Sunday, April 30, 2006

How I ballast my track

Five methods have been used over the years to fix the ballast in place:

1. Using cement
2. Using PVA adhesive
3. A combination of the above two methods
4. Using tile adhesive 
5. Wet mix plus dry dressing for station areas

Each approach has its merits and, it seems to me, that the choice of method is largely dependent on cost, purpose and desired outcome.
The ballast I have used for my 45mm gauge mainline trackwork was 2mm alpine pink gravel, which I found in my local garden centre in 'handy packs'. Each pack is sufficient for around 12 - 18m of track dependent on the method used.

In station areas, I have used a wet mix of cement together with a dry-mix top-dressing of sifted soil, sand, crushed coal and bird grit.

For the 32mm copper mine feeder, I have used bird grit into which was mixed crushed pottery shards. My reasoning being that the miners will have used the freely available sandstone spoil from their mining as ballast.

1. Using cement

I have found this method the simplest to apply. The bag of gravel was mixed with builders' sand in the proportion 2:1 (ie a bag of gravel to half a bag of sand). To this was added the equivalent of half a bag of cement. In other words gravel:sand:cement = 2:1:1. I wanted a fairly strong mix to ensure the gravel would be held in place.

Once the ingredients had been thoroughly mixed, it was trowelled on to the track:

The ballast was then carefully brushed into place with an old paintbrush:

Care was taken to ensure there was not a build-up of ballast on the inside of the rail and also that it was kept well away from point mechanisms.

Once the ballast was in place, it was given a gentle dowsing from a watering-can fitted with a very fine rose. It's important not to wash the ballast out from where it has been carefully brushed or to dowse it so much the cement powder is sluiced away. I could have let nature take its course if light rain was forecast, but a downpour would have been disastrous before the cement was set.

The resulting ballast give the appearance of narrow gauge track which has been ballasted cheaply with grit rather than gravel.

The cement method is fairly easy to do. It's not particularly time-consuming once the ingredients have been mixed and is not too expensive. A bag of cement is relatively inexpensive compared with PVA and the addition of a portion of sand makes the ballast stretch further. However, the end-result is less attractive than the pure PVA method (see below) and resembles ballasting seen on poorly maintained narrow-gauge railways. I have found it easy to lift the track with this method, the ballast crumbles away.

Update four years on
I thought you might be interested in how this approach has weathered over the past four years:
As you can see, the ballast is probably in place (hard to tell) and a healthy growth of moss has covered it all. Personally, I'm more than happy with the moss as it gives the track a slightly neglected look, which is fine for the type of railway I'm modelling (a struggling British light railway in the early 1930s). Those wishing to model something more affluent (eg a Swiss railway) may prefer something less verdant.

2. Using PVA

The PVA approach gives a result which resembles carefully ballasted and well maintained track:
The gravel is used 'neat' from the bag and, as with the method above, is trowelled and brushed carefully into place:

Once the ballast in in position, it is dowsed in water from a watering-can fitted with a fine rose - being careful not to sluice away the gravel. The gravel is then flooded with diluted exterior / weatherproof PVA adhesive, to which a small quantity of washing-up liquid has been added to encourage it to flow into the gravel.

 The PVA method takes longer and is more expensive. The PVA adhesive is not cheap and even though it is diluted, does not stretch far. However, the result (in the short term - see below)  is more attractive as it resembles carefully maintained, well-ballasted track. Once set, the ballast becomes a fairly solid mass and hence I will imagine it will be more difficult to lift the track. I have read elsewhere that attempting to use interior PVA, although less expensive, is not a wise move as it very soon deteriorates outdoors.

Update four years on
Again, how has this method fared?
As can be seen, there's some, but not a great deal, of the ballast left in place. Whether the frost has broken it up, or birds has dislodged it, I can't be sure. But considering the additional effort which needs to be put into this approach I'm not convinced the longer-term return makes it worthwhile.

3. The combined PVA/Cement method
When I extended the line (see How I built the extension), I mounted the track on wooden supports covered in roofing felt. To ballast this track I used the cement method as above but rather than watering it in, I dribbled watered-down PVA over the whole thing and then watered it in. The whole lot dries rock-hard and does not seem to want to go anywhere. This has now become my default method of ballasting.
Although this track has not been in place as long as the rest of the railway, after a year and a hard winter, the ballast seems to be holding-up well. When I needed to lift some of this track to add some more sidings (see Progress Report 28), the ballast crumbled away readily, though was a little more persistent in clinging on to the sleepers than the cement-only method.
In four years' time, I will post another update on how well this approach is doing.

