Three Way Turnout

[Armchair Modeller]

You're not going to forge a good relationship like that.

regards

Alan

Alan, are you suggesting that in the future, if I meet Keith somewhere like a bar, I might be a stretcher case? Hopefully I can give him the slip.

Flymo - Yes, the phosphor bronze may be worth a try. Do editors know about the 3 foot rule though? Looking through magazines they even make 2mm scale look like 7mm or larger!

One other idea I forgot was in another MRJ (7mm scale again). Soldering a looped wire under the rail to join the point blade to the adjacent rail, with an additional loop of wire soldered sideways into holes drilled through both rails.

[Armchair Modeller]

Well, we have made limited progress, but this is proving to be a mildly frustrating project so far!

GE-3-way-Turnouta.gif

The turnout has 2 kinds of point blades - 2 hinged (not in orange) and 2 which flex (in orange)

Dealing with the hinged ones first, I tried several of the ideas listed. Most failed pretty quickly, The only one that seem reliable is the pivot method. To be honest, I think it could be made fairly invisible in the ballast, so that doesn't worry me too much.

The unhinged ones are the big problem as the rail only has about 40mm of flexible length. It takes a lot of force to move them from side to side. After a lot of trial and error, I came up with the idea of using 0.5mm brass tube for the stretcher bar. I drilled a hole in the rail (lots of broken drills, I fear!) and inserted an L shaped piece of 0.3mm wire through the hole and into the stretcher bar. I then soldered it up. The wire would be a snug fit inside the recess in the stock rail. This joint proved to be pretty strong.

The next step was to try and work out a way of insulating the centre of the stretcher bar. I tried several things that fitted inside the tube, like plastic rod and thread soaked in superglue, but these all failed pretty quickly due to the stress in the stretcher bar finding the weak point.

Finally, I tried soldering the two parts of the stretcher bar to a thin bit of copperclad. I haven't done too much testing yet, but this seems to be the most promising from a strength point of view. Here are a couple of photos of the components all soldered together.

aDSCF4393.gif

aDSCF4392.gif

The copperclad could be filed down a lot further to reduce its visibility.

The stretcher bar should be around 0.3mm to scale, but 0.3mm tube is very fragile and anything put in the tube would have little structural strength.

Not sure where to go from here just yet, apart from testing the new stretcher bar to destruction.

[Knuckles]

0.3mm TUBE? Blimey. Finescale then!

I can't currently offer any help but I can watch your thread and cheer you on with some wire wool pom poms.

I need to make a 3 way equal point at some point so I'm trying to pick up a trick or two.

[Jol Wilkinson]

A M,

please change your pseudonym. You are doing a great disservice to the real armchair modellers who unselfishly limit themselves to reading the modelling media, attending exhibitions and posting their views on social media.

Making stuff shows complete disloyalty to the brotherhood (I was never allowed to join).

Jol

[Knuckles]

A M,

please change your pseudonym. You are doing a great disservice to the real armchair modellers who unselfishly limit themselves to reading the modelling media, attending exhibitions and posting their views on social media.

Making stuff shows complete disloyalty to the brotherhood (I was never allowed to join).

Jol

LOL, nice one.

I guess we should hack his account and change his name to Active Modeller.

[Armchair Modeller]

I model armchairs in my spare time so leave my name alone!!!!!!!!!!!!!!!!

[Winander]

Dealing with the hinged ones first, I tried several of the ideas listed. Most failed pretty quickly, The only one that seem reliable is the pivot method. To be honest, I think it could be made fairly invisible in the ballast, so that doesn't worry me too much.

As I have to do this, may I ask how you restricted vertical movement and if the tube is buried in the ballast, how do you keep the hinged part aligned with the fixed? Although I doubt the latter would be difficult, it's just something else to have to do on installation. My thoughts were to solder etch to the fixed rail that protrudes under the hinged rail and use a vero pin up through it into the rail that would take care of vertical movement and alignment and make the unit self contained. Subject to not soldering the entire thing solid, that is!

regards

[Armchair Modeller]

Many thanks for your interest Winander.

Before I answer your points (sorry for the pun ), here's another idea for hinged point blades - towards the bottom of post no 28

http://www.rmweb.co.uk/community/index. ... try1163187

Might try something like that too.

Regarding my own humble efforts, I didn't feel I needed to try the pivot method at this stage because it seemed very likely that it could be made to work. I thought I would mess around trying the best of the others first.

