Layout Design- Eliminating S Curves

Started by jward, December 21, 2012, 01:41:48 PM

Previous topic - Next topic

jward

One of the biggest mistakes people new to the hobby, and even some old heads, make when designing a layout is in the use of s curves. these can creep into a plan unless you are careful to recognize them and eliminate them. I have alluded to the s curve problem many times in previous posts, here I will attempt to illustrate the problem and the solution.....

the problem with s curves is that on cars and locomotives with body mounted couplers, the ends of the cars swing out. on an s curve, they swing in opposite directions. This is not a problem unless the limits of the coupler swing are exceeded, then you have a derailment as the couplers inability to swing further forces one or both cars off the rails. using a section of straight track between curves of opposite direction will bring the coupler swing problem down to manageable levels. The rule of thumb here is to use a straight at least as long as your longest car. On a practical matter, a 9" straight should be sufficient. and if you run cars no longer than 50 (scale) feet 6" will do.

The problem can sneak in when using track switches as crossovers between track. The numbered switches such as #4 or $6 have a section of straight track beyond the frogs, so their use in a crossover poses no problem. But the standard 18r switch included in some train sets, and in the track expander packs, are designed to be a drop in replacement for a standard 18r curve. Thus, the rails curve the whole way through the switch, and using two of them as a crossover gives you an 18r s curve. inserting the straight track between the switches alters track spacing, so a better solution must be found.


In the illustrations below are a typical crossover arrangement, and steps that can be taken to eliminate problems.



Here we have a crossover coming out of a curve. I have seen this in many plans including ones published in print. This configuration includes a wicked double s curve, referenced at X and Y.





By flipping the direction of the crossover, we are able to eliminate one of the s curves, but the other still remains, referenced at X.






Moving the crossover back into the curve eliminates both s curves, as well as keeping the crossover in the same direction as the original.


I hope these illustrations will help others avoid potential problems with their layouts.



Jeffery S Ward Sr
Pittsburgh, PA

NarrowMinded

Thanks Jward,

I didn't have a problem with S curves but I did run into a problem with space for a double loop I wanted to put under the tree.

Your configuration hold two switches in a much shorter space which I think will work out nicely

Thanks again

NM-Jeff

RAM


What I have been told about S curves is the the space between the two curves must be longer than the longest car that you have.

Doneldon

Quote from: RAM on December 21, 2012, 10:21:37 PM

What I have been told about S curves is the the space between the two curves must be longer than the longest car that you have.

RAM-

Not exactly. What you need is a straight section as long as your longest car.

                                                                                                            -- D

jward

Quote from: RAM on December 21, 2012, 10:21:37 PM

What I have been told about S curves is the the space between the two curves must be longer than the longest car that you have.

that is the rule of thumb. My experience is that it can be fudged a little bit. the key is how far to the side the couplers can swing. the actual distance of the straight section between the curves also depends on the radii of the curves. with say  the 35 1/2 inch ez track curves, you may not even need a straight between the curves, because the coupler swing isn't much. with 18r, there must be something between the curves.

another factor is the lengths of the coupled cars. cars of similar length will have similar swing. but couple an sd40-2 to a shorty ore car, and you will probably get a derailment on 18r because the sd40-2 swings wide, while the ore car doesn't swing much at all. Thus, the sd40-2 will probably pull the ore car off the track when it swings out on the curve.
Jeffery S Ward Sr
Pittsburgh, PA

rogertra

Quote from: jward on December 21, 2012, 01:41:48 PM






Moving the crossover back into the curve eliminates both s curves, as well as keeping the crossover in the same direction as the original.


While practical for model railroaders, it's unprotoypical because the diverging route has now become the main route through the curve.  The prototype rarely will do this on a main track.  In fact, even in this example, the inside track's switch would be a custom build so that the curve is aligned to be the main track while the straight would be slightly off set so as to become the diverging route.

It's all about good engineering practice regarding the lay of the points favouring the main route.  In this case, the main route is through the curve and thus the switch would need to be custom built for this to be prototypical.

