NMRA S-4.2 Questions

[narrowminded]

This has been bugging me for a little while now.  Nn3 runs on Z track, yes?  Then why does the Nn3 back to back spec carry a measurably different dimension from the Z back to back spec in S-4.2?  http://www.nmra.org/sites/default/files/standards/sandrp/pdf/S-4.2%202015.01.19.pdf

Example:
"B" Target dimension (back to back).  Nn3 is .208", while Z is .214".  The plus tolerance remains at .002" for both while the minus tolerance is .002" for Nn3  and  .007" for Z.  The width for Nn3 is less by a real amount yet would fall within the much more generous tolerances of Z specs.  What it amounts to is Nn3 is at  .210"/ .206"  and Z is at  .216"/ .207".   It would seem they should be the same as they're intended to run on the same track.  Anybody know? Or absent that, maybe, could we collectively make up a good story?    C'mon, I know we could do it.  I've seen it happen.

Seriously, I'm setting specs and tolerances for wheels right now and am inclined to stick with the Z specs as this will be used on Z track and will also be used on Z chassis offerings.  I won't need the broadest tolerance allowed by Z specs but am concerned that it will be sloppy if utilizing the Nn3 spec, even when used in an Nn3 application.  Afterall, it all should run on Z track, yes?    My inclination is to set the back to back target dimension at .212" with +/-.002" tolerance.  Or .214"/ .210".  Any naysayers, please speak up.

It is intended that the flanges will all be RP-25 and there is no discrepancy in the specs between Nn3 or Z flanges so that's easy.  Tread width is still up in the air with a good possibility that there will be a standard offering, slightly wide, and then a fine or semi-fine option.  I think I should start a thread for opinions/ input on that but really need to decide this back to back, "B" dimension.

[nkalanaga]

I think the difference is that Marklin's standards ARE very sloppy, while the NMRA felt that many American narrow gauge modelers would be laying their own track.  Thus, while the narrow gauge standards are compatible with Marklin's Z scale, they don't allow as much slop.   

[narrowminded]

Thanks.  I see that and the same thought crossed my mind BUT... (and this is where I start going sideways) since they are all supposed to run on any NMRA speced Z track, and the Z back spacing spec is maxed at .216", and should run well on that, and then the Nn3 spec is maxed at .210" which suggests to me that it's now at least .006" looser fit.  Figuring that the track must accept the max spec width of .216", then add the track tolerance and it's even more.  I can make any of those numbers, holding +/-.002" (which will be more like +/- .001") and really just want nice performance, hence this post.  I'm leaning, as I said, to a target of .212" then allowing +/-.002" maximum and think that's tight enough to track reliably and yet not be too loose between the rails.  Maybe I should target .210" +/- .002".  This is where I am stuck a little.

My own testing included switch brands Micro-Trains and Rokuhan and it ran well through both of those brands as long as the flange profile was right (RP-25).  An N profile just regauged for Z track was EXTREMELY sensitive, especially in the Rokuhan switch.  It could be made to work but only within about .002" back spacing.  It proved out that the numbers do matter as the RP-25 flange thickness difference between N and Z is only .006".  Add the .002" that was the difference that seemed to work, and all of a sudden it seems that the back spacing will also matter and the narrow end of the spec seems like it might be too much, especially if allowance for the track work is considered.  This is how I'm getting to what I think will be a good, nice running number, at the .212" target but am still very confused by the spec variation.  I'd feel better knowing what the reasoning was because it sure doesn't seem to make any sense.

