM it is less reliable and more prone to derailments. This belief is based on the logic of simple geometry. However, there are other factors coming into play, which possibly make P4 more reliable than OO. The depth of a P4 flange is 0.38mm. In simple geometry a wagon of 40mm wheelbase could not derail unless the track twisted at a rate of 0.38mm in 40mm, i.e. approx 1 in 100. Now this is a severe twist by anyone’s standards equivalent to twisting a 4ft long baseboard by just over 35 degrees. Well baseboards just do not twist that much and good quality (12mm) plywood does not have local defects that bad. The problem comes from adopting Peco recommended practice of using wood fibre insulation board or other fibrous materials, which come under various trade names. These materials should not be used on any account for model railways not even as track sub-base supported by plywood or particle board. These materials will swell if dampened with water based paints or glues as I have in the past found to my cost. Localised swelling of these materials can easily cause a greater than 1 in 100 track twist.
Unlike OO there is only one P4 standard for wheels and track (S4 standards excepted). Peco points are a compromise for all the different OO wheel profiles currently in use by modellers. With P4 no compromise is necessary and because of the much narrower flangeways in switches and crossings there is insignificant wheel drop in common crossings making for much improved reliability.
With Peco track the rails are mounted vertically. Coned wheels only make contact at the gauge corner of the rails. I am using P4 flexi track from Exactoscale, which has the rails prototypically inclined inwards at 1 in 20 to match the coning of the wheels. This results in a wider area of contact between wheel and rail giving improved adhesion and electrical contact. Improved adhesion gives a better ride with less hunting and consequently improved reliability.
I have not found the quality of ready to run models from Hornby or Bachmann to be a problem with respect to parallelism of axles in the vertical plane. If pin-point axles are running in plastic bearing any minuscule irregularity will be gradually eliminated by wear on the two bearings carrying the greater load. If the bearings are of brass or other bearing material some compensation can occur automatically provided the pinpoints are not a tight fit in the vee of the bearings. If the track twists then the two diagonally opposite bearings carrying the greater load will slip down over the pinpoints. If for example the axle has 0.5mm of side play then for a 45-degree coning of the bearing, the axle could move up by 0.5mm. Two axles thus allowing a 1mm defect or 1 in 40 twist to be negotiated. If we add the flange depth to this calculation a 1.38mm defect or 1 in 29 twist could be negotiated before derailment occurred. Surely no one lays track that badly!
My fleet of rolling stock will for the most part be converted propriety r-t-r models and I have no plans to provide compensated suspension. It is therefore necessary for my track be to a high standard with respect to alignment, longitudinal level and cross level to achieve the best possible reliability of running. To achieve a high standard does not necessarily require a high degree of skill but it does require consideration of the methods of construction from the floor of the model railway room to the top of the rails.
At every stage I have sought to minimise movement but at the same time I have adopted what I believe to be practical methods. Movement or warping at any stage in the structure will have a negative impact on the quality of the track.
The floor
My model railway room is built on a concrete slab for stability. Suspended timber floors move as people walk across them and they gradually deflect more and more as the load increases during construction of the layout. Ideally a model railway should be built standing on a solid concrete floor.
Supporting structure
All timber will shrink and warp to some degree. That is a fact of life. Using large, thick timber will not solve the problem. Ingenious solutions of using sandwiches of plywood and timber blocks are a step in the right direction but for anything more than small layouts these methods involve a significant amount of work. I have chosen to build a steel frame from Handy Angle™ and Dexion™, which will not be susceptible to shrinkage or warping. It is quick and easy to erect. It can be adjusted by slackening nuts and bolts and it is easily recycled for the next layout. It is also cheap. All of the material used for Littlehempston was second hand and the whole lot including nuts, bolts and gusset plates cost A$100 (£50).
Above: Completed supporting structure built from Dexion and Handy Angle.
Baseboards
There are numerous methods of baseboard construction but it is my view that best results can be achieved if timber is avoided altogether and the layout is built in one permanent piece. I believe that the best method of construction is traditional open frame but using Handy Angle steel instead of timber.
However, it was necessary for me to consider the possibility that we might relocate at sometime in the near future. With this in mind my layout needs to be ‘moveable’ rather than ‘portable’. Baseboards have been constructed to a larger size than would otherwise be required for a portable layout. Primarily the aim has been to simplify construction and reduce the number of joins the tracks have to cross.
