I have seen many posts from various members about this subject. It is difficult to trouble shoot via the inter net, because trains are dynamic models, and it is sometimes necessary to actually see what is happening to cause the derailing. This will be a long post, but I hope that I can put in a comprehensive tutorial that anyone can use to trouble shoot derailing problems on the railroad. By the way, I model in ho. I think the same dynamics apply from z scale to LGB, but I have no experience with scales other than ho, so you guys modeling in other scales will need to scale up or down accordingly.

First off, I'm going to suggest that there are three areas where we can find the cause of the derailing of our cars and locomotives.
#1 tracks.
#2 incompatibility between rolling stock and radius.
#3 rolling stock problems.
I've found the problems will generally fall into one of those areas, and the plan is to put up a comprehensive post to cover all contingencies if possible. On second thought, I think I'll do a general over view in this post and then do the detailed analysis in each area in subsequent replies.

How to figure out if a problem is track related, rolling stock related, or incompatibility between rolling stock and track radius.

First, does all rolling stock derail regularly at the same place on the layout? If the answer is yes, you probably have a track problem.

Is the problem section curved or straight track or is it a switch?

Does only some of the equipment derail at the same place, but other equipment works fine there? If the answer is yes, does the rolling stock that has problems have something in common? For instance, do 40 foot and 50 foot cars work fine, but 60 foot-85 foot cars have problems? Yes? Probably a radius problem.

Do you have one or more cars that derail in a particuar spot if they are coupled directly behind a specific type or size engine, but seem to work fine when hooked to other engines or with other cars between them and the engine? Yes? Probably a compatibility problem between radius and equipment.

Do your 4 axle Gp & f type type locomotives and 4 axle freight cars work fine through a section of track(maybe even a straight section), but 6 axle diesels, steam engines, and 6 axle passenger car trucks derail? Yes? It could be tight radius problem or a bump or dip in trackwork that knocks the long trucks off the rails.

These are all of the questions I can think of to ask to help narrow down the source of problems. If anybody thinks of other questions that should be asked, post them and I will edit this post.

I am now going to do 3 replys to this post discussing track problems, radius incompatability problems, and rolling stock problems respectively. If I overlook anything that should be included, just post it and I will add it to the appropriate post.

My goal is to have a thread in the tips section with 4 posts that will be comprehensive instruction on how to trouble shoot, find, and repair any problem anyone might encounter in keeping their trains on the tracks.


I think track can be catagorized as straight, curved, switches, or crossovers.

I am going to link some web pages from the NMRA. The first page is a page from the NMRA store, and I'm linking the nonmembers page. All of the items in the NMRA store have both members prices and nonmember prices, the non member pages show nonmember prices. You can go to the NMRA site and navigate to the member store pages to get the member prices if you are a member, but this information is intended for nonmembers as well as members to find.

The first tool I would recommend for anyone building a layout is this one
<!-- m --><a class="postlink" href=""></a><!-- m -->

The NMRA gauge can test track gauge, flangeways & frog clearance, tunnel portal clearance, loading dock and station platform clearance, wheel gauge, and coupler height. There may be some other things it tests as well that I'm forgetting. It is as close to a universal measuring tool as I've seen. It is the essential tool to test just about everything that has anything to do with track. You can get them at hobby shops for the nonmember price or maybe a dollar more, but to me it is an essential tool for anyone building a model railroad layout. They offer them in sizes from n scale to O scale (not Lionel O gauge). If you model Nn3 or Z, Lionel, Marx, or LGB I don't know of anything for you to use. Modelers in those scales may help out with suggestions here. Get one in the appropriate scale for the layout you are building.

Things to look for in trackwork:

1: Make sure all of it is in gauge. I can only think of 2 causes for commercial track to be out of gauge. You could get a piece of defective track (I've never heard of this problem). I think it is a very rare situation. If you buy used track at a swap meet or on E bay, you might get damaged track. The other and I think most common problems for commercial track being out of gauge is improper installation or assembly of track. All of the commercial track except the old "True Track" (out of production for many years) use plastic tie strips. There was some track manufactured in the 1960's that used a paper/fibe tie strip with staples to hold the rails to the ties. You don't want to use any of that stuff. The only thing it is any good for is as a source for bare code 100 rail in ho scale. I don't know if it was offered in any other scale. The most common way to knock commercial track out of gauge is to use track nails to mount your track to your roadbed and then drive the nails in too tight! If you use track nails to mount roadbed, make sure you leave a little clearance between the bottom of the nail head and the top of the tie. You could make a "feeler gauge" out of a business card, and make sure when you put in your nails you have clearance to slip the business card under the nail head and above the ties. What tends to happen is the nail is driven in too deep and the tie bends pulling the rails together.

