Kadee Couplers

[melmerby]

Hi all

 

Without going back through all the posts has anybody succesfully fitted Kadee 18s as a straight swap on Hornby Corridor Gresleys?

 

I have been having a dabble recently (after relying on the Hornby "as supplied" close couplers) however none too successfully

First the pin is way to low, at least 1mm below rail height as supplied so I have had to bend them all up. Nothing unusual I believe.

Secondly there seems to be a diagonal pull on the Hornby close couplling mech as I am having derailments on straight track. What happens is the wheels of the second axle of a leading bogie on a coach starts to climb the rail, then both second and first axles derail.

I can reverse the coach in the rake with the same result.

 

Any suggestions?

 

Cheers

 

Keith

Edited December 2, 2016 by melmerby

[ejstubbs]

Some Kadee couplers come with nylon 2-56 screws. I have found that if you drill & tap the holes for 8BA, then the 2-56's will fit fine. It shouldn't work, but it does. Probably because there is enough give on the nylon.

 

The Kadee #246 tap set - which does result in a proper 2-56 tapped hole - is actually quite reasonably priced.   Gaugemaster list it.

 

(I've also mentioned this point previously on this thread, as it happens!)

Edited December 2, 2016 by ejstubbs

[BR60103]

I just measured a 2-56 bolt and it's 2mm over the threads. NMRA data sheets say .086".

If the original coupling was held by a screw, I try to use that. Moulded on couplings are a headache; by the time there is a mounting surface, there may be no material across the end of the bogie.

I posted some pictures earlier.

http://www.rmweb.co.uk/community/index.php?/topic/56954-replacing-a-Hornby-non-nem-large-tension-lock-coupling-with-something-else/&do=findComment&comment=1923420

Edited December 3, 2016 by BR60103

[Suzie]

...Secondly there seems to be a diagonal pull on the Hornby close couplling mech as I am having derailments on straight track. What happens is the wheels of the second axle of a leading bogie on a coach starts to climb the rail, then both second and first axles derail...

 

This is a problem with the close coupling mechanisms. The answer is not to use Kadees but to use a rigid coupling type like the Fleishmann or the  rigid bar type couplings that restricts the mechanism from moving side to side when it should not.

 

Reducing friction of the wheels, or increasing the strength of the spring in the close coupling mechanism may help.

[brian777999]

There is nothing like actually getting in there and doing it is there ? So I went to Jaycar and their smallest self tapper was far too big so I had to make do with what I had at home. I have a very large selection of M2 socket caps of various lengths. Using a tap drill of 1.6mm I found that the M2X4mm screws will go into the plastic quite nicely and hold firmly. The advantage of using sockets caps is that you can  rest the socket cap on the hex head driver while apply pressure to cut a thread into the plastic ; this is much easier than doing the same thing with slotted or Phillips head screws.

 

After a bit of experimenting I thought I had the correct place for the #20's. I made up two coaches  which looked really good (photo 1). But after pushing them around the test track I found I had buffer lock so I had to redrill the holes in the bogies so that the knuckle was further out. This did run well without buffer lock but the coaches look too far apart for my liking (photos 2 and 3).

 

I am going to experiment tomorrow with the 141 and 146 Kadees. The bogies have had a suitable section cut away to enable the draft boxes to fit neatly on the end. I have glued the draft boxes together tonight to make them easier to handle tomorrow. Do most people glue them together ? I did try just holding them together with the M2X6mm screw which passes right through the hole into the bogie but after it is tightened up then the coupler will not move from side to side.

Edited December 3, 2016 by brian777999

[Dunsignalling]

Hi all

 

Without going back through all the posts has anybody succesfully fitted Kadee 18s as a straight swap on Hornby Corridor Gresleys?

 

I have been having a dabble recently (after relying on the Hornby "as supplied" close couplers) however none too successfully

First the pin is way to low, at least 1mm below rail height as supplied so I have had to bend them all up. Nothing unusual I believe.

