Unitrack Derailments

[freescopesdad]

I have experienced derailments of 2 locomotives (Atlas GP38 and BLI 2-8-2) at one specific location. I have traced the issue to 2 pieces of Unitrack whose joint is slightly uneven vertically. Filing the track ends slightly at a 45 degree angle before installation was one modeler's solution. However my track is already glued down and the area has been finished. Is there a way to achieve the smoothing of this problem joint without removing track? Open to any and all suggestions. Ron

[learmoia]

What if you took a square file and went at the joint with one of the 90 deg corners of the file?

V
1

Should yield 45 degrees on each side of the joint..

[freescopesdad]

I'm going to practice with a couple of spare pieces of Unitrack, maybe start with a nail file or a piece of emery board.

[StewRRFan]

Make sure the connectors are pushed down tight.   I have had issues when not all the way down with derailments.  Good luck

[peteski]

If nothing else works take a flat smooth-cut file and gently file down the the raised railhead down. Lay the file down across both rails and don't file perpendicular perpendicular to the rail. Do it at an angle, and don't use too much pressure.  Nickel silver is a fairly soft brass alloy so it should not take much to file it down.  If you are worried about the the other adjacent rail, stick some masking tape over that rail.  Smooth-cut file is not very aggressive - it should not cause any damage.

[Maletrain]

The actual Unitrack metal rail joiners, contained in the plastic roadbed snap-together parts, often get bent open a bit when the track sections are snapped apart repeatedly as we experiment with track.  That can let the rails get slightly our of alignment. 

I suggest that you first see if the high rail can be pushed down flush with the low rail.  If so, you can probably slightly crush the metal rail joiner to get the alignment right.  I like a 2.5 mm wide straight blade screwdriver bit to do that.

If it won't move, or you can't make it stay down, then filing it is your only solution unless you remove the track sections and replace the rail joiners.

If you do file it, and the rail moves vertically, then file along the rail from the low side toward the high side, being careful not to file the low side lower in the process.

I am a little surprised that you get derailments on a vertical step in the rail.  Is it on a curve?  If not, look for some sideways misalignment that the wheel flanges can pick.

[djconway]

Seeing that you are derailing at a vertical step,  the first thing I would check is that the higher rail is fully in the rail joiner.  I know its Unitrack, but after repeated joining and separations the joiners can work loose.

[arbomambo]

Diamond sharpening stone…
Comes in various sizes, various grits…
Use it like a bright boy; it will shim your railheads to the same height with no ‘step’ and, rather than producing pits in the top of the rail, will actually polish it.

[peteski]

The other think which might be worth trying is apply some pressure to the high end of the rail.  Kato Joiners are very deep (their base is way below the bottom of the rail), so maybe the rail can be slightly bent down to make it even with the other rail.

[robert3985]

The logic goes like this:

(1) - If every loco derails at the same location, it's the fault of the track

(2) - If only one loco derails at a location, but all the others run okay, it's the fault of that loco

(3) - If a couple of locos derail at the same location, but all the others run okay, then it's the fault of the locos

(4) - Modifying track so one or two of your locos run okay, while all the others already run okay, is counter-intuitive...fix the offending locos, don't modify the track.

Soooo...using that logic, if all the rest of your engines run across that spot without problems (at least not de-railing) then, if it were me, I would check to make sure the locos were (a) in gauge (b) flanges are not bent (c) wheels don't have build-up of track junk/oil/cat hair etc. (d) truck rotation is smooth and unobstructed (e) coupler dongle is at the correct height (f) nothing on the underbody is extra low and catching on the uneven track joint

If you check your two engines and there's nothing operationally different about them compared to all your other engines, then I might consider filing the joint with a large, sharp bastard file to get the rails even at the joint...or just compressing them with flat-nosed pliers at the joint to force them to be even...or more even. You can also do the flat-nose-plier trick for side-to-side alignment.

I am not of the opinion that rail needs to be polished to run reliably. Every turnout I build gets at least a few strokes from my large bastard file before installing to ensure the railheads are all level, and I've never had a pick-up problem or dirty track, or accumulation of anything that prevents them conducting electricity reliably...in over 30 years.

That's my two-bits worth...

GOOD LUCK!

Cheerio!
Bob Gilmore

[dem34]

Yeah. I remember going nuts once trying to fix an Atlas #7 (6) so an ancient Kato SD40 would go through properly. Only to find afterwards that while the front truck wheel gauge was good. The rear truck was a whole mil off and of course, was going to pick every time.

