HELP! Kadee couplers unreliable in unit train!


Ignorance is Patriotic
Hello everyone,

I have a major problem and I am frustrated to no end. I have spent the last year or so assembling a unit train, 60 cars to be exact, using Eel River Models' (ERM) 62' beer car. Finally, the other day was its first test run at a local modular layout here in Denver.

Much to my dismay, the 60 car train wouldn't stop uncoupling and disconnecting; when I backed the train up to pick up some cars that had uncoupled, the problem worsened. It seems that every time I go in reverse, then move forward, the train will uncouple in several places and segment into 5 or 6 pieces. Every time the train begins to move after being stopped, or after being put in reverse, the grasp between the Kadee couplers is lost. What's more, when I use my hand to move a string of cars that have fallen off the back of the train, the cars that I move will uncouple and another piece of train disconnects. When I move yet that other string of cars, lo and behold, the last five cars uncouple and the train is literally in five or six separate segments. Simply put, I can't keep this train together as it travels over the layout.

I am using Kadee #26 couplers, as per the instructions on ERM's 62' beer car. ALL OF MY COUPLER HEIGHTS MEET EXACTLY THE NMRA HEIGHT USING THE KADEE HEIGHT GAUGE. THE COUPLERS ARE CENTERED USING THE CENTERING PLATE. I have hundreds of freight cars using Kadee 5's, and have never had any problems. I can run 100-car trains of my other rolling stock with no major uncoupling issues; as I mentioned above, all of my coupler heights are standardized. Using Eel River Models' beer car, and using Kadee #26 "long shank" couplers, this 60-car unit train is in a complete mess, the train will not stay together and seperates at every chance it gets: when moving forward (picking up slack) after being stopped or put in reverse, at slight curves, etc.

The uncoupling issue does not seem concentrated to a few specific cars, the problem is affecting all 60 cars randomly and frequently. I have inspected all cars several times, and the coupler heights are standardized and the coupler efficiently centers using Kadee's brass centering plate.

I have provided the following graphical timeline of the issue, step-by step. Let's start with STOPPING THE TRAIN, THEN MOVING FORWARD...

Step 1: The train is stopped, at a complete standstill. The train them moves forward. Instantly, the Kadee couplers lose grip with one another when the slack is picked up:


The train is uncoupled. When I move the string of cars to reconnect to the train, another uncoupling occurs 5 cars down the line, in the same manner.

Let's now put the train in reverse to pick up some uncoupled cars, shall we?

STEP 1: Here we can see the couplers being compressed, shoved together, as the train is being pushed in reverse:


STEP 2: Now we move forward. As slack is picked up, we can see the "claws" of the Kadee couplers are not meeting in a gripping manner. Sometimes they decide to "click" inward and re-establish a bond, but most times they decide to fall outward, losing their "handshake" so to speak. This is clearly going to happen here:

STEP 3: Slowly, as the train is moved forward, the couplers decide to part ways...


The train is uncoupled and, as I look back to the end of the train, I notice uncouplings at several places throughout. The train is now in 4 or 5 pieces...

CAN ANYONE TELL ME WHAT IS GOING ON HERE? These couplers should be in perfect working order, but this uncoupling problem is so pervasive and happens so frequently that I am about to give up on this train.

The only explanation I can conjure is that there's something peculiar with the physics of the car and the #26 coupler, a problem with the design of the car, or placement of the coupler mounting bolster at a certain measurement so as to exploit a vulnerability in the #26 coupler's operation.

Assuming the "long shank" couplers contributed to the problem, I attempted to outfit some cars with a standard #5 coupler, but they don't work due to the car's design. ERM placed the coupler bolster, or mounting pin, recessed too far back into the coupler pocket to accomodate a standard shank coupler; a long-shank coupler such as #26 must be used just so the coupler can clear the end of the pocket. Looks as though I'm stuck with the #26's, but this is not working. This train will not stay coupled together.

I have debated using shelf couplers, such as those used on tank cars, to not only assist with this uncoupling problem, but also to help me solve uncouplings due to vertical shift as cars rise and fall over uneven trackwork and between modules. Question is: does anyone make a "long shank" shelf coupler?