2: Do transitions for every curve both horizontal and vertical. What is a vertical curve? When you go from a flat level run of track ans start up a hill, or have been climbing a grade and level out on top before going down again, you have encountered a vertical curve. If you have a "whoop de do" in your benchwork or roadbed surface that causes a dip or bump in the track, you have a vertical curve. Sometimes you might have a spot on your track where 4 axle diesels and rolling stock goes through fine, but 6 axles diesels, passenger cars, and steam engines seem to derail from time to time, even on a straight section. Look closely while running the equipment that has a problem very slowly across the problem area and see if the front or rear wheel set of one or more trucks tends to lift off the rail. A bump or dip in the rails on a straight section may derail long wheelbase trucks sometimes, a bump or dip in a curve will do it just about everytime. If you use cork or even some of the new rubber roadbed, make yourself a long sanding block. Get a 1x2 8 inches long and sand it smooth to eliminate any slivers. Now cut a piece of 100 grit sand paper to just fill the 2 inch wide side (you will need a 1x3 for O scale) and glue the strip of sand paper to the sanding block you just made. Now before you lay any track on the freshly installed roadbed, run you sanding block over it to get any bumps out of the roadbed and to bring high spots down to eliminate dips. Also make sure switches are always installed on level "ground." Switches will always cause problems if the base under them is uneven. If you are going to have a switch taking a siding off the high ballasted mainline onto dirt for an industry, keep the entire switch up on the ballasted "high line;" and then let the siding transition to dirt after it comes away from the switch.

To transition from level track to a grade or a grade to level track, I would recommend using plywood subroadbed correctly. I think the new foam grade making stuff from Woodland Scenics probably has grade transitions manufactured in, but I've never had any experience with the product so I can't comment on it either way. When you use a plywood base under your roadbed, DON'T EVER start your grade transition at the point where 2 pieces of plywood butt together! If you do that I can just about guarantee you that you will get a vertical curve without any transition. You may not have derailing issues if the problem is not on a curved piece of track, but you will probably get mysterious uncoupling of your train from the first locomotive that crosses through that vertical curve. The dynamics of what is happening is the locomotive tips up as it starts it's climb, and the rear knuckle coupler drops down and slips out of the coupler behind it. Secure your plywood subbase a few inches before the grade is to start, and bend the plywood into the grade. If the grade is not too extreme, the plywood will form a natural transition into the vertical curve at the start of the grade. In the same way, when you top out at the top of the hill make sure your transition is in the plywood and not where the 2 pieces join. What happens when the train goes "over the top" at a poorly transitioned grade is just opposite of what happens at the bottom, but the results are almost the same. In this case the rear coupler of the engine or a long car swings up outt of the grasp of the knuckle on the car behind. I said the result is almost the same. If you uncouple at the bottom of the grade, the part of the train that uncoupled stays at the bottom of the grade while the rest of the train climbs the hill. If you have an unwanted uncoupling problem at the top of the grade, the uncoupled cars will run uncontrollably at fairly high speed back down the hill, and may get up enough speed to fly right off the track!

For transitions from straight to curved track in a horizontal plain, you must use flex track. Layouts built with sectional track will not have transitions, and they will work, but some long equipment that might make it through a tight radius with transitions may derail when trying to go through the same tight radius without transitions. One trick that might work if your layout is made up of 18 inch minimum radius curves, is to use a section of 22 inch radius at the beginning and end of each curve. If you are already using a 22 inch minimum radius, I don't know of any commercial sectional track bigger than 22 inch radius. This thread was copy pasted from the old site. Since I originally wrote this, manufacturers have come out with 24 and perhaps 26 inch radius sectional track in ho because of the extra length of the modern super power diesels. There have been many methods discussed on how to make transitions by bending yard sticks or bending flex track, or bending what ever. There are also mathematical formulas to tell you how to compute a spiral transition into a curve, but I can't remember those either. A method I was shown that is easy and works great for me is to start and end the first 3 inches or the last 3 inches every curve with the track offset about 1/2 the width of the track outside of the nominal curve radius. In effect the true centerline of the curve will seem to line up on the inside rail of the straight track leading into the curve.

By the way, all commercial turnouts are made with a transition curve built in to any diverging route. In the case of a wye turnout, both routes will have transitions built in from the manufacturer, and the same thing is true for any commercial curved turnout.