My reservations about going straight for the pivot method were based on the premise that it would be too visible, but I have since mellowed. Using the same 0.5mm tube and 0.3mm wire that I used for the stretcher bar experiment, the pivot would be pretty well invisible, especially with a bit of good, old fashioned ballasting. The pivot centre would have to be right at the blunt end of the point blade, near enough, to look authentic. I wouldn't recommend drilling a hole to insert a pin into the bottom of the rail - It would be difficult to get a small hole straight and at the right angle. You would also be likely to break a lot of drills as the C&L n/s rail I am using is very hard.

I was thinking of bending an L shaped piece of wire, soldering one side of the L to the underside of the point blade. The other side of the L would fit into some 0.5mm tube fixed vertically under the rail. To keep the pointy end of the point blade down, I would try bending the L in the wire at slightly less that 90 degrees. The bottom end of the wire would have to be retained in the tube by a soldered washer or suchlike on the underside of the track.

I haven't thought too deeply about how I would hold the tube in place yet. My turnout would be very much a one-off, so it matters little whether the tubing is inserted in the baseboard or attached in some way to the turnout itself. For mass production, I would be thinking of creating jigs and making a metal bracket to solder under the sleepers to hold the tube in position (I am using copperclad sleepers). Pretty well all the soldering could then be done before final assembly, making it less likely that you would solder the whole thing solid.

Most of my problems are unique to the 3-way turnout I am trying to build though. A plain, ordinary turnout would probably be very easy, by comparison.

Having said all that, I am now wondering how much of the movement of a pivoted pint blade was pivoting and how much was flexing. My gut feeling is that the point blade would have to be held pretty firmly by the fishplates to make sure it stayed properly upright and aligned with the closure rail. The few photos I have found suggest maybe a bit of both pivoting and flexing took place, but I can't be sure.

[Armchair Modeller]

I forgot to mention how the stretcher bar testing is going. Amazingly well, so far! Despite tugging and pushing the honorary point blade back and forth many times, the joint is still intact.

Just to reinforce my next experiment a little further, I will add a metal washer to the end joint on the inside of the point blade - just to give it that little bit more strength. 2mm finescale valve gear washers appear to fit nicely. My next attempt will be on a properly filed down point blade, just to make sure that a drastically thinned down rail can still cope with the stresses imposed by the stretcher bar/tie bar arrangement on the design I am following.

The one big concern is that if a stretcher bar did fail on a completed model of the while turnout, it could be a nightmare to try and put it right.

[Winander]

Before I answer your points (sorry for the pun ), here's another idea for hinged point blades - towards the bottom of post no 28

http://www.rmweb.co.uk/community/index. ... try1163187

Might try something like that too.

An interesting idea, but I don't fancy attempting it in 4mm to the foot. I've heard that 7mm modellers even paint keys in the chairs to look like wood .

I was thinking to cut the foot and web of the rail away (leaving the head) and solder something in line with the web to form the hinge pin thus avoiding the need to drill the rail. Like you, I anticipated that to be a difficult exercise and recognised that the fulcrum had to be right at the end of the rail. Leaving the head intact means the top of the rail does not have a joint. My current idea is to use a steel panel pin as I use steel rail, although since it will be painted, a piece of brass will do.

I am enjoying this discussion as it is forcing me to finalise a design that was only vague, so thank you.

regards
Richard

[Armchair Modeller]

Thank you too for prompting a few ideas. I am really just exploring the possibilities and don't have a fixed idea of how things will turn out. It will just evolve over time. Do make sure you show us how you get on!

[Armchair Modeller]

Looking again at the GE 3-throw switch, I did a quick mock up of what the complete turnout might have looked like in Templot. Much to my surprise, my best guess minimum radius, based on the short section shown in the GE diagram, came out at around 54.8 inches for the RH turnout and 80 inches for the LH turnout in 4mm scale. The angle of the crossings came out at 1:7.75 and 1:6. I used 12 ft straight cut switch blades, as per the diagram. Any attempt to reduce the length of the turnouts quickly led the Templot turnout well astray from the GE diagram. OK the geometry of the GE turnouts might not quite have followed Templot rules, but the 1:7.75 with an 80 inch minimum radius did surprise me as a bit large for a supposed space saver.

I now have the dilemma of should I follow the GE diagram rigorously or not. If I make the radii much sharper, then the switch blades will need to go down to 9ft. Because the roads diverge so much quicker, there is likely to be room for pivoting blades on both roads, which would almost certainly be an advantage.

