Castle Aching

[Edwardian]

I have been pondering the article on super-elevation and transition curves by Nigel Digby in March's Railway Modeller.

 

I thought it might be fun to put transition curves either side of my 3' radius exit curve.

 

Mr Digby gives the offsets for a 2' transition curve. He rounds everything to the nearest 0.25mm. This slightly troubles me, given my cack-handedness, which tends to translate to half an inch either way for anything I attempt to measure, but we'll see.

 

Mr Digby provides an equation for those of us wanting a different radius. 

 

Applying his methods, I arrived at the following table, showing the offset at 50mm intervals from 0 (beginning of the transition curve) to 100mm (the beginning of the 3' radius curve).

 

Length (mm):       0             50          100           150          200         250          300           350           400         450         500

 

Offset (mm):        0               0          0.25         1.25          3.00         5.75        10.00        15.00        23.50       33.50      60.00

 

God knows if it's right, but we'll see when I try to plot it at the weekend. 

Edited April 27, 2017 by Edwardian

[Stubby47]

James,

 

Plot those in Excel and you'll see an anomaly at 80mm.

 

 

Edit : 18.5 seems to give a better curve, instead of 23.5

 

Edited April 27, 2017 by Stubby47

[Edwardian]

James,

 

Plot those in Excel and you'll see an anomaly at 80mm.

 

transition curve.JPG

 

Looks a bit odd all round.  Still, I have checked my figures and I believe they accord with Mr Digby's method.

 

In the meantime, I have looked at calculating for 2.5cm intervals, thus:

 

300          325         350         375         400        425  

 

10.00     12.50     15.00      19.25     23.50     28.50

 

No idea what I'm doing, mind you.

Edited April 27, 2017 by Edwardian

[Stubby47]

I'm guessing you're plotting the expanding spiral from a continuous 3ft radius, to a point where the spiral can meet a straight line ?

 

I'm further guessing that if you were to continue the 3' curve, the offsets are radial distances between the fixed curve and the spiral ?

 

If so, the transition from 3' curve to straight in 100mm (10 cm) is not a very long distance at all (loving the mixed measurements ) - I'd have expected a yard or so (or a metre) to allow the curve to fully transit to the straight.

 

(and being cack-handed is a blessing, not a curse ).

 

Edited April 27, 2017 by Stubby47

[Edwardian]

I'm guessing you're plotting the expanding spiral from a continuous 3ft radius, to a point where the spiral can meet a straight line ?

 

I'm further guessing that if you were to continue the 3' curve, the offsets are radial distances between the fixed curve and the spiral ?

 

If so, the transition from 3' curve to straight in 100mm (10 cm) is not a very long distance at all (loving the mixed measurements ) - I'd have expected a yard or so (or a metre) to allow the curve to fully transit to the straight.

 

(and being cack-handed is a blessing, not a curse ).

 

ned.jpg

 

Now I did tell you that I didn't know what I was doing.

 

Having re-read the article and established the correct length of the transition curve, I have amended the table accordingly.  The offset figures are still correct as per Mr Digby's method, however.

 

The fact that I cannot visualise distances (at least those beyond a scale 6') when given in metric does not help.  If he'd done it in inches I would have realised the length was impossibly short.  Duh! 

Edited April 27, 2017 by Edwardian

[Stubby47]

Again, plotting both sets of new figures gives a sharp kink for the first set and an almost straight line for the second.

 

 

[Edwardian]

Again, plotting both sets of new figures gives a sharp kink for the first set and an almost straight line for the second.

 

I can see I am going to have to do this by eye; my usual method, as neither mathematics nor computing are my friends.

 

Supposing that my transition curve at least starts and ends approximateIy where Mr Digby's method suggests it should, I wonder what radius point I join to it?  That would be point number 1 on the plan.

 

Honestly, I can quite see why people buy set-track.  

Edited April 27, 2017 by Edwardian

[Stubby47]

'By eye' always works for me

 

A transitional curve for the straight track of the point will mean the curved track of the point exits from a fixed curve to a (virtual) straight. 

 

Whatever point you choose, this will be the same.  If you treat the platform line as the 'main' line, then continue the curve of the plain track in the same radius as the point, rather than a transition from straight. You can then transition this curve into your 3' curve.

