See "reverse running"

Ooops in sympathy.

Bullet proof track is easiest done if planned and drawn, or at least full size templated and checked for fit, and only then laid.

parabolic curves

We had a problem like that on one curve going into a helix on the Webbwood branch.  We built the curve with spline and didn't pay enough attention to making sure that the radius was consistent.  We were shooting for 30 inch radius and ended up with something closer to 24 in the middle, and because of the combination of superelevated curve and grade and helix in the same place had frequent derailments.  We ended up ripping out the spline and replacing it with a carefully laid out 30" radius curve on plywood, without superelevation, to make sure there was a minimum of stress on the cars.  It has worked fine ever since.

Since then I built a 30" radius splining jig to make sure we achieve and maintain our minimum radius when we are building long, tight curves.  I should get a picture of it and post it sometime.

Another old tip...

Do not ballast until after you know it's reliable.

Planning?

Yeah, I agree on the 'planning' idea -- works great provided the tools are up to it. I used to do plant engineering and specialized in installation jobs best described as 'packing elephants into valises'. Tried that approach with this project -- might have worked if I had a reasonable amount of space to work in or less grandiose delusions of what I could accomplish with what 'senior management' allowed could be used. But no.... that is NOT how this one has come down. I have tried pretty much every layout design tool out there and they all have serious problems trying to work within the allotted space for anything non-trivial. In the end I found that buying commercial turnouts and flextrack and playing with them in the space worked better than any other method. In matters of taste we are all quite mad -- its my nickel and my retirement project so if there was some waste, so be it. Not like it really mattered... (A different story when one watches a forklift carry a million dollar custom packing machine through a hole and into a space you designed...)

Did a careful enumeration of the problems... And a lot of it was the parabolic curve issue -- being just a bit too glib in how the flextrack would assume certain curves and using commercial turnouts where they really did not belong. And a few places where gapping in curved flextrack really didn't work without something more robust to hold the curve. And in the end I decided that some of my trackwork was just plain ugly and should be taken out behind the barn and shot, not waiting for sunrise.

So there is a bundle of rail downstairs sitting on the layout and a couple of tools from Fasttracks... got a new learning experience.. I ripped out all the offending track and turnouts, scraped the plywood clean. Will make the east end of the upper deck much simpler and the west side coming into Sydenham more complex... more curved turnouts and maybe a double slip -- and too bad about the control panel -- it was 3/4ths wired up and will be a complete change. Will solder the parts to a printed circuit prototype board once the turnout count settles down.

So, detailed design is wonderful and works sometimes. But I suspect it is over-rated. Sometimes cut/fit/swear/repeat is the only method that works. And after all, it is a learning experience...

An AMAZING curved turnout and track template making tool

that would fix all your problems is called TEMPLOT. And it's actually FREE !!!

http://www.templot.com/

Check out the link and get blown away!

Templot

Would that life were so simple...

I have Templot (from before the author decided to make it freeware) and find it useful although a bit tedious. Good for working out implementation details but not grand architectural renderings. Now that I have decided to hand lay the turnouts on the upper deck to get the degree of curvature I need, this will be my detail tool. But not any more or less useful than 3rdPlanit or Cadrail (or others) for planning what has ended up being a point to point (to point) layout folded around itself to fit into the three dimensional volume I have to work with.

It is the old collision between taste, ability and requirements. For some tasks only the equivalent of a classic 4x5 technical view camera is correct. For others, maybe fingerpaint... And sometimes there is no substitute for marking the lines directly on the wood -- like when making sliding dovetails.

cut/fit/swear/repeat   I

cut/fit/swear/repeat

I think I'm going to put that on a plaque over my workbench.

Long passenger cars

Funny you mention some issues caused by long passenger cars.  I have a passenger train consisting of heavyweight Pullman cars (85 foot, I think) with six wheel trucks.  I have a #6 LH turnout coming off my main line that then turns right onto a siding that parallels the main line.  That's actually a configuration that seems quite common.  

This is no problem with normal sized freight cars, but the S curve created by this configuration has caused some grief with the passenger cars.  In actual operation, the truck of the leading car is into the right turn a bit before the rear truck is out of the left curve.  Even though my minimum radius is over 30 inches, this  S combination seems to create a geometry that binds couplers and causes derailments.  I thought that my 30 inch radius would keep me safe.  Lesson learned.  I guess it's time to re-arrange some track.  Oh, well.  Look how much smarter I am now!

John C

A couple thoughts

I believe the Canadian comes with long-shank couplers (not sure, mine won't be ready for a few more months), but Jason Shron from Rapido has provided some recommendations for tighter curves.  You can put one long-shank coupler on each car and join them long-to-short to get around tighter curves and even go to all long-shank to get even tighter without derailments.  It wouldn't look as good, but some of us just don't have the space for broad curves.

Secondly, though it sounds like you haven't had success with planning tools, I have printed curve templates with easements of different lengths and radiuses that I print 1:1 to make sure I get nice easements and the right radius throughout the curve.  It works great for me and they are reusable.

Yep

Thanks for the comments, guys. I agree with the plaque -- 'cut/fit/swear/repeat', indeed. This is a 'hobby' after all. When I started down this road I was going to use flextrack and commercial turnouts (even though the fancy ones are bloody expensive). That might have worked if I had plenty of room -- but between my delusions of grandeur and the 7'x10' space allocated to this project by senior management and my reaction to the size of 'N' equipment compared to my hands... it has been pretty interesting. I have read a lot of good information and advise -- especially my well-thumbed copy of 'Armstrong Track Planning'. Problem with all of this is integrating it with the reality of the space in front of me - there is always that all too human sense that 'x' doesn't really apply to me or I can get away with it. I will be doing the coupler transplant on my Canadian -- the instructions are hysterical, by the way. Was my first question to Rapido when I decided to order the thing...(a very long time ago). What I am realizing is that there is a point where you just cannot get there from here -- the constraints of commercial units mandate certain kinds of geometry. I ran into this doing the big yard -- had to chop the #4.5 turnouts down to a point that they had to be soldered together into one big multi-branch. So I guess I have been edging into the hand-laid arena for some of my trackwork for some time. Now waiting for the rail roller as almost everything that needs to be built is curved. Templot2 will be useful for this -- it wasn't before. And it has been really useful to use my digital camera up close to the troublesome areas to film what is happening -- like the 'S' curve issue (see Armstrong for what he says about that one...). Anyhow, thanks for the support. Nice to know I am not the only loon in the hobby who screws up and has to do it over.

Not sure about this

If your turnout angles are to sharp to run well (and you have lots of reverse curves with no straights in between), then hand-layng them won't make things any better.  Even curved turnouts need the same precautions and transitions.

Longer car end Coupler fouling is a transition curve problem rather than absolute radius.