I am pretty certain that shorter 3-throw switches did exist. I have a 1:500 OS map showing a GN one in Nottingham London Road that must have been shorter. A guestimate of this suggests 9ft blades and a minimum radius on one road of 47.3 inches and a radius on the other of 57.5 inches, using Templot as my guide. As I am basing this on an OS map, it can only be a very rough guess, of course.

Time to pause for thought!

[grovenor-2685]

The GE switch drawing looks symnetrical to me. Why have you assumed different radii and crossing angles for the LH and RH sides? The more usual case is for them to be equal as small differences give problems with checking the crossings.
Regards

[Armchair Modeller]

Hi Keith. Many thanks again for your contribution.

I just imported the GE plan into Templot and tried to draw the turnout over the top of it. Symmetry didn't work, as the rails beyond the switch didn't follow the plan. The curves change quite dramatically as the crossing angle changes. Only particular combinations of switch length and crossing angle gave a result for me that fitted the diagram. My total lack of expertise may not have helped.

It could of course be that the plan is distorted, but its all I have to go on. I don't think the drawing is particularly distorted though as the straight rails in the middle do appear straight. Maybe Templot works in a slightly different way to GE track designers?

In my ignorance I thought that the turnout could not be totally symmetrical though, as the switches start one sleeper apart,

If it has to be symmetrical, I guess I would have to go for something between the two radii I calculated.

[grovenor-2685]

Hi Keith. Many thanks again for your contribution.

I just imported the GE plan into Templot and tried to draw the turnout over the top of it. Symmetry didn't work, as the rails beyond the switch didn't follow the plan.

What plan? have you got more than just the switches as shown above?

The curves change quite dramatically as the crossing angle changes. Only particular combinations of switch length and crossing angle gave a result for me that fitted the diagram. My total lack of expertise may not have helped.

It could of course be that the plan is distorted, but its all I have to go on. I don't think the drawing is particularly distorted though as the straight rails in the middle do appear straight. Maybe Templot works in a slightly different way to GE track designers?

Almost certainly! I don't think Templot has any specific provision of 3-way switches so you just need to put both sets of switches starting on the same timber, its the switch blades for the curved routes that matter, the straight switch blades are just shortened back one timber.

In my ignorance I thought that the turnout could not be totally symmetrical though, as the switches start one sleeper apart,

Note that the switches are symnetrical as each set is composed of one long switch and one short switch.

If it has to be symmetrical, I guess I would have to go for something between the two radii I calculated.

I would let Templot do all the calculating, just overlay a LH switch over a RH switch with the toe and crossing noses together, and try 1:6, 1:6.5, 1:7, see which looks best.
The original 3-way templates from P4/Studiolith were done as B6 but I don't know where they got details from, look OK though.

Its also possible to have 3-way switches where the straight route is at one side and these end up with a quite different geometry.
Regards

[Armchair Modeller]

Thank you for your patience with me Keith! The only plan I have is the one we have all seen. I have nothing beyond that.

Apologies about suggesting the turnout is asymmetric. It must have been some kind of optical illusion - or I need a visit to Specsavers. I played around in Photoshop and overlaid a mirror image of the drawing onto the original. It is undoubtedly symmetric.

I have played around in Templot again, using 12ft switch blades. I can more or less get Templot to match the diagram with anything from V-6 to V-7.75 (there is not a lot of difference at the switch end of the closure rails) before the program automatically makes the switch blades shorter or longer. However, I can only get the geometry of the curved rails to match the plan if I roam the turnout one sleeper to the right - i.e. switch blades in line with the short switch blades on the plan.

I guess that something non-standard is going on at the sharp end of the longer switch blades? Certainly there is little or no discernible alteration of the angle of the stock rails on the plan until they get to the knuckles. That would make the track out of gauge, which can't be right. I guess they did the classic modellers fudge of widening the gauge slightly before the knuckle to accommodate the long switch blade width?

Again, I really don't know much about all this. Your expert eye and comments are certainly making me think though!

[Armchair Modeller]

.......and just for a bit of fun, a first stab at what the Nottingham London Road 3-way with slip might have looked like. As with my other work, I claim absolutely no expertise in the matter whatever, but it was fun trying to draw it!

nlr1.gif

It is not the finished thing - I should really do some refinements to the Templot drawing - particularly to the checkrails, but at least its a start.

[Martin Wynne]

before the program automatically makes the switch blades shorter or longer.

To prevent that you can either:

1. use F9 to adjust the crossing angle instead of F5, or

2. while using F5, right-click and select lock switch.

Martin.

[Armchair Modeller]

Thanks for more helpful Templot advice, Martin.