2: Do "S" curves correctly. I deliberately did not say to avoid "S" curves. I think a train snaking through a correctly done "S" curve is a beautiful sight! The problem created by S curves has to do with the dynamics of rolling stock being pulled and pulling on the ends. If you locomotive has just come out of a right turn and is going into a left turn, the locomotive will be pulling the car following to the left; but the rest of the train will be pulling that same car to the right. The car will want to take a short cut and cut the corner, but there are no tracks there. The result is a derail or cars laying over on their sides. The solution is to put a straight section in the middle between the right and left turn so the train goes right, straight, then left. The length of the straight section should be equal to the longest car that is going to go through the S curve. There is one exception to this rule. If you are modeling a shortline or branchline that is served by a peddler frieght with one passenger car or combine running at the end of the train instead of a caboose, you can make the straight in the S curve long enough for the longest freight cars that will regularly use the trackwork. Because there is nothing behind the longer passenger car or combine to pull on the back of it, it shouldn't have any problem with the S curve although you will have operate at restricted speeds. Then again, peddler freights operate at low speeds anyway, and most branch lines or short lines running a peddler freight will not have the best trackwork in the country anyway.

Something else that has come to my attention since posting this thread originally is that the guard rails on Atlas turnouts are too far from the rails they are protecting. The stock Atlas switch will allow wheels to fall off the rails and land on the ties between the rail and the guard rail. The fix is simple. Just cut a piece of .015" styrene to the same height as the guard rail and glue it between the guard rail and the main rail. Do it to all guard rails on Atlas switches and then paint the "shim" to match the guard rails.

I'll start this off with the results of some experiments that I did on my first layout. In fact the description of the experiment and the results might make further discussion of this point unnecessary.

On my first layout, I used flex track, but I made 22 inches my minimum radius for the mainline. I had some sidings that were 18 inch radius, and then I had a couple of industries that were a tight fit (didn't want to waste any real estate) so I installed some 15 inch radius snap track.

This layout was built in the early 1980's, and I was using Sd45's & Sd40-2's for mainline power. I had Gp7's for local switchers, and Sw7s for the yard. The plan from the very beginning was that the Gp7's or the Sw7's would always switch those industries with 15 inch radius curves leading into them, never an Sd45.

Strictly as an experiment, I tried running my Athearn Sd45 through the 15 inch radius curves, and it worked fine although it looked ugly doing it. Then I tried coupling a 40 foot boxcar to the Sd45 and running it through the 15" curve. The 15" curve went to the right out of a #4 turnout. The Sd45 was able to push the boxcar through the turnout and into the 15" curve without any problems until the locomotive started into the 15 " curve. As soon as the Sd45 went into the 15" curve the boxcar tipped up on its side. The left side wheels stayed on the track, but the right side wheels raised up about 1/4 inch off the rails. Once the locomotive came out of the curve, the boxcar dropped back onto the rails. When I pulled the Sd45 back out of the siding with the boxcar in tow, the same "up on 2 wheels back down on all 4" happened again. The final experiment was to couple a Gp7 to the Sd45 and then couple the boxcar to the Gp7. when I did it that way everything worked fine.

I learned about dynamics of locomotive length and coupler restrictions vs radius from that experiment.

If you have a curve that a particular freight car tends to derail in every time when it is coupled behind a particular engine, try a different car. If it does it consistently, it may be that the curve is too tight for that particular locomotive. Also the dynamics will change with freight car length. You may find the a 40 foot car will follow that big locomotive through the tight curve fine, but a 50 foot car or a 60 foot car has problems.

This is something that Fred Wright has posted on a number of threads dealing with minimum radius and rolling stock. If I'm remembering correctly, Fred has suggested that your minimum radius should be 3-3 1/2 times the length of the longest locomotive or car that you want to run. As an example an 85 foot passenger car in ho will be almost 1 foot long. The traditional recommendation for 85 foot passenger cars is a 30in minimum radius, but using Fred's formula, 36-40 inches would be a better choice. As model railroaders we seem to have a constant battle going on between the amount of room I want for the radius I want versus the amount of space I have. I will discuss talgo mounting vs body mounting of couplers in the next post, but one trick the manufacturers use to help long passenger equipment negotiate tighter radius curve is to mount the couplers to the trucks rather than body mount them. The body mounted couplers will always track toward the outside of a curve, the distance to the outside depending on the overhang of the particular piece of rolling stock. Talgo mounted coupler (coupler mounted to the truck) will always track to the center line of the curve. As a result, cars with talgo trucks will negotiate tighter radius curves when being pulled than cars with body mounts. Talgo's don't push worth beans! As soon as you try to push one the coupler wants to shove the truck off the track to the outside of the curve. If you want to run a passenger train, but don't have the room for the radius needed for full length cars you have 2 choices. You can run "shorty" passenger cars, or you may try cars with talgo trucks. You just can't back a talgo equipped passenger train very well at all. Also, if you locomotive has body mounted couplers, you may need to modify it to allow more swing on the locomotive couplers to allow it to track without pulling the cars off the track. The other thing you can do is make a "transition car" out of a baggage car or express box car or reefer which are typically shorter than the 85 foot full length passenger cars. Body mount the coupler that hooks to the engine and talgo mount the coupler that pulls the rest of the train.