Secondly there seems to be a diagonal pull on the Hornby close couplling mech as I am having derailments on straight track. What happens is the wheels of the second axle of a leading bogie on a coach starts to climb the rail, then both second and first axles derail.

I can reverse the coach in the rake with the same result.

 

Any suggestions?

 

Cheers

 

Keith

The trick is only to have Kadees (or the original TL couplers for that matter) on the outer ends of the train - to perform correctly, adjacent CCUs within sets need to be locked together to form a rigid unit.

 

I set up a set of Gresleys for a friend a while back and IIRC the Hornby "close" coupler heads still left a bit of a gap between corridors so I alternated them with the similar but shorter Roco item.  

 

Also, any corridor end covers should be removed for coach-to-coach coupling.  

 

John

Edited December 3, 2016 by Dunsignalling

[melmerby]

This is a problem with the close coupling mechanisms. The answer is not to use Kadees but to use a rigid coupling type like the Fleishmann or the  rigid bar type couplings that restricts the mechanism from moving side to side when it should not.

 

Reducing friction of the wheels, or increasing the strength of the spring in the close coupling mechanism may help.

The wheels are very free running, so no problem there AFAIK

 

The trick is only to have Kadees (or the original TL couplers for that matter) on the outer ends of the train - to perform correctly, adjacent CCUs within sets need to be locked together to form a rigid unit.

 

I set up a set of Gresleys for a friend a while back and IIRC the Hornby "close" coupler heads still left a bit of a gap between corridors so I alternated them with the similar but shorter Roco item.  

 

Also, any corridor end covers should be removed for coach-to-coach coupling.  

 

John

I was trying the #18s as they give nice close coupling without problems on my layout's curves. (mosly 36"-48" with the odd 30") The buffers just rub on curves without locking and are slightly apart on straights - the original Hornby CC is way too wide.

The other advantage the coaches can just be lifted apart.

 

Looks like the Rocos might be the best in-rake option.

 

Cheers

 

Keith

[Dunsignalling]

The wheels are very free running, so no problem there AFAIK

 

I was trying the #18s as they give nice close coupling without problems on my layout's curves. (mosly 36"-48" with the odd 30") The buffers just rub on curves without locking and are slightly apart on straights - the original Hornby CC is way too wide.

The other advantage the coaches can just be lifted apart.

 

Looks like the Rocos might be the best in-rake option.

 

Cheers

 

Keith

If using the Roco/Alt. Hornby heads, they will uncouple manually just by lifting the end of one coach by about a quarter inch.

 

John

[ejstubbs]

ejstubbs : I was going to use the Bachmann bogies but they come with 14mm wheels and I wanted to get the buffers as low as possible.

 

You could always put 12mm wheels on the Bachmann bogies, as you say you have done with the Tri-ang bogies.  My understanding, though, is that 12mm wheels are unprototypically small for coaching stock.

 

I've read (here in fact) that the Tri-ang bogies are a BR type, and thus wholly inappropriate for the clerestory coaches.  OK, the coaches themselves don't actually represent any particular prototype vehicle, but there certainly wouldn't have been BR bogies around in the days when clerestory coaches were being regularly built!  Since my clerestories are supposed to be on an LMS line, I used LMS bogies.

 

FWIW, here are some pics of the Bachmann bogies fitted to my clerestories:

 

 

 

 

I'd be the first to agree that the current Kadee fitting on the Bachmann bogies is rather ugly, but it does me for now.  It's all glued together, by the way, no screws used - Kadee gear boxes take solvent cement very well.  It should be possible to make the fitting less clunky by cutting away the flat area at the front of the bogie (where the Bachmann TLC was screwed on), mounting the gear box further back in the bogie (there's room above the axle) and using one of the Kadee couplings with a longer shaft.