In a couple of places I have used wet-mixed concrete to in-fill where some of the blocks had sunk (see Progress Report 24). After a year, moss is beginning to attach itself to this ballast so I anticipate that eventually it will be indistinguishable from the surrounding track.

4. Using Tile Adhesive (method developed in 2014)

After six years, I've discovered that the above methods work for a while but eventually the bond between the materials breaks down. In some places the ballast was fixed down well whilst in others it had been washed away by the elements.

After relaying some ceramic floor tiles in the kitchen I had half a bag of tile cement remaining. I'd found the previous tile cement had been very effective, adhering to the old tiles extremely well. I figured it was the mix of fine sand and adhesive in the tile cement that provided a far more effective bond. If it works for tiles, why not for trackwork? I'd also noticed there was a whiff of PVA adhesive (or similar) emanating from the tile cement - so maybe it includes an additive to aid adhesion.

As with my previous approaches, I dry-mixed the tile cement in equal measures with potting grit.

This was then trowelled on to the track ......

..... and brushed into place with a 1" paintbrush.

The ballast was then watered with a watering-can to which a few drops of washing-up liquid had been added to help break down surface tension.

In addition to the newly relaid section of track, the adjacent track .........

...... and other sections where the ballast had been washed away...........

..... were also re-ballasted.

A few more shots, now the tile cement has had a chance to harden off (around two weeks after the ballast was applied) ......
The track beside the stream just outside Peckforton Station towards Bulkeley.
The approach to Peckforton from Beeston Castle
The cross-over between Bulkeley and Bickerton (looking towards Bulkeley)
The approach to Beeston Market
A close-up of the approach to Beeston Market

Time will tell whether this approach is more enduring than my previous attempts, but early signs are that it has a tougher finish which will hopefully protect it from the worst of the weather. The true test will probably be the extent to which it survives the frost in the winter months.

5. Wet mix and dry dressing (for station areas)

 The disadvantage of using the above methods in station areas is that there always seems to be some loose ballast which clogs up the pointwork. So I decided to use a wet-mix approach for station areas. However, because this did not look realistic when it dried, I then gave it a surface dressing of dry-mix fixed in place the SBR adhesive (a waterproof PVA cement additive).

The wet mix is used to infill between the tracks and provide a foundation for the top dressing.

The top dressing is a dry-mix of sifted soil, sand and crushed coal with bird grit (largely crushed sea shells) for the ballast in and around the tracks.

Once the dry-mix has been brushed into place, it is wetted and diluted SBR adhesive is dribbled on.

To tone down the contrast between the patches of different coloured dry-mix, it was then given a wash of highly diluted black cement dye. In muckier areas (eg where locos take on coal and water), the dye was diluted less to provide a stronger colouration,

This seems to provide a good representation of the sort of ballasting which would be found in station yards.

For more information see How I ballasted Beeston Market station yard.

Monday, April 24, 2006

How Did I Create the Stream?

You will see from Progress Report 7, that I use a sump for the end of the stream, rather than having a pond. Firstly, this part of the garden is shaded from the sun by the house and so a pond would receive insufficient sunlight to keep it healthy. And secondly, I did not fancy having a breeding ground for mosquitoes so close to the house.

Initially for the stream bed, I purchased a fibreglass, ready-formed stream, complete with pump. It was OK as a temporary measure but when I realised I would need to bed it into the landscape, I recognised its limitations. It was also very expensive and very short.

I then consulted various books and websites. I had originally intended to construct the bed of the stream entirely from concrete, but the books suggested this was not advisable as it cracks easily and ceases to be watertight. Ready-made plastic or resin sections of waterfall and stream seemed horrendously expensive - I worked out that even a short two metre stream would cost over 600 GBP (at 2006 prices). Using PVC or rubberised pond-liner seemed a better option, but would be difficult to blend into the landscape. Eventually, I opted for the method described below which aims to combine the reliability (and cost effectiveness) of a flexible liner with the creative potential of concrete. I have had one section in place for over a year now and it seems to be weathering nicely (UPDATE 2018 - the stream has been in place for twelve years now, and so far, so good).

The photos for this post show an extension I added to the original stream.

Stage 1
I planned where the stream will flow - making sure there was a gentle fall from top to bottom.

At the bottom of this photo, you can just see the top of the original stream. Also, you can see the junction-piece ready for extending the pipework.

Stage 2
I dug out the bed for the stream with a trowel, digging approximately 6 inches (15cm) deeper than the actual stream bed.