I've thought about using dummy couplers too, but again, not sure if anyone makes a long-shank version that will clear the edge of ERM's coupler pocket. Another drawback to dummy couplers is that I would like these cars to be compatible with other rolling stock using Kadees, McHenry's, et al.

I don't know what I'm doing wrong, guys. These 60 ERM cars have passed vigorous coupler inspections, and exactly meet standards in every way. Theoretically, this train should work and remain coupled, but it breaks apart into several segments so easily; the slightest jar will cause several uncouplings throughout the train.

Can anyone offer advice or direction as to where I need to go with this project? Thanks for any help, and apologies for this LONG posting. Can you tell how frustrated I am?

Your friend and fellow Model RR'r,

Try and compare the #26 to your #5's those don't look right to me?I've been useing KD'S for over 20 years and never had a problem like that?HUMMMMMMMMMM? Maybe they are lose in the box?and tourq some?The only change I've done is start replaceing the brass spring with the new wiskers,as the old ones get broken.Yup I slam them to hard every now an then when shoving up to and track end!!
Something looks wrong to me also. It looks like the couplers are slightly twisted as they try to couple and so can't complete the coupling. My suggestion would be to contact Kadee Service Dept. and tell them of the problem. They may be able to suggest a solution. I have never used the#26 couplers, but have used several different kadee types over the years and they usually work flawlessly. Hope you find a solution to this!
This may be a dumb question, but I am not familliar with that make of car. Are they dicast metal? The coupler box looks like bare metal on the sides. Anyway, I thought they may have become magnetised some how so the couplers are repelling each other. If you don't use magnetic uncouplers, you may want to try trimming some trip pins to see if that helps.
Just my 2 cents. Hope you can figure out the problem.
No, the coupler pockets are plastic. The only metal on the cars are the Kadee couplers and the wheelsets.

I sent an email to Kadee asking for their help in solving this issue; it is, after all, their product allowing all of this uncoupling in the first place. In all my time using their products, I have never seen their couplers simply lose grip with one another for no apparent reason. The coupler can be in a "locked" position, then when the train moves forward, the claw of one coupler "misses" the claw of the other, sliding right by its side, leading to an uncoupling.

I have spent hundreds of dollars on the #26 for this train, and related Kadee products. I wonder how much help they can be, though, when they don't have direct access to these freight cars and can't see the problem occuring first-hand. I willl let all of you know the degree of assistance I receive from Kadee, and how strongly they stand behind their products.

In the meantime, though, I would still like to solicit for your advice/recommendations, perhaps even alternatives to Kadee. So much for reliability...
Last edited by a moderator:

I have installed the #26 couplers on different freight cars, and tested their ability to remain coupled after being compressed and pulled apart. Turns out that even on these different cars, the #26 couplers had a tendency to separate. I then tried the same test with the #5 coupler, and the problem did not surface.

Therefore, I have narrowed down the problem: There is an inherent design flaw with the #26 coupler that will lead the "claws" of the coupler to brush past each other on the outside rather than grasp each other from the inside, as can be seen in this picture:


This picture illustrates the claws about to bypass each other rather than bite into each other as intended. When these couplers come apart, they are going to slip away from one another.

The "trick" to finding this lapse in performance is to place both couplers in a certain position; it will take lots of different combinations, but there is a certain positioning of the couplers that will exploit this weak spot and cause the claws of the coupler (or the ridge at the very end of the knuckle) to bypass one another rather than lock inwards with each other. This, of course, will lead to an uncoupling.

After being compressed, the couplers will give up slack and each will have a tendency to move to its respective right. There is just enough room between the coupler connection to allow one coupler to slip out of the bond, and a separation occurs. Backing a 60-car train of #26 couplers in reverse will cause all sorts of different coupler positionings between the cars, so it should be no surprise to me that a portion of coupler connections in my train were broken. The couplers simply "slipped away" from one another when the train moved forward.

As I mentioned earlier, this problem does NOT surface with the #5 coupler, and I'm not sure why; it seems the #26 is an exact twin of the #5 besides the longer shank. The problem most definitely is the #26 coupler. Perhaps Kadee did not make the claws long enough for a guaranteed, flawless bond?