However, you sadly see this sort of incorrectly applied design error in many model railroads.

What you should do, in both cases, is move the X-Y cross-over one car length down the straight tracks.  That's what the prototype would do.





rbryce1

#6
Our Christmas layout contains 2 "S" curves, and both work satisfactorily.  One is a direct "S" curve on 18" radius tracks, and we run a 10 car freight train only with a max 6" car length and a small Bachman 2-6-0 steam engine.  No difficulties, but I would have derailments with any longer cars.

The second is an "S" curve with 22" radius curves and a 5" straight section (an EZ Track 3" and 2" track section connected together) breaking up the "S" curve.  I can run my EM-1 and McKinley Explorer train on this arrangement fine, but again, it would not run without problems without the 5" straight section.

So even though I break a couple of the rules and I am having no problems, that is not to say I would not have problems if I went to any other arrangement like longer rolling stock or a longer engine with more wheels on the 18" curves.  I believe I would then have problems.  I lucked out, as there was not any more room for more track in these two locations on my arrangement.

jward

Quote from: rogertra on December 22, 2012, 02:41:50 AM


While practical for model railroaders, it's unprotoypical because the diverging route has now become the main route through the curve.  The prototype rarely will do this on a main track.  In fact, even in this example, the inside track's switch would be a custom build so that the curve is aligned to be the main track while the straight would be slightly off set so as to become the diverging route.

It's all about good engineering practice regarding the lay of the points favouring the main route.  In this case, the main route is through the curve and thus the switch would need to be custom built for this to be prototypical.

However, you sadly see this sort of incorrectly applied design error in many model railroads.

What you should do, in both cases, is move the X-Y cross-over one car length down the straight tracks.  That's what the prototype would do.






two points here.

first, on smaller layouts it may not be practical to move the crossover where you say it should be. we are all space challenged here compared with the real thing.

second and most importantly, the prototype could and often did run the main through the curved route. examples include Grafton, wv on the b&o with its legendary curved crossover, mckeesport, pa on the p&le where the branch took the straight route (in the middle of a bridge no less), the west end of the pittsburgh station on the prr, and roshester, pa also on the prr where the branch no only took the straight route but also used a crossing to come off the second of 4 main tracks.  looking at photos of british railroads it appears to have been common practice there as well.
Jeffery S Ward Sr
Pittsburgh, PA

M1FredQ

Your "S-Curve " problem came at an excellent time. I am taking time off this Christmas and will be re-doing our layout and we will re-work the "S_Curve"

as you suggest. The S- we currently have does cause problems and derailments

as the speeds approach 50 scale MPH.

Thanks for the tip!!!!!!!!1

rbryce1

#9
Quote
While practical for model railroaders, it's unprotoypical because the diverging route has now become the main route through the curve.  The prototype rarely will do this on a main track.  In fact, even in this example, the inside track's switch would be a custom build so that the curve is aligned to be the main track while the straight would be slightly off set so as to become the diverging route.

It's all about good engineering practice regarding the lay of the points favouring the main route.  In this case, the main route is through the curve and thus the switch would need to be custom built for this to be prototypical.

However, you sadly see this sort of incorrectly applied design error in many model railroads.

What you should do, in both cases, is move the X-Y cross-over one car length down the straight tracks.  That's what the prototype would do.


I really think Jeff is providing really good input for someone building a MODEL RAILROAD, not a "prototype".  We need to try and remember that model railroads do not have half the countryside to build on, and sometimes you just need to deviate from the prototype to have a feature work.

rogertra

Quote from: jward on December 22, 2012, 09:52:34 AM
Quote from: rogertra on December 22, 2012, 02:41:50 AM


While practical for model railroaders, it's unprotoypical because the diverging route has now become the main route through the curve.  The prototype rarely will do this on a main track.  In fact, even in this example, the inside track's switch would be a custom build so that the curve is aligned to be the main track while the straight would be slightly off set so as to become the diverging route.

It's all about good engineering practice regarding the lay of the points favouring the main route.  In this case, the main route is through the curve and thus the switch would need to be custom built for this to be prototypical.