 

The bogie is mounted to the coach using an M5 nylon button head machine screw.  The screw is held in place on the bogie with a nylon nut; you could probably manage without that, which would allow finer adjustment of the coach height on the bogie using nylon washers or thin plastikard.  The screw is secured within the coach body using a nyloc nut, to reduce the risk of it coming undone.  There is a nylon washer between the nyloc nut and coach floor.    Basically, there is no metal-plastic contact in the bogie mounting except at the nylon washer (which is a sacrificial/replaceable item), to reduce the risk of erosion of the plastic bogie chassis and coach floor.  The holes in the coach floor have to be (carefully) drilled out to take the M5 screw.  IIRC the holes in the Bachmann bogies will take the M5 screw without modification.

Edited December 3, 2016 by ejstubbs

[ejstubbs]

Without going back through all the posts has anybody succesfully fitted Kadee 18s as a straight swap on Hornby Corridor Gresleys?

 

I have been having a dabble recently (after relying on the Hornby "as supplied" close couplers) however none too successfully

First the pin is way to low, at least 1mm below rail height as supplied so I have had to bend them all up. Nothing unusual I believe.

 

Sorry, that does not sound usual or right to me.  I use NEM Kadees on my Hornby LMS coaches and the trip pin is exactly the right height with no modification needed.  Photo below of a #18 fitted to my Hornby LMS full brake:

 

Secondly there seems to be a diagonal pull on the Hornby close couplling mech as I am having derailments on straight track. What happens is the wheels of the second axle of a leading bogie on a coach starts to climb the rail, then both second and first axles derail.

 

I would agree that the Kadees are not, out of the box, a good match for the Hornby close-coupling mechanism.  I believe that this is primarily due to the fact that the Kadee coupling head pivots on the end of the NEM shaft.  Note that this is not the same as the knuckle pivot, which obviously has to pivot to allow coupling and uncoupling.  The photo below shows the pivot I mean:

 

 

As Suzie says, the Hornby close-coupling mechanism works best with a fairly rigid link between the coaches.  I have had some success by immobilising the coupling head pivot with a drop of superglue either side (be careful not to get superglue on the knuckle pivot!)  This makes the coupling more rigid, and helps to activate the close-coupling mechanism on curves.   By doing this I can run my coaches coupled together using the longer #19 NEM Kadees round 2nd radius curves without any problems.  On less trainset-like curves the #18 might also work.

Edited December 3, 2016 by ejstubbs

[ejstubbs]

Moulded on couplings are a headache; by the time there is a mounting surface, there may be no material across the end of the bogie.

 

 

I agree this can be a pain.  However, are we not model engineers?  Photos below of Kadees fitted to ancient Airfix bogies:

 

 

 

Cutting away the Airfix TLC coupling mount leaves a yawning gap in the front member of the bogie.  However, the Kadee gear box can be glued in the gap.  A bit of plastikard web is then canopy glued to the bogie pivot cross member, and to the Kadee gear box, to restore structural rigidity/strength.

 

(As it happens, the Kadee ended up a fraction too low with the setup shown in the photos.  With the strengthening web made from a single piece of plastikard, rather than two pieces glued together in a stepped configuration as it was when I took the photos, the coupling sits at the correct height.)

 

By the way, I think this a good reason to use the gear box mounted Kadee couplings, rather then just trying to screw or glue them to what might be left of the bogie.  IMO Kadees don't take well to being mounted rigidly anyway.  They're definitely not designed to be used that way.  (The good old #5s which can be bought without a gear box are intended to be fitted in existing gear boxes on RTR stock, which is a fairly common feature across the pond but more or less never seen in the UK.)

 

I have glued the draft boxes together tonight to make them easier to handle tomorrow. Do most people glue them together ? I did try just holding them together with the M2X6mm screw which passes right through the hole into the bogie but after it is tightened up then the coupler will not move from side to side.

 

Just slacken the screw off a fraction of a turn until the coupling swings freely.  (This is another good reason to use a machine screw in a tapped hole, rather than a self-tapper which can become too loose if you slacken it off after over-tightening.)  I actually find that the whisker couplings benefit from being fitted with the 10thou washer in the gear box - the couplings seem a bit too droopy otherwise.