You can see I also chiselled into to blocks supporting the trackbed as I needed more width for the stream at this point.

Stage 3
The pipework is laid at the bottom of the trench. You may prefer to lay this to one side, in case leaks need to be repaired. I'm working on the principle that plastic pipes are almost indestructible and are more likely to be damaged by careless digging - hence laying them beneath the stream itself where there will be no digging.

The rocks are there to hold it down for the photo and were removed during the next stage.

Stage 4
Soft sand was laid over the pipe, to provide a bed for the stream. Where possible, it was moulded up the sides of the stream to ensure sharp stones don't puncture the liner.

Stage 5
I used some foam packaging which was wrapped around some electrical good to act as an additional cushion for the liner. This is not essential but I had it to hand and hate to see waste.......

Stage 6
I next laid the liner which as you can see is initially much wider than it needs to be - easier to cut of excess than try to extend something which is too narrow.

Stage 7
Now came the creative part. I carefully selected chunks of sandstone to line the sides of the stream. Because the stream was very narrow, I had to cleave some of the blocks. I suppose you could use pebbles, stones or gravel but as I had some chunks left over from building the rockery I used these - and it ties in with the locality.

Once I was happy with the position of the rocks (and this took more than a little trial and improvement), I trimmed back the excess plastic liner and the foam 'underlay'.
I then tested the stream by turning on the pump. When making the original stream, I had to test it by pouring water from a bucket as the pump I had then was too weedy. I quickly realised the flow was going to be too great for this new narrower stream and so decided to split the pipe at the point I had joined the new length to the old. I now have an outlet flowing into the original stream and an outlet flowing into the new stream.

Stage 8
The messy bit! I donned a pair of rubber gloves for this - I prefer the 'hands-on' method. A mix of 4 parts sand to 1 part cement was slapped, slopped and splodged behind, under and between the rocks to fill the cracks, provide a bed for the stream and, in places, create some new rocky outcrops. If I hadn't had any rocks, I could probably have moulded my own from concrete. I use an old 2" household paintbrush to smooth down the cement - it can also provided in interesting finish if it is stippled on the surface. When concrete is 'green' (ie when it has partially dried out but before it is hard) it can be sculpted and scribed.

I tend to then leave it for at least a week to harden. If the heatwave had continued, I would have covered the concrete with damp cloths to prevent it from drying out too quickly. Similarly, if it rained heavily I would have covered it with polythene.

I have no idea how durable this approach will be - a hard winter might result in some frost damage. However, I anticipate it will be relatively easy to repair any cracked concrete and the liner should prevent leakage (should!).

Views of the stream
The new extension
Where the new joins the old
The original stream

Update - July 2012

Over the past six years, the stream has become blended into the landscape. Moss and Mind Your Own Business (aka Baby's Tears) now line the banks and dip into the flow, giving the stream a more natural look.

From time to time I have to clear out the waterways of encroaching undergrowth - but this takes only a few minutes a couple of times a year. There is some seepage of water, which means that if I run it for around half a day I have to top it up with around a gallon of water. I can't figure out where the water is escaping as nowhere seems to become soggy, but this rate of seepage is something I can live with.

I am now considering adding a mill beside the main stream. It will be positioned on the further bank and will require its own dedicated siding - similar to that found on the Southwold Railway. As it will be on the furthermost bank of the stream I will need to build a bridge - and a simple timber trestle seems the most appropriate (again similar to that on the Southwold. I am also interested in building a fairly realistic mill race and sluiced leat to feed the mill. I'd like to be able to regulate the water flow in some way so the wheel doesn't end up spinning madly (and unrealistically) - but that may prove more difficult than I imagine!

Update - April 2014
 The planned mill siding has now been added (see How I constructed the mill siding)

UPDATE - July 2017
 After supposing that I had a leak in the stream (which there wasn't), I remodelled the section of the stream beside mill (see How I revamped part of the stream). I widened the stream here and incorporated a small section of mill-stream, ready for the installation of the mill building

UPDATE - November 2017
The mill has now been constructed and a sluice gate added to control the flow of water to the wheel. 

The stream has been in operation now for over twelve years and, as far as I can tell, has not sprung a leak, so it would seem that this system is fairly watertight (in both senses). A couple of times I suspected there was a leak but it turned out that the perforated lid of the dustbin through which the water returns to the sump get clogged with debris and so the water overflows rather than recycling. I now clean out the sump twice a year and find there is still very little water loss while running the pump.