However, you sadly see this sort of incorrectly applied design error in many model railroads.

What you should do, in both cases, is move the X-Y cross-over one car length down the straight tracks.  That's what the prototype would do.






two points here.

first, on smaller layouts it may not be practical to move the crossover where you say it should be. we are all space challenged here compared with the real thing.

second and most importantly, the prototype could and often did run the main through the curved route. examples include Grafton, wv on the b&o with its legendary curved crossover, mckeesport, pa on the p&le where the branch took the straight route (in the middle of a bridge no less), the west end of the pittsburgh station on the prr, and roshester, pa also on the prr where the branch no only took the straight route but also used a crossing to come off the second of 4 main tracks.  looking at photos of british railroads it appears to have been common practice there as well.

Your comment about the prototype running through the curve of the switch is well taken but what you fail to mention is that when they do that, the geometry of the switch is changed to favour the curved route.  They don't just insert an ordinary switch into the curve, they custom build a switch for the location.  Modellers just take a left or right hand switch and use the curve leg of the switch as the main route which, although it works on a model railroad, is NOT how it's done on the prototype.  Having built protoype switches, I know how it's done.  :-)

Of course, there are modellers, who model railroads and then there are model railroaders who model other people's model railroads.

While one way is more accurate, it doesn't matter which one you are providing you're having fun.

rogertra

#11
Quote from: NarrowMinded on December 23, 2012, 02:00:50 AM
I snicker every time I see someone start talking about things being prototypical on model railroads, I know they are out there but aside from amusement parks there aren't a lot of continuous loop railroads either, so If we flip a switch to make things run smoother like Jeff shows great! we spend less time rerailing cars.

Thanks again Jeff

Your assumption seems to be that everyone runs roundy-go-roundy trains, which is far from the mark.

There are many railroad modellers out there that DO run their model railroads as close to the prototype as possible.  We don't all run the latest techno-toasters pulling wooden freight cars or use 4-4-0s to haul autoracks.  Some of us, many of us, run our railroads based on strict eras and the rules of the day and run point to point.  If we do have through staging, the railroad will still run point to point with the staging yard acting as a terminal for all trains that enter it.  We tend to run our trains with a definite purpose, like the prototype does and use various car forwarding systems to handle freight cars and passenger cars in a realistic manner.

However, if you get a kick out of coupling up that 4-4-0 to a rake of autoracks and run that around the same circle of tracks for a given number of laps and then you run the "beer" train and later the "meat" train and then the "circus" train or the train made up of cars with all the state flags on them, then that's great as well, we all participate in the hobby to get what we want from it.  There's no right nor wrong way.

You'll also not see standard left and right hand switches in use with the main track going through the diverging route.  I take great pains to avoid that   :)

The same for locomotives.  No articulated locomotives, not even 2-8-4s as they were most uncommon in Canada.  The only exception are two RS-1s and the Bachmann  Russian 2-10-0s.  The RS-1s as it's assumed they are owned by an American subsidiary and the 2-10-0s, which have all been "un-Russianed BTW) are for use on a light rail branch line.

Personally, when it comes to modelling, I'm one of the former.  My GER, although freelanced, is set firmly in the summer of 1958 and operates (Or will) using TT&TO and CC&WB for freight forwarding.  I try hard to see that there are no anomalies in the locomotive, freight and passenger car fleets, automobiles, trucks and even the signage on buildings.  You'll see no Helvetica Medium font on my buildings.  :)

I want people to look at the GER and realise that it's set in the late 1950s in Southern Quebec.
 

jward

just wondering how this devolved into an argument over operations? this thread was intended to help others eliminate a problem with their layouts. if you want to debate an unrelated point, start your own thread, don't hijack mine.
Jeffery S Ward Sr
Pittsburgh, PA

NarrowMinded

Sorry Jeff,

I deleted my off topic comments.

NM-Jeff

jward

nm,
it wasn't you i was referring to.

thanks.
Jeffery S Ward Sr
Pittsburgh, PA