 

Hope that make sense...

[Edwardian]

 

I'm further guessing that if you were to continue the 3' curve, the offsets are radial distances between the fixed curve and the spiral ?

 

 I'd have expected a yard or so (or a metre) to allow the curve to fully transit to the straight.

 

 

 

Thanks again, Stubby. To attempt an answer:

 

(1) I do not understand "the offsets are radial distances between the fixed curve and the spiral ?"  Because I am stupid.

 

(2) See amended tables - the length of transition curve is 500mm, still perhaps rather shot? Mr Digby suggests a 500mm length of transition curve for 2', 3' and 4' radius curves, which surprised me.

 

 

Whatever point you choose, this will be the same.  If you treat the platform line as the 'main' line, then continue the curve of the plain track in the same radius as the point, rather than a transition from straight. You can then transition this curve into your 3' curve.

 

Hope that make sense...

 

No, not to me.  Because I am stupid.

 

I thought to myself, "here is something in the scary world of track that I can understand and apply". Now I discover this is not so and it is all impossibly complicated.

 

Do you mean I could, moving right to left, have: (1) 3' radius curve, (2) transition curve, (3) curved 3' radius point into the platform road?

Edited April 27, 2017 by Edwardian

[Stubby47]

Not you - I'm not explaining it very well...

 

1 )

Imagine a circle.  From the centre, there are lines leading out to, and beyond, the edge of the circle (radials).

Now imagine a spiral leaving the circle edge and getting progressively wider and crossing those radial lines. The distance between the circle and the spiral is the offset (the red bits).

 

 

2)

500mm is probably adequate, but the longer the curve the more fluid the change in curve.

 

3) If the point is already a 3ft curve, then continue the plain track without a transition, which should give you a smooth flow.  (which negates the whole point of your theorising  )

Edited April 27, 2017 by Stubby47

[Nearholmer]

I was going to say 'forget all the formulae; do it by eye', the I thought that sounded too anti-scientific, then i saw what Stubby had written.

 

Do it by eye.

 

If you select a position and orientation for your 'king' point, which seems to be Point 1, then work outward from it in all three directions, you will be able to eyeball it as you go ....... if it looks right, it is right ..... you aren't going to be running 1000 tonne trains at 200kph through it.

 

K

[Edwardian]

 

A transitional curve for the straight track of the point will mean the curved track of the point exits from a fixed curve to a (virtual) straight. 

 

Whatever point you choose, this will be the same.  If you treat the platform line as the 'main' line, then continue the curve of the plain track in the same radius as the point, rather than a transition from straight. You can then transition this curve into your 3' curve.

 

 

 

I think I get this now.

 

On the plan, the road leading to the siding is straight, so I have a straight left-hand point.  That is why you say whatever point I choose it will be the same?

 

The curved road is the left road to the platform, and this is whatever radius selected, with no transition.  It was suggested that I work to a minimum 3' radius, so this could be a left-hand 36" radius point.  But is that necessarily the curve I want? Would it lead to the straight platform road at the correct angle?

I wonder what the effect of a curved point at this, ahem, point would be?

I was talked into building the points at one, ahem, point, but I don't think it will ever get done that way.  As it is, points 4 and 5 (shown read on the plan, are FB, so still might have to be built (unless I cheat with Peco Code 75 FB points and alter the sleepering to something closer to the plain track spacing, but that's by way on an aside).

The 3 blue points are bullhead, and I'd really, really like to cheat and buy ready-made SMP, same as the plain track.

 

 

Not you - I'm not explaining it very well...

 

1 )

Imagine a circle.  From the centre, there are lines leading out to, and beyond, the edge of the circle (radials).

Now imagine a spiral leaving the circle edge and getting progressively wider and crossing those radial lines. The distance between the circle and the spiral is the offset (the red bits).

 

transition curve sketch.JPG

 

2)

500mm is probably adequate, but the longer the curve the more fluid the change in curve.

 

3) If the point is already a 3ft curve, then continue the plain track without a transition, which should give you a smooth flow.  (which negates the whole point of your theorising  )

 

Yes, me.  I am very stupid when it comes to this sort of thing.  Always have been.