If you have one locomotive or freight car that derails consistently at one spot on the railroad, but everything else you operate works fine at that spot, is it a track problem or a rolling stock problem?

If the problem piece of rolling stock is a unique piece that is longer than anything else, has longer wheelbase trucks than anything else, etc., you may have a track radius problem or a defect in the trackwork causing the derailments. If the problem piece is one of 2 or 3 or more identical pieces, and only one has the problem, you probably have an issue with that one piece of rolling stock.

Sometimes a particular car will work ok if it is just turned around in the train. That is fine if you want to watch you train run over the layout, but it is not acceptable if you want to do prototype operation. To do prototype operation, your rolling stock has to function correctly without derailing in either direction whether being pushed or pulled through any point on the railroad. In this section I will focus on tuning up rolling stock for it;s most efficeint operation.

The first thing I will mention is that most toy train sets come with what are called "talgo" trucks. That just means that the couplers are mounted to the trucks instead of to the body. That type of coupler mounting is used to allow the train to negotiate tight radius curves. It may work with Lionel & LGB large scale trains due to the very heavy weight of those models. I don't have any first hand experience with either one of those trains, so I can't speak about them. With ho and n scale trains, the dynamics of pushing on a car with talgo trucks is that the coupler will have a side force exerted on it that will try to push the coupler to the outside of the curve. With the relatively light weight of the individual car combined with a relatively heavy weight of the train it is pushing against, talgo trucks will always tend to derail to the outside of the curve you are pushing them through, and sooner or later they will derail 100% of the time they are being pushed. The problem is compounded because all of the best "model railroad locomotives" will have body mounted couplers, and most r-t-r or kit built model railroad rolling stock also will have body mounted couplers. When rolling stock with body mounted couplers goes through a curve, the coupler will swing over toward the outside rail. The talgo coupler will track through the center of the rails. Basically we have an incompatability problem, and that incompatability must be addressed before we go farther. The trainset locomotive will also have talgo trucks, but because the trainset locomotive is light in weight, usually with a poor quality, poorly designed drive system, they are not very good pullers anyway. Fortunately the fix is easy. The couplers will snap out of the trucks easily. You will also want to replace the plastic wheel sets with metal wheels for better operation so we can fix it all at once. Remove the coupler and wheels from the toy train trucks. Cut off the mounting tab that holds the coupler, and body mount the coupler. That brings me to my first tool recommendations.

<!-- m --><a class="postlink" href=""> ... t&ID=82838</a><!-- m -->

<!-- m --><a class="postlink" href=""></a><!-- m -->

The first tool in combination with a metal wheel set will make your cars free rolling.

To body mount couplers, most cars today come with some sort of knuckle coupler similar to the Kadee #5. Most of the new couplers will fit a Kadee coupler box, or you can just mount a Kadee #5 coupler (or Sergents or whatever your preference). Choosing your couplers is beyond the scope of what I want to address here. Take a suitable coupler box to mount to your rolling stock, and assemble the coupler/coupler box. Drill a #45 hole in the bottom center of your freight car's end. Tap the hole for a #2-56 screw, and fasten the coupler box to the underside of the car with a suitable length 2-56 screw. that brings us to the second tool on the list, the Kadee coupler height gauge. The Kadee gauge will fit over a section of track, put you car on the tracks and roll it over to couple with the coupler on the Kadee height gauge. When it is right, the coupler knuckles will line up perfectly, and the trip pin will just slide over the top of the little ledge at the bottom of the gauge between the rails.

Now we will presume that all of your rolling stock is compatible in terms of coupler heights and all of it has body mounted couplers. but this one car still derails at one point on the layout. We need to give the car a complete tune up.

The first thing I do is set the car on a piece of track on a perfectly flat level surface. With this test we are checking to see if the car's frame is warped in any way. What you are looking for is for all of the wheels on both sides of the car to sit properly on the rails. You might also want to check the car sitting directly on your flat surface to see if all flanges sit on the surface with no space between any flange and the surface. Athearn 4 bay open hoppers are famous for having warped frames, by the way. If a frame is warped, we can sometimes put it on a flat surface and put weight on the frame like a heavy book and leave it sit overnight to see if it will straighten out. This can only be done with a disassembled kit.