 

I don't glue my gear boxes together.  If nothing else, it would be extremely inconvenient when experimenting with different coupling lengths & coupler head heights.  And once I've got the right combination, the screw holds it all together perfectly well.  Even if I'm not using a screw because the gear box is glued in place, I find that the clip fit between the gear box and the lid is plenty secure enough for normal running.  If you do contrive to bash the coupling hard enough to pop the lid off then arguably its better for the assembly to break open in a predictable way, rather than ending up damaged so badly that the whole thing needs replacing.

 

(You are aware that the lid of the gear box is a clip fit on to the box itself, aren't you?  If it won't clip in place then you may have it the wrong way up.  But then again, I think if you have the lid the wrong way then the lip on the lid would interfere with the coupling, so that arrangement wouldn't work anyway.)

Edited December 3, 2016 by ejstubbs

[brian777999]

ejstubbs : until I started to put together the draft boxes I was not aware that they are a clip fit. I am aware of that fact now but they are already glued together...oh well, no real problem as this is all an experiment.

 

I do know that the Triang bogies are not correct as are 12mm wheels but as you have pointed out, the coaches themselves do not represent any prototype so these ''mistakes'' do not worry me too much. I was more concerned about getting the buffer height correct as I initially intended to use 3 link couplers.

 

I am hoping that the draft boxes for the 146 couplers will solve the problems I was having with the #20's being screwed directly into the bogie. I agree that the Kadees were not designed to operate this way.

 

If I was doing all this again I would probably use the Bachmann bogies but I have put too much time and energy into converting the original Triang bogies so I am not giving up now   I had to make four brass bushes for each coach on my lathe so that the bogies could be fitted at the correct height. Then I had to turn the axle tips of each Bachmann set of wheels so that it would run smoothly in the Triang bogies. I did not find out until later that the Hornby wheels would have been a perfect fit !

 

Do you drill and tap when using machine screws in plastic ? I do have metric taps but I did not think them necessary when drilling/screwing into 1.5mm thick plastic.

Edited December 3, 2016 by brian777999

[SD45T-2]

If anyone is looking for 8BA screws please look at this link:

 

http://www.canford.co.uk/Products/14-186_8BA-SLOTTED-SCREW-Rnd-MS-BNP-6mm-pack-of-100

 

The usual disclaimer applies.

[ejstubbs]

Do you drill and tap when using machine screws in plastic ? I do have metric taps but I did not think them necessary when drilling/screwing into 1.5mm thick plastic.

 

Yes, I use the Kadee #246 drill and tap set, which I found worked out cheaper than buying drills and a 2-56 tap separately.  If you're using metric threaded screws then of course a metric tap would be preferable.

If I was doing all this again I would probably use the Bachmann bogies but I have put too much time and energy into converting the original Triang bogies so I am not giving up now 

 

I know that feeling!  The Airfix bogies that I converted to Kadees took up far more time than could really have been justified.  And now I've gone and bought myself a rake of the Hornby RTR non-corridor coaches anyway, having finally found some at a non-hair-raising price.  Hey ho...

 

Good luck getting your coaches finally fettled.

Edited December 3, 2016 by ejstubbs

[melmerby]

Sorry, that does not sound usual or right to me.  I use NEM Kadees on my Hornby LMS coaches and the trip pin is exactly the right height with no modification needed.  Photo below of a #18 fitted to my Hornby LMS full brake:

 

 

All I did was remove the Hornby couplings (I was using their version of the Roco type) and plug in the #18.

I did one coach to coach first to try

First piece of pointwork it came to the coach was off the track when the forward facing pin hit the a rail.

Looking at the coupling side on you could see it went below rail top. (as did the trailing one but that just skidded over the rail.)

Bent them both so they were just clear and all went fine.

I then did the next coach to coach with the same problem, more bending required.

Eventually I fitted a rake of 6 all with the pins adjusted and they went around curves fine, however as they went onto a straight the derailments started, second or third coach due to wheels riding up.