 

The diagram does, however, speak a thousand words.  Mr Digby plots his offsets from the outer side of the curve.  You diagram makes it look as if I could put the track down on a plotted 3' curve (I believe I have a Tracksetta!), but then progressively ease it.

 

"If the point is already a 3ft curve, then continue the plain track without a transition, which should give you a smooth flow."  - does this mean that I lay a 3'radius curve and then add a 3' radius (left-hand) point to it?

 

I am not sure there is anything suitable for that 'off the shelf'

 

Which brings me back to the idea of a straight left-hand point, with the radius left into the platform road I would have an off the shelf choice of 36' or 48'.  

[Caley Jim]

...... That would be point number 1 on the plan.

 

Honestly, I can quite see why people buy set-track.  

This plan looks to me very much like it has been drawn with Templot, in which case you can use that to put in a transition curve without having to get get involved (and confused) by any mathematical machinations.  You can even have a turnout on a transition curve.   Or,as you suggest, do it by eye!To parody a quote about statistics 'Some people use mathematics as a drunk man uses a lamp post - for support rather than illumination'.

 

Jim

[Edwardian]

I was going to say 'forget all the formulae; do it by eye', the I thought that sounded too anti-scientific, then i saw what Stubby had written.

 

Do it by eye.

 

If you select a position and orientation for your 'king' point, which seems to be Point 1, then work outward from it in all three directions, you will be able to eyeball it as you go ....... if it looks right, it is right ..... you aren't going to be running 1000 tonne trains at 200kph through it.

 

K

 

Thanks.  That makes sense in terms of the transition curve, but how do I determine the radius of the Point No.1 left hand fork to the platform road?

 

Wargaming might be the better option, after all! 

[Edwardian]

This plan looks to me very much like it has been drawn with Templot, in which case you can use that to put in a transition curve without having to get get involved (and confused) by any mathematical machinations.  You can even have a turnout on a transition curve.   Or,as you suggest, do it by eye!To parody a quote about statistics 'Some people use mathematics as a drunk man uses a lamp post - for support rather than illumination'.

 

Jim

 

I believe it was, but not by me, and I have no idea how to translate it to full size, or what points is shows.

[Stubby47]

I think I get this now.

 

On the plan, the road leading to the siding is straight, so I have a straight left-hand point.  That is why you say whatever point I choose it will be the same? - Not quite - whichever RTR point you use at #1 will have a fixed curve from toe to curved track exit 

 

The curved road is the left road to the platform, and this is whatever radius selected, with no transition.  It was suggested that I work to a minimum 3' radius, so this could be a left-hand 36" radius point.  But is that necessarily the curve I want? Would it lead to the straight platform road at the correct angle? - you might want a transition curve into the platform, to ease the overhang of the rolling stock - help prevent them hitting the platform

I wonder what the effect of a curved point at this, ahem, point would be?

I was talked into building the points at one, ahem, point, but I don't think it will ever get done that way.  As it is, points 4 and 5 (shown read on the plan, are FB, so still might have to be built (unless I cheat with Peco Code 75 FB points and alter the sleepering to something closer to the plain track spacing, but that's by way on an aside).

The 3 blue points are bullhead, and I'd really, really like to cheat and buy ready-made SMP, same as the plain track.

 

 

 

Yes, me.  I am very stupid when it comes to this sort of thing.  Always have been.

 

The diagram does, however, speak a thousand words.  Mr Digby plots his offsets from the outer side of the curve.  You diagram makes it look as if I could put the track down on a plotted 3' curve (I believe I have a Tracksetta!), but then progressively ease it.

 

"If the point is already a 3ft curve, then continue the plain track without a transition, which should give you a smooth flow."  - does this mean that I lay a 3'radius curve and then add a 3' radius (left-hand) point to it? - Yes

 

I am not sure there is anything suitable for that 'off the shelf' - Ah, nor do I , I just use Peco

 

Which brings me back to the idea of a straight left-hand point, with the radius left into the platform road I would have an off the shelf choice of 36' or 48'.  

 

 

I feel I'm making things worse...

[Nearholmer]

You are making the points? Then you might be better starting by deciding what the angle of divergence of the point should be to suit your railway, and I suggest that 1:6 or 1:8 might be about right.

 

Then you can fair the curve of the switchblade, and eyeball how you want that curve to run out on the right.