I tend to use the mouse for everything I possibly can, rather than F keys. Your advice will no doubt be of great use to others though.

[Armchair Modeller]

I have just about convinced myself now that I can build a workable model of the 3-way turnout.

I did an experiment with a crudely filed switch blade. The joint with the stretcher bar held OK when I did some testing. I would be too embarrassed with the workmanship to do a photo - but then it didn't have to look nice. One made for the final track will of course be more carefully prepared - it will have to be very precise, or it won't fit.

I have to order some more chairs and sleeper strip before I can go ahead with the real thing. The idea is to do a 3-way combined with a single slip, as in the diagram a few posts up. I repeat it here...

nlr1a1.gif

This was inspired by a nineteenth century formation at Nottingham London Road (Great Northern). I have no photo or scale drawing - just OS maps, including a 1:500 map which shows a reasonable level of detail. I suspect the formation didn't last too long, as a 1900 track plan I found recently in 'The Engineer' suggests a simplified setup with just a single slip.

With no drawing or photo, I shall probably have great fun working out where to place all the sleepers. I suspect the real thing may have had to be a bit of a bodge. Some of the timbers will be hidden in deep ballast anyway, so only the chairs will give anything away. I managed to squeeze 12ft switch blades in the 3-way, but the RH end of the single slip is a 9 footer - nothing longer would fit.

The main thing is just do do it and (hopefully) get it working satisfactorily. That would give me a real sense of achievement. If all is well, it will go on the new baseboard for my layout. If not, then the experience of building it alone should make the effort worthwhile.

[John Palmer]

With complex formations I often have difficulty seeing exactly what problems may be encountered when, in Templot, you have multiple templates overlaid. I could be wrong but I think you might encounter such a problem with this formation in the area indicated by a red ellipse on the attached copy of your plan. Are there two crossings here in very close proximity? If so, I think you may have difficulty getting adequate support for the wheel passing over the left crossing of the the two - its wing rail is certainly going to be very short, and fabrication of the vee could also prove tricky. Before committing yourself to this geometry, would it help to develop the Templot plan so you can see the formation as it will finally appear?

Problematic 3-way.jpg

[Armchair Modeller]

Hi John and thanks for your comment. An expert eye is always useful.

In this instance, I think I am covered though - see similar examples in the photos below

nlr1a1a.gif

nlr1a1b.gif

[Martin Wynne]

In this instance, I think I am covered though - see similar examples in the photos below

Hi,

Your track plan is missing this check rail:

missing_check.jpg

regards,

Martin.

[Armchair Modeller]

Yes, sorry Martin. I am being very lazy and doing a few manual adjustments to the plan to cover things like that. There are in fact one or two other things like extra wing rails to add as well. I need to do one or two slight adjustments to the stock rail position and the long point blades, based on the GER drawing to keep everything in gauge too.

I am sure an expert could get a perfect drawing in Templot without the need for manual adjustments, but I am not that good and it would take me ages. My 'insurance policy' is to use Masokits chairs and soldered construction, so I can adapt things relatively easily as I go along. If I find some horribly catastrophic problem, I am happy to learn from it (the hard way!) and try again.

[Armchair Modeller]

A rainy day, so I took another look at my drawing.

I found a LSWR 1:7 single slip drawing in the S4 archives (23.6.3.1). I decided to try another version of my 3-way/single slip in Templot using the LSWR drawing as a template. There is a drawing of LSWR 3-throw switch in a topic about LSWR track on RMweb. I overlaid the switch drawing onto the single slip in Photoshop, along with the mirror image of the single slip to show me where the third side of the 3-way would go. I then imported this into Templot and drew a Templot track plan over the top of it.

I could not get the 3-way totally symmetrical as the right hand curve interfered with the slip switch. Changing to a 1:6 on the right hand curve did make everything fit. I am happier with this than my original drawing, particularly as the LSWR single slip drawing shows a lot of detail I was missing. The angles are very slightly different - but given that my original source of information was an OS map, accuracy would always be slightly questionable.

LSWR-Single-slip-3way-web.gif

Again, I haven't drawn all the bells and whistles on this drawing. My pencil has since done the work on a paper copy, adding missing gaps in rails, adding checkrails and wing rails etc. As far as I can see, it should be possible to make the design run satisfactorily, especially with the professional bodges shown in photos in a recent posting above. I have also marked where the insulation gaps have to go.

The LSWR 3-throw switch does not have joggles, unlike the GER one I based my first attempt on.

As before, I will have to wait for components before I can make a start, so it may be a few weeks before I can make any significant progress.