Swapping coaches around made no difference, it was always the second or third in the rake of 6 on a straight after negotiating a curve.

 

Keith

[brian777999]

I finally had some luck. The 146's turned out to be too low but the 141's are perfect as you can see from the photos below. The coaches will negotiate radius 3 curves with no buffer lock. I do not have any smaller curves with which to test them but I think they would be okay on radius 1 and radius 2 as well. I had to remove part of the front bogie so that the lip of the draft box seats neatly against the front of the bogie. The white shim is some 0.4mm plastic to keep the draft box level as the front of the bogie is slightly raised. Now I have to do the other 7 coaches but I don't think Kadee produce a bulk pack of 141's.

 

Who makes a nice neat cosmetic hook to put in the holes I previously drilled and cut in the coach ends for 3 link couplers ?

Edited December 4, 2016 by brian777999

[ejstubbs]

Who makes a nice neat cosmetic hook to put in the holes I previously drilled and cut in the coach ends for 3 link couplers ?

 

Lanarkshire Models & Supplies

Dart Castings

Brassmasters

Coast Line Models

 

are ones I know of.  I've used the Lanarkshire ones and found them perfectly fine.  Dave Franks is a nice guy, too!  (He posts on here occasionally.)

Edited December 4, 2016 by ejstubbs

[ejstubbs]

All I did was remove the Hornby couplings (I was using their version of the Roco type) and plug in the #18.

I did one coach to coach first to try

First piece of pointwork it came to the coach was off the track when the forward facing pin hit the a rail.

Looking at the coupling side on you could see it went below rail top. (as did the trailing one but that just skidded over the rail.)

Bent them both so they were just clear and all went fine.

I then did the next coach to coach with the same problem, more bending required.

Eventually I fitted a rake of 6 all with the pins adjusted and they went around curves fine, however as they went onto a straight the derailments started, second or third coach due to wheels riding up.

Swapping coaches around made no difference, it was always the second or third in the rake of 6 on a straight after negotiating a curve.

 

Keith

 

I'd like to apologise if my original response to your query seemed to be questioning what you said was happening with your stock.  When I wrote "that does not sound right" I meant that it sounded as if something was wrong with either your coaches or your couplers, not that your observation of the problem was flawed.  Sorry if you felt that was what I was implying.

 

I'm not aware of there being a known problem with the height of the NEM pocket on the Hornby Gresley coaches.  That said, I've not experienced, or heard of, a whole batch of Kadee couplers having their trip pins too low by as much 1mm either.  Your problem does sound unusual and a little odd.  None of which is likely to be much comfort to you, of course.

 

If you're happy bending the pins yourself then that does sound like a solution.  I'd still be puzzled about the root cause, though.

 

As for the derailments you are experiencing: are you sure that these aren't caused by buffer lock?  The fact that the coaches derail coming off curves on to straight track does sound like that could be a cause.  Wheels "riding up" sounds like the coach is being tipped over; again a fairly typical result if buffer lock is occurring.  It might be that the #18 coupler is a fraction too short to prevent buffer lock (I use the #19 on my Hornby LMS coaches - but then I have 2nd radius curves on my layout).  It might also be worth trying the modification to the Kadee coupler that I described in my previous post on this particular subject - it does seem to help the close coupling mechanism to work more effectively, which can in turn help to avoid buffer lock.

 

Another way to avoid buffer lock which is frequently mentioned on this forum is to run the coaches with the buffers retracted.  Apparently they were actually run like this on the prototype once formed up in to rakes.  I've read that you can achieve this with sprung buffers by slipping a short length of cable insulation over the shank of the buffer behind the buffer beam.  I've not tried this myself so can't vouch for its effectiveness.

Edited December 4, 2016 by ejstubbs

[Dunsignalling]

I'd like to apologise if my original response to your query seemed to be questioning what you said was happening with your stock.  When I wrote "that does not sound right" I meant that it sounded as if something was wrong with either your coaches or your couplers, not that your observation of the problem was flawed.  Sorry if you felt that was what I was implying.