[randgust]

My primary experience with Unitrack has been on the big Ttrak setups at Altoona.   We'll have over 100 module units there every year.

There's consistent problems with a couple things -  kato #4's on picking the points (where the switches need the point socket filed in) on junctions and vertical head variation between modules.  You're completely dependent on those joiners on module-to-module connection, and 'your mileage may vary' on the quality of track and construction from modeler to modeler.

For the most part, straight track to straight track will work even when pretty badly out of vertical alignment, just look awful when a train bangs across it, but do that same stunt on a curve to straight connection and you're on the ground.  There really aren't easements and always a joint there.  And there can be a lot of vertical slop despite Kato's joiner design.   I'm the 'track inspector' a lot of times, I want to feel NOTHING on 'curve to straight' joint and will keep tweaking table height and shimming until its right.  it's a 'run your finger' through the rail joint and if you can feel anything,  chances are it needs adjusted.

You can also get vertical peaking at a joint due to irregular subroadbed where it feels ok but isn't, check with a straightedge or ruler. 

The vertical slop factor can be greater than flange depth, too, so be aware that you may be perfectly in gauge, but one locomotive piles up and another is just fine, and you can't spot it easily.  But six-wheel trucks and long steam locomotives with lower flanges will have fits where everything else works just fine.   I had a miserable problem on my own modules with a mild 'hump' at a module joint where a long-wheelbase truck or steam locomotive would randomly derail, finally had to tear it down to the wood to eliminate it.   You at Altoona know EXACTLY where that was... drove me nuts.

[freescopesdad]

Thanks all for the suggestions. The offending track is on a 14" radius curve where one piece of track is glued solidly to the foam, while the adjacent section has some vertical play. I am going to put some caulk under that track and weight it down overnight to see if that solves the problem.

[Maletrain]

I am going to quibble a bit with "Robert3985's rules" for diagnosing derailment causes:

On one of our club layouts, my pair of Scale Trains Dash-9s derails at the same location, too frequently but not always.  All other locos pass through that location repeatedly without derailing.

Checking the two derailing locos' wheels, they are properly in-gage and do not have any detectable flaws in their abilities to roll, pivot, etc.  However, their treads are narrow, just fitting into the "NoGo" slot on the NMRA gage.

Checking the track gage shows that it is substantially too narrow - by about the width of the code 80 railhead.  And that is a lot!  (This is flex track, not Kato Unitrack, so I am probably hijacking this thread a bit.)
 
Comparing the locos that derail with those that don't, all I see is the thinner wheels on the Dash-9s.

It is not clear to me that the thinner wheels on the Dash-9s are the main problem, although it does somehow seem that thicker wheels are better able to mitigate the too-narrow track gage condition.  Those Dash-9 wheels go through other parts of our problematic trackwork without any problems, even where other locos and rolling stock do have problems.

Where the Dash-9 locos derail is on the upper part of a down-sloped curve, and the derails are to the outside of the curve, not far after the track gage becomes too narrow just after the beginning of the down slope.  I am thinking that the flange to the outside of the curve is riding on the outside railhead and the second loco pushes it to the outside of the curve and over the railhead if there is any glitch in the pickup of the lead loco, so the derailments are not a sure thing, and are not predictable.  The same train my go through many times, then start to derail frequently, but still not every time.  The wheel tread should not be off the rail on the inside of the curve before the flange on the outside of the curve is out past the railhead.

So, my take on my situation is that I should fix the track and see if that mitigates the problem, before deciding to tackle the loco wheel replacement to get wider treads (somehow).

The previous coordinator for this layout used guard rails on the side of the track away from the direction of derailments.  There are many, and some of them are quite long.  They do seem to have solved the problems in those locations.  So, along with trying to file the inside of the railhead a bit on the inside rail of the curve when the track gage is too narrow, I plan to resort to trying guard rails before giving up on Scale Trains or trying to find wider replacements for their drive wheels.

Unless somebody here with more experience has a better idea.   

[mmagliaro]

If everything is glued down, this is actually a "permanent" installation that happens to use Unitrack, rather than an actual "portable" installation like modules for train shows.  You could just push the offending higher rail down and then just solder the joint while you hold it in place, so it will never move out of alignment again.