 

One thing to avoid is a sudden change of radius within a few inches to the right of the switch-tips, because that can cause sudden changes in what is happening at the wheel/flange, leading to the flange climbing over the rail.

 

A very pre-templot approach!

 

K

[Nearholmer]

Er ....... oh ....... they are Peco points?

 

Plonk the "king" point down, then go by eye.

 

K

[Donw]

I looked at that article and thought to myself that will cause some headscratching.  I used a simpler method with is derived from that use by John Armstrong a US layout designer. Basically the transition needs to be two coach lengths to be effective so I would draw the plain radius circle ( e.g 3 ft)  mark a point one coach length about 9inch back along the curve. I mark the end point 9inch down the straight and about 1.5 inch out from it using a flexible piece of thin wood I align it to the curve and to the straight using the marks this will naturally take up a nice curve.

However you will note that we have lost 9inch of the straight length so a pain on a cramped layout.  

Simpler method from the strart point of your curve use a 5ft radius curve for 10inch a 4ft radius curve  for the next    8inch then the 3ft curve you will not notice is is not a true transistion but it will look good and do the same job.

 

If you are going into a turnout at the end of the curve a 48inch turnout from a 3ft curve will work if the main is the curved route and the straight bit is a siding. If the diverging route needs to be on the inside use a 54/36 or a 48/32 curved turnout .

 

Once you have got the idea you can largely do it by eye but the thin bit of wood is most helpful as it naturally takes up natural curves.

 

The reason why transition curves make sense full size is it controls the forces generated by the change in direction

 

The reason why transition curves make sense on a model is it reduces the maximum coupling offsets.

Consider two long coaches on a curve the ends betwen them are both off centre but they are off centre on the same side thus the couplings are in line but at a slight angle

Consider two coaches approaching a non transistion curve as one enters the curve the end will swing out off centre but the coach it is attached to is still on the straight dead centre thus there is an offset between them

 

Consider the two coaches approaching a transistion curve the end of the first coach will start to swing out off centre but less so on a genrous curve as the curve increases the second coach will be on the generous bit of the curve and its end will be starting to swing out as well so the amount of off set is reduced.

A couple of minutes with two coaches and a bit of flexitrack should show you what I mean.

 

Lastly full size turnout have a transistion curve within many of them on a typical size the first part of the blades etc. will be at a larger radius that the closure rail the bit through the crossing can be straight or at the closure rail radius. On modern turnouts (post grouping is modern to me)  the blades themselves were curved or semi curved with gave a transition within the switch part. These transitions were needed to cope with the forces generated without them trains would be severely restricted in speed through turnouts.   

 

Don

 

 

ps. Model turnouts vary those building there own use prototype forms but often sharper than would be used full size. Manufacturers usually either use a simple curved turnout or something in between. These shortened turnouts do not cause problems in running but don't look as good as ones built to prototype designs.

 

pps I have given figures used for 4mm 

Edited April 27, 2017 by Donw

[Edwardian]

Thank you everyone. 

 

The plan was one I sketched out.  The Templot image was made on the basis of my sketch by the Man Who Deleted Himself.  I just received the image; no idea of what to do with it.

 

Once upon I time I was persuaded to build the points, but, no, I don't want to build any points.  Progress, if it can be called that, is glacial enough as it is.  15 months in and no track laid. You could measure my progress in geological time.

 

But does this plan require a hand-built 'King' point?

 

Point No. 1

 

To me, Point No.1 looks like a straight left-handed point.  I thought that, rather than go from the straight toe of the point immediately into the 3' radius curve, I could insert a transition curve.  I assume that this much makes sense as a plan.

 

I note Don W's comment "If you are going into a turnout at the end of the curve a 48inch turnout from a 3ft curve will work if the main is the curved route and the straight bit is a siding. If the diverging route needs to be on the inside use a 54/36 or a 48/32 curved turnout".

 

SMP options include:

 

36" Radius Point Left or Right Hand - £ 27.50

 

48" Radius Point Left or Right Hand - £ 27.50

 

54"/36" Radius Curved Point Left or Right Hand - £ 47.50

 

48"/36" Radius Curved Point Left or Right Hand - £ 47.50

The cost of the curved points are clearly far greater, but, if I understand the plan and Don's comment, the siding route is straight at the platform road follows a curve path to the left.  That suggests, to me, that a 48" Radius Point Left is the appropriate choice.