 

I'm not aware of there being a known problem with the height of the NEM pocket on the Hornby Gresley coaches.  That said, I've not experienced, or heard of, a whole batch of Kadee couplers having their trip pins too low by as much 1mm either.  Your problem does sound unusual and a little odd.  None of which is likely to be much comfort to you, of course.

 

If you're happy bending the pins yourself then that does sound like a solution.  I'd still be puzzled about the root cause, though.

 

As for the derailments you are experiencing: are you sure that these aren't caused by buffer lock?  The fact that the coaches derail coming off curves on to straight track does sound like that could be a cause.  Wheels "riding up" sounds like the coach is being tipped over; again a fairly typical result if buffer lock is occurring.  It might be that the #18 coupler is a fraction too short to prevent buffer lock (I use the #19 on my Hornby LMS coaches - but then I have 2nd radius curves on my layout).  It might also be worth trying the modification to the Kadee coupler that I described in my previous post on this particular subject - it does seem to help the close coupling mechanism to work more effectively, which can in turn help to avoid buffer lock.

 

Another way to avoid buffer lock which is frequently mentioned on this forum is to run the coaches with the buffers retracted.  Apparently they were actually run like this on the prototype once formed up in to rakes.  I've read that you can achieve this with sprung buffers by slipping a short length of cable insulation over the shank of the buffer behind the buffer beam.  I've not tried this myself so can't vouch for its effectiveness.

I'd say 2 x #18s are too short in this application. That will be confirmed if they don't couple up easily.

 

The description of the coaches tipping makes me think the derailments stem from adjacent corridor connectors pressing too tightly against one another as the train exits the curve rather than buffer locking.

 

I reckon Kadee 18s give pretty much the same spacing as a pair of Rocos, possibly a tiny bit less.

 

On the Gresley set I dealt with, I needed to alternate 1 x Roco and 1 x Hornby lookalike to make everything behave (I also retracted the buffers).

 

My suggestion, if using all Kadees, would be to alternate 18s and 19s and see what happens. 

 

John

Edited December 4, 2016 by Dunsignalling

[47137]

I tried alternating number 18 and 19 couplers on some Roco ferry vans and this would have been perfect except the close coupling mechanisms let the heads of the number 18s foul the buffers. So I ended up with number 19s throughout. This is with radius 2 curves. It's easier to start with number 20s and see what happens and work down from there.

 

Richard.

Edited December 4, 2016 by 47137

[melmerby]

ejstubbs & Dunsignalling:

 

The problem is definitely not buffer locking. Must check what the corridors are doing though.

On more than 36" radius curves (which most of mine are) the buffers do not overlap completely and just buffer up.

On the curves they derail after they have travelled at least a coach length along the straight before it goes pear-shaped.

It happens even at slow speeds and you can watch it happen.

Note; it is not every curve either! The actual radius seems quite critical and not necessarily the tighter ones.

 

The tightest radii I have anywhere is the inside track of a Peco curved point which is nominally 30" but probably slightly less.

Nothing as low as a number 2 which is about 17"!

 

#18s will couple fairly easily when on a straight, they just need to buffer up gently, but will not couple on any sort of curve.

 

I've just ordered a pack of 50 Roco short couplers to experiment with as well.

 

Keith

[BR60103]

I've 2 wagons this year (out of 2) where the NEM mount let the Kadee sit too low. The trip pin dangles between the rails and catches and the head is too low. Generally, if the trip pin is too low, the whole unit is too low.

 

Extra problems I've had: nut or bolt head strikes bits of the underframe.

Coupler head strikes the buffer beam or even the buffer as it turns.

Cosmetic 3-link coupler dangles into the Kadee.

[Dunsignalling]

ejstubbs & Dunsignalling:

 

The problem is definitely not buffer locking. Must check what the corridors are doing though.

On more than 36" radius curves (which most of mine are) the buffers do not overlap completely and just buffer up.