The second test I do with the car still sitting on the track is to check for the frame or body being bent, is the side of the car parrallel with the track?
Are the side frames of both trucks parrallel with the track?
Is there any tendency for either truck to appear to "crab walk" along the track due to one or more axles not being in perfect 90 degree alignment in relation to the truck side frames.

Now roll the car along the track section, is there any sign of a wheel not being centered on the axle indicated by one or more corners of the car going up and down as the car rolls?
Does the car or any of the wheels show signs of wobbling when it rolls on the test track? I have a Bachmann Plus F-7 unit that has one wheel on one axle that is not straight. When that locomotive runs down the track, it wobbles. If you find a wheel that either isn't concentric or wobbles, the only fix is to replace that axle & wheel set.

Now we need the next tool.

<!-- m --><a class="postlink" href=""></a><!-- m -->

This is The NMRA gauge. It is one of the most valuable tools you can have in your modeling box for model railroading. It is almost a universal measuring tool for model railroads.

We are going to use it to check the wheel gauge on our car. We need to check each axle individually. If you look closely around the outside of the gauge, you will see some small slots cut in the gauge that are marked for testing wheels for gauge. If both wheels on one axle won't fit into the grooves because the inside of one wheel interferes, the gauge is too tight. If it doesn't fit because one wheel is too wide for the gauge, the wheel set is too wide. Often times the wheels are an interference fit on the axle and by removing the wheels from the truck and twisting them in opposite directions while either pushing in or pulling out, you can adjust the gauge of the wheels.

The next test requires our next link.

<!-- m --><a class="postlink" href=""></a><!-- m -->

This link is to the NMRA recomended standards for weight of cars. find your scale on the chart and the length of the car and you can see what it should weigh. You can get inexpensive postal scales, or an inexpensive scale designed for weighing food portions on a diet. Both will work for what we are trying to do. If your car is too light, add weight to bring it up to the NMRA standard. On closed cars it is easy, you can glue pennies inside to bring up the weight. On open cars it is more of a challenge. You might use a load to add weight. I've used small lead shot gleued in voids in the frame with epoxy to add weight. You might change to metal trucks like a set of Kadee freight trucks to add some weight. You can even take a car apart and replace a steel or zamac weight between the car body and frame with sheet lead to add weight.


There is one final test that need to be done on our car before we can declare that it is "tuned up."

Turn the car upside down and check the trucks for proper torque or tightness. We want a three point mounting for our trucks suspension system. To get that we want one truck just tight enough that it swivels freely, but doesn't rock. The other truck we tighten the same, then back off just enough to allow the car to rock side to side slightly.

We are done. If you have carefully followed every step in checking out you problem car, and done any corrective repairs needed, you should be rewarded with reliable operation.

Jim Thompson mentioned one other cause of rolling stock problems that I forgot. His post is in the thread a few posts down from this, but I'm going to rewrite it here in order that these first posts at the top of this thread might truly be comprehensive. I brought this post over from the other site, but I left the replies over there, so Jim's original post is still at the other forum.

You may have a car that derails, and when you check it out, everything seems to be in gauge and in spec. as far as weight, truck tightness, etc. The other thing to check is for interference between the truck and the frame, undercarriage, or details. This is especially important with cars with highly a detailed undercarriage, or one which you may have superdetailed the under carriage of a standard kit. The radius that the prototype uses is much bigger than anything we run. It is very easy when super detailing the undercarriage of a car with brake rigging, etc. to have a part that interferes with your truck swivel when negotiating the tightest radius curves on your layout.
Wow a very informative read. I will have to read it a couple times to take it all in.

New Adventure <!-- l --><a class="postlink-local" href="">viewtopic.php?f=46&t=9245</a><!-- l -->

Great White North
Thanks to Russ for recreating this comprehensive trouble shooting guide!

This thread was moved to The NEW Academy on August 13, 2009 and is now closed to further comments. Some posts may have been removed or edited for the sake of clarity. If you have any feedback on this topic, please create a suitably titled post (e.g. Question about TOPIC in the Academy) in the Upper Berth.

Note that there is a follow-up thread here --> <!-- l --><a class="postlink-local" href="">viewtopic.php?f=3&t=1726</a><!-- l -->

for The NEW Academy

Forum Jump:

Users browsing this thread: 1 Guest(s)