My 60 ERM cars have a long shank coupler pocket, a #5 will not fit. I am forced to use a long shank coupler, and I'm not ready to switch over to the plastic McHenrys. (this is a heavy train, 60 cars weighing 8 oz each). I am forced to use Kadee #26, but what can I do when I have identified a flaw, or "weak spot", in their design?
Last edited by a moderator:
To me it looks as if the knuckle is not closing far enough. Maybe there was a bit of flash in the casting or something to keep it from closing all the way.
I have noticed on a couple of occasions that I have had to take a small round file to clean out the burrs inside the knuckle. Your pictures look as though the couplers are not closing properly which may indicate some flashing.
It definitely is a flash problem on the knuckle itself. Your pictures clearly show it.

I went down to my workbench a minute ago and opened the only pack of 26"s I have and compared them to some #5's. The 26's have a large amount of flash on them, especially on the knuckle and coupler face. The 20 series are the insulated couplers which means the shank is metal but the coupler knuckle and coupler body is engineering plastic. The coupler then is made from a two part molding, and if the dies don't come together perfectly there is flash in the parts.

Just remove the flash as GB sez with a file or a #11 blade and repeat the test. Bet the coupler functions as it should. Just make sure you clean the knuckle inside and out as well as the coupler body.
Last edited by a moderator:
I say just start the process now of switching to EC87's, considering they couple like a real coupler, the only downside will be having to have one coupler "open", but hey its realistic!

Gentlemen, I have discovered the problem, and it has nothing to do with flash on the coupler. Turns out the #26 coupler is shaped somewhat differently than the #5, and its design makes it inferior to the latter.

I decided to compare both couplers, side by side, under a magnifier to learn why the #5 worked reliably while the #26 didn't. The answer?

The "thumb" of the #26 coupler is thinner than the thumb of the #5 coupler, as can be seen below:


As can be seen above, the #5 coupler on the left has a thumb that is substanitally thicker than the #26 on the right, it is obvious to see this difference under close inspection. What does this mean? Well, obviously the #26 coupler has a lot more "leeway" and due to its thinner thumb, does not securely hold the claw of the opposing coupler into the connection. With a larger distance between the end of the thumb and its own claw, the #26 easily allows the claw of the opposing coupler to "slip out" or "slip away"; the gap is simply too wide. The #5, with its thicker thumb, creates a more cozy connection and doesn't allow any room for the claw of the opposing coupler to escape the bond. The mechanics of this scenario are obvious and understandable when the comparison above is studied.

What have I decided to do? Well, here is my solution:


I have bent the thumb of the #26 inward so it more closely grasps, or hugs, the claw of the opposing coupler. There is absolutely no room at all for the opposing coupler to escape, as I have narrowed the clearance gap that before was too wide.

My only question is: the #26 is plastic, and even after I bend the thumb, I fear it can easily retract to its original position just as easily as I bent it out of position. Is there any way I can secure the bend so it remains strong and doesn't gradually drift back to its Kadee factory position?

As for Kadee, does anyone feel they are aware of this problem? Or am I just imagining this problem? For what it's worth, although I have not tested the train on the club layout yet, I have spent considerable time playing with the "renovated" #26 couplers... the bond is secure, no longer will they come apart so easily as they did before. I sincerely hope I have solved this problem. Now, how do I ensure the thumbs don't bend back to their original positions?
Last edited by a moderator:
Let them know of your findings. Include the above photo. You may have found a design weakness they were not aware of.
By just bending the thumb, you have de-formed the plastic still allowing the plastic's memory to try an put it back to the original molded position. Carefully heat the thumb with a soldering iron tip and form it to desired shape and position and see if that works.
In the side by side comparisons you can also see the knuckle on the #5 closes farther than on the #26. The tip of the knuckle is past the top of the trip pin on the 5 and even with the left edge on the 26. An interesting solution though, I have often wished the couplers didn't have so much slack in them.
Hey Switchman,

I had similar problems with other kadee couplers and solved it by greasing them with graphite powder.
Try , maybe this is your problem.
Have a nice day.