On the curves they derail after they have travelled at least a coach length along the straight before it goes pear-shaped.

It happens even at slow speeds and you can watch it happen.

Note; it is not every curve either! The actual radius seems quite critical and not necessarily the tighter ones.

 

The tightest radii I have anywhere is the inside track of a Peco curved point which is nominally 30" but probably slightly less.

Nothing as low as a number 2 which is about 17"!

 

#18s will couple fairly easily when on a straight, they just need to buffer up gently, but will not couple on any sort of curve.

 

I've just ordered a pack of 50 Roco short couplers to experiment with as well.

 

Keith

It's also worth checking smooth movement of the bogies and (probably more important) the close coupler linkages - an odd bit of moulding flash in the wrong place can stop everything straightening up cleanly when the vehicle exists the curve.

 

When this happens, everything takes a little longer than it should to realign, and releases with a tiny jerk. One coach destabilizes its neighbour and off we go. The effect is variable hence the point of derailment along the straight probably won't be consistent. Try taking one coach at a time out of the train in turn and see if the problem disappears. If it does, you have your culprit and can sort it out.

 

John

[ejstubbs]

#18s will couple fairly easily when on a straight, they just need to buffer up gently, but will not couple on any sort of curve.

 

I think Kadee themselves state that coupling and uncoupling only really work reliably on a straight section of track.  For example, in this document they say, amongst other things: "It will be a trial and error effort and you'll find many limitations of uncoupling and coupling on a curve."

 

Just a thought: do you have gradients on your layout?  The Kadee coupler operates almost entirely in the horizontal plane; there's not a large amount of accommodation for angular displacement between items of stock in a vertical plane.  If you have gradients then it's vaguely possible that the coupler heads are getting out of vertical alignment sufficiently to cause one coupler head to lever the other one upwards just enough to derail its vehicle.

 

Otherwise, I'm struggling to envisage how Kadee couplers could exert the torque necessary to cause wheels to ride up as you describe, at least without some other parts (ie buffers, corridor connectors etc) of the vehicles concerned interfering.  And that shouldn't be possible on straight track!

 

Dunsignalling's suggestion to check the free movement of the close coupling mechanism on all your coaches does seem like sound advice.

 

Another thought: you say that your coaches buffer up on your curves.  Could they be buffering up to the extent that the buffers get fully compressed?  If the inside buffers are more than fully compressed ie the coach bodies are actually pushing directly on one another, that would exert a tilting force on the coaches which could lead to derailment.  It wouldn't necessarily explain your observation that the derailments happen after the train has exited the curve - although I have seen my own stock do something like this; the initial problem occurs on the curve, but the the curve itself kind of supports the wheel in a semi-derailed condition until the track straightens up, at which point the wheel drops off the rail head altogether.  I'll admit that that doesn't sound like what you describe, though.

Edited December 5, 2016 by ejstubbs

[Suzie]

The issue with the 'close coupling' mechanisms is that when going from curved track to straight track the spring in the mechanism is required to pull the coupling in. This is working against the train trying to pull it out so the coupling does not easily return to the straight position, and there is therefore a side force on the bogie twisting it. This side force is enough to force two of the flanges against the rails and therefore it is likely that at least one rides up on the straight. Adjusting spring tension may help, as may lubrication of the close coupling mechanism (these usually have a lot of sliding face area). The spring tension required will vary depending on where the vehicle is in the train of course with a heavier spring required at the front of the train.

 

If the problem is marginal adjusting the back to backs to the max might help by eliminating the ability of the bogie to turn within the gauge.

 

Kadees should be body mounted rather than bogie mounted (Talgo mounted in Kadee language I think) and this will eliminate the problem (in terms of if the mechanism does not return to straight there will be no turning force on the bogie), but if you want to retain close coupling you will probably have to completely re-engineer the mechanism.

 

This effect can be clearly observed at slow speed when you will see that on returning to straight track the gap does not close up completely and the coupler arm remains to one side rather than straight.