 

Point No.2

 

Now this does look like a curved point.  Expensive.

 

I should explain that the distance between the platform road and the one beside it, the loop, can be quite narrow, i.e. 'the six foot' could actually be 6'.  The gap between the loop and the outside line is wider.  The latter is the engines shed line and allowance must be made for the thickness of the outer wall of the train-shed that will run between it and the loop.

 

Anyway, the available options off SMP's peg are:

 

54"/36" Radius Curved Point Left or Right Hand - £ 47.50

72/48" Radius Curved Point Left or Right Hand - £ 47.50

36"/28" Radius Curved Point Left or Right Hand - £ 47.50

48"/36" Radius Curved Point Left or Right Hand - £ 47.50

 

A left-hand of the last of these would seem to continue the 48' curve of the preceding straight point (No.1), allowing an acceptable 3' curve into the loop. Or am I missing something again?!? 

 

Point No.3

 

This looks like a 'Y' point on the plan.  I note that SMP produce: 36" or 72" Radius Wye Point.  Is one of these likely to be sensible?

 

The sensible course of action would be to send for their 'Layout Planner Full Size Rail Prints Of The Range' at £ 5.50.  I am resistant to that, because I included this in my order for plain track, FB rail and sleepers, but it was not included in the delivery!

 

So, a straight, a Y and a curved gives me a cost of £102.50 for the bullhead sections. Not insupportable. I'd like to avoid using 2 curved points, however! 

 

For the 2 FB (red) points in the sidings, I was hoping that "Peco SL-E189 00/H0 Finescale Large Radius Left Hand Points" would do the job, as these seem to be available for as little as £12 a pop, so less than half the price of the SMP straight and Y points. That economy eases the pain somewhat.

 

Any comments, suggestions and corrections are very welcome.

 

Thanks

Edited April 27, 2017 by Edwardian

[Edwardian]

I feel I'm making things worse...

 

Actually, you have been a great help, thank you, as has Kevin, Don W and Caley Jim.

 

I's a bit thick, 'tis all.

Edited April 27, 2017 by Edwardian

[Compound2632]

I think one point Stubby made was that taking the platform line to be the main line, all other lines diverge from it, so technically point 1 is a right-hand point. Not sure if that helps?

 

But this is a light railway so main-line thinking may not apply... unless we're going to see a Midland D307 Stores Sleepers wagon sneaking in.

[Edwardian]

I think one point Stubby made was that taking the platform line to be the main line, all other lines diverge from it, so technically point 1 is a right-hand point. Not sure if that helps?

 

But this is a light railway so main-line thinking may not apply... unless we're going to see a Midland D307 Stores Sleepers wagon sneaking in.

 

I say "left-handed" for Point No.1, as the right-hand path appears to be straight, whilst the left-hand path diverges in a curve. Whether that is acceptable prototypically speaking is quite another matter!

 

And the West Norfolk is not a Light Railway!

 

Pay attention!

 

Or I'll find a Midland Stores Sleepers wagon to run!

[Compound2632]

And the West Norfolk is not a Light Railway!

 

Pay attention!

 

Or I'll find a Midland Stores Sleepers wagon to run!

 

The West Norfolk is not a light railway

The West Norfolk is not a light railway

+ 98 further repetitions.

 

Indeed, light railways only came in after the act of 1898 which was too late to save over-capitalisation of lines like yours.

[Stephen 28]

Bonjour, c'est Routier du Nord ici! Mon ami Kevin 'e say ze ideal Norfolk line ees at Mont St Michel, naughty boy! But 'e has raison, oui! Go to where ze createurs des chateaux com from, n'est ce pas? Forget you obsessions wiz Wells et 'Unstanton - pouf! I geef you Andelys, Entretat, Carteret - formidable!

]http://forum.e-train.fr/viewtopic.php?f=5&t=76745IMG_0976.JPG

Superb, even more distracting and off piste than the usual Castle Aching topics, especially when only possessing very rusty GCSE French so roughly one word in three making sense. Some lovely termini and track layouts obviously designed with model railways in mind - more zig zag reverses than your average line.