The baseboards have been constructed using 12mm plywood for the tops, sides and cross members. The boards sit on the steel frame supported by packing pieces at their corners. Adjacent boards rest on common packing so that their vertical relationship can be restored should they need to be moved in the future. Boards are bolted together but I have not made any attempt at horizontal registration. I constructed the boards as carefully as possible but I found it impossible to achieve a level surface to the standard I was seeking. Fortunately track will only be fixed directly to the boards in the hidden sidings. Throughout the whole of the visible section the track will be 4.5 inches above the surface of the baseboard. I carefully checked the surface of the boards in the hidden sidings with a spirit level and straight edge to identify low spots. These low spots were raised with layers of cardboard from Cornflake packets, glued down with PVA.
Preparing the track bed
Throughout the visible section the track bed is supported on wooden blocks. These blocks are glued to the baseboard surface and screwed from underneath. They are spaced at a maximum of nine inches apart. I took a lot of time using a builder’s spirit level and a steel straight edge, taking great care to ensure that the tops of these blocks were as near to being perfectly level as possible. On curves there are two blocks at each location fixed to the baseboard surface with a third sandwiched between, which has been fixed in position for longitudinal level and cant (superelevation).
Above: Completed support for raised track bed. The embankment will carry three tracks. The nearest will not be canted. The middle piece of wood is inclined to carry the two canted tracks. The rebate in the right hand piece has no significance and is due to the wood having been recycled. The blocks of wood are glued to the baseboard with PVA and screwed from below.
Above: Blocks of wood stand ready to support the track bed.
I have used 7mm plywood for the track bed. I considered using 12mm plywood but it would not have been flexible enough to be twisted through a transition as the cant increases or decreases. Also my layout will include three underline bridges. If I had used 12mm plywood topped with a track sub-base the depth of construction of the bridge decks would have become excessive and unrealistic. The track bed on each baseboard is cut from a single piece of plywood. Joins only occur where baseboards meet. This strengthens the structure and minimises irregularities in the track.
The plywood track bed has been screwed and glued to the timber supporting blocks. This task must be completed for the whole layout before any track laying is even contemplated. It is not uncommon for modellers to construct one or two baseboards at a time and complete the track laying on each before commencing another baseboard. This approach is satisfactory when working with coarse OO standards but not when attempting to achieve the highest possible standard for P4. Once the track bed is installed one must crawl all over it with a straight edge and spirit level. Special attention must be given to the areas around the baseboard joins. Low spots were identified and again packed with thin cardboard. Only when I was satisfied with the entire layout did I move on to the next task.
The Sub base
For a sub base I am using a layer of cork. The cork has been obtained from the local hardware store in the form of 6mm thick floor tiles. There was a discussion in the BRMA e-group, which I initiated, regarding the most suitable glue for fixing the tiles down. Members were generally of the opinion that Bondcrete™ or PVA were the best option. It was however pointed out to me by another member that PVA takes a long time to dry and needs to remain weighted down during this period. It was suggested that I should use a contact adhesive such as Selley’s Kwik Grip™ (Identical to Evostic in the UK) as this would enable the work to be completed much quicker and with a better result.
Again I crawled all over the layout with a straight edge. Even after all the previous effort there were still high and low spots! High spots were dealt with by sanding down the cork, while one low spot needed an additional intermediate support to be installed between the track bed and the surface of the baseboard.
Again only when I was satisfied with the entire layout did I move on to the next task.
Above: The plywood track bed has been attached to the wooden blocks and a cork sub-base has been added but not yet painted. The track is loosely laid Peco code 100, just to get something running.
My previous layout was ruined when diluted PVA being used to fix ballast soaked into the fibreboard sub base causing it to swell. Having proceeded thus far with my new layout taking extreme care to get everything, as near perfect as possible I did not want to have this layout ruined the same way. There seems to be divided opinion as to whether cork will absorb water and swell. For the sake of a coat of paint it is not worth taking the risk. I have painted the cork with two coats of a brown acrylic paint. In addition to sealing the cork it provides two other useful advantages. Firstly it is easier to draw on with a pencil than the bare cork and secondly if any ballast should become detached the brown colour, which is not unlike track colour will be exposed.
Laying the track
Finally at long last I have reached the stage where I can lay the track. I have decided to use point kits and flexible track supplied by the P4 Track Co. In my next article I will describe my experiences laying the track at Littlehempston.