Check the sharp ends of the points. I encountered several snap switches (bought in 1970s) with squared-off ends that wouldn't fit all the way into the stock rail notches.
If that's the case, the cure is to gently file each point to a sharp point. That eliminates the 'step' that flanges use to climb up and over the rail.
Snap switches are built to radius, as Crandell mentioned. If they are built to the locomotive's minimum radius, even the slightest deviation from perfection will put the wheels on the ties.
I, personally, have the ultimate solution. I hand-lay all of my specialwork - and can take the time to assure proper gauge and clearances. As for the points, I DON'T run my fingers along the rails into the twin razor blades of a facing-point switch.
Chuck (Modeling Central Japan in September, 1964)
Make sure the turnout is level. Even if it is level, I have cured some derailment problems by slightly shimming the the outer rail right at the point the wheels try to climb out.
First, I wish you hadn't said it is a snap switch. They are curved throughout, if I have that right about your version...? That changes things because the closure rails on a standard N. American turnout straighten by the time they get to the guards. It may be that this one loco just can't negotiate that one turnout.
However, turnouts' rails are sometimes not all at grade. Some come from their tooling with high or depressed frog rails, as an example, or they are simply torsioned clockwise or counterclockwise when looking at them down their main axis from eye-level at the points end. It sounds like you may have a dip in the curved closure rail, or there is a height disparity somewhere. This assumes your gauges, whether rails, flanges, or flange paths between guards and rails, are all to spec.
-Crandell
There was a discussion on here a few years ago about the Atlas turnouts being removed from the molds too soon. There is a slight warping of the center of the turn out, the cure is to put a track nail in the dead center to make it lay flat. I also had the same problem and the extra nail cured it, you may have to put a 2nd nail in so take a straight edge and lay it along the rails to check, if you see daylight thats the problem. By the way this is both gauges N and Ho. Hope this helps, Jim.
It might be that with time the underlying benchwork has expanded, contracted or warped a little but, just enough to cause the turnout -- always the most derailment prone part of a layout anyway -- to be ever so slightly twisted. If you take a flat car and really press down on it with a finger as it runs through the turnout (both routes) do you sense it twisting or changing its level?
Steam locomotives and certain rigid 6 axle diesels such as the old Athearn FM Trainmaster or ALCO PA are always the first to find these problems.
Dave Nelson
Welcome to the foms.
Have you used an NMRA track gauge on the turnouts? Since they are older turnouts use and handling may have knocked them out of gauge.
Also, are they Atlas Snap Switches or their numbered turnouts (#4,#6 and #8). I have seen many posts and experienced myself, that the Snap Switches are occasionally out of whack. Matter of fact I had a thread recently due to one locomotive I have derailing on two turnouts when it goes in one direction. All others, steam and diesel have shown no problems in either direction. I have found a combination of several things beirn just a little out, when put together, are causing the problem. Slightly warped loco truck and turnout just a little out of gauge and off it goes. Still have a couple of other things to check to see if they will help.
I think you will find that #6 perform better, especially with longer cars and locos, as they have a lesser angle than Snap or #4's. I plan to use #6's off the main on my next layout, will problbly use #4's in the freight yards to save space.
Good luck,
Richard
Crandell,
I have run it through crawling but I get fixated on that leed driver riding up ove the rail and everything else following. I am going to light it up good and my camera shoots HD video so maybe then if I cant see it maybe experienced modelers can help. Thanks again!
The derailments may have nothing to do with side-play. It might be interference with a guard rail, tight gauge, points not lying tight against the stock rail as more of the axles pass over them on their way to the frog (hinged ones get sloppy after a while, allowing the points rail to tilt over. Maybe an axles on the steamer has gone wide or narrow in guage for some reason, maybe even just handling.
I think every loco I have, except maybe my Sunset Brass 2-10-4, would go through a genuine NMRA #5 turnout. Certainly they all take a Peco Code 83 #6, including that brass CPR Selkirk.
Have you lit the affected location brightly and passed your steamer through it both ways, and also in reverse both ways, to see what is going on? You need to do this slowly, maybe a scale 5 mph, and stop the loco as soon as you see a wheel or axle start to lift. That MAY be the trouble spot, but often steamers are showing you where the symptom is, and not where the problem is. The problem may be forward or aft. A lead axle may be lifting on a high frog, and since the loco is turning with the closure rails, the next axle will want to pop out. Or, the rear of the loco is being levered some way such that the axles forward can't negotiate the gauge properly and pop out. Happened that way on my Lionel Challenger with a #5 EZ Track turnout years ago before I switched to flex, Pecos, and hand laid turnouts.
I inherited my dads layout and trains recently after his passing. Most if not all of his equipment is new old Atlas track from the late 70's and all the rolling stock and locos as well. He built a layout a couple years ago and I am in process of fininshing it. My problem comes in that I am getting alot of derails at the turnouts. Mostly with the steam loco's. Even crawling the drivers want to climb over the switch. I am have changed most of the them to #6 and it;s gotten better.
I am curious though are there just some locos that wont have enough side to side slop in the drivers to ever use anything less than a #6?
Thanks!