[narrowminded]

OK, I think I see it when I look at the NMRA S-3.2 track specs linked here. It still doesn't strike me as a necessary or even a genuinely helpful spec, holding +/- .002" while hand soldering two individual joints/ rails to PCB ties, but it's there.  I wonder if any hand laid track actually holds that tolerance across the board, even working with decent fixtures.   I'm not a gambler but I'd take bets on this.   And the good news is, it doesn't have to.

http://www.nmra.org/sites/default/files/standards/sandrp/pdf/s-3.2_2010.05.08.pdf

The manufactured Z track spec is a little looser yet has the better chance of actually holding those dimensions/ tolerances as a function of good tooling in a controlled manufacturing environment.  But that's just a guess and I can easily work within it.  As I look these over and on the assumption that most applications I might have to deal with will be actual Z track, I think I'll go with my .212" target with the +/-.002" max tolerance.  And if someone actually had Nn3 track, AND that was at the max tight end of the .002" spec, AND the wheelset was at ITS max tolerance, it should still run, albeit with near bearing fit clearance before using the flange root radius to buy some more room.  And worst case, in my assembly arrangement it would be easy to remove .002" from a stub axle to drop the desired amount and only needed IF the track was at the extreme end of the NMRA Nn3 tolerance AND the wheels were at the extremes of my chosen tolerances, AND in the flange binding direction and THEN, actually giving troubles.  Z track, which is what I think is the volume, should be fine with these chosen dimensions that are tighter than required while still comfortably within the Z scale specs and affords a generous tolerance coming and going.  It should run very nicely, all day, every day.  

Thanks for listening as I think out loud.   Add anything at all if you think it might be helpful... or just want to do a little out loud thinking of your own.

[nkalanaga]

My biggest problem was on dual gauge track, where the common rail flangeways had to handle both standards.  Some of those were a little tricky, but it can be made to work. 

[narrowminded]

I subscribe to the never say never school of thought but I can say, as close as I can ever get to saying never, it would be that I will not be doing dual gauge track.

When you did that was it N scale with the third rail inside?  That I could see doing as the flanges in Z are narrower and a decent tread width would prevent the wheel from dropping into the frog when there's a little extra side play.  Now, fine scale tread?  That's the second one in a few sentences that gets close to saying never.

[nkalanaga]

Yes, American dual gauge, standard and "3 foot" gauge.  The biggest pain wasn't the frogs and guardrails, but the standard gauge points.  Using code 55 rail there isn't really room for the points to move if one uses solid rail, or room for a rail joiner for pivoting points.  I used solid rail, and filed a notch in both sides of the base and the outside of the head, so that the web acts as a "solid pivot".  It works, but I'd just as soon not have to build any more.  #8s are about the smallest practical size in code 55.


All of the turnouts have at least one side of a frog in the common rail, so they have to accept the narrower Nn3/Z wheels.  On the other hand, some N scale wheels are no wider than the Marklin Z wheels, so that wasn't a real problem.

In code 40, #6s work fine, and there's quite a bit more room for the points.  They're relatively easy, but I wanted the larger rail for my standard gauge mainline, so that I could run everything I had, even if the flanges were a little large.

[narrowminded]

There's a special place in heaven for folks who can build a dual gauge switch.

Maybe a better way to ask the wheel gauging question would be, is anybody using anything but an NMRA Z gauge tool when checking their Nn3 equipment?  If not......

[VonRyan]

There's a special place in heaven for folks who can build a dual gauge switch.

Maybe a better way to ask the wheel gauging question would be, is anybody using anything but an NMRA Z gauge tool when checking their Nn3 equipment?  If not......

I just use the handlaid track on my layout to check if wheelsets are ok. As long as the trains run without issue, all is well in my book.

[narrowminded]

I just use the handlaid track on my layout to check if wheelsets are ok. As long as the trains run without issue, all is well in my book.

And that's what I would suspect most do and especially through switches.  And I'm sure that the dimensions aren't all over the place but also suspect it isn't +/-.002" everywhere.  And that's also not to be disparaging of the amazingly nice work that hand track layers do, like dual gauge switches.   And from a manufacturing standpoint, that's what tolerances are for.  You have the target number, the ideal, and then you have the tolerance which should represent the dimension beyond which the part would not work acceptably.  That becomes a reject or one for rework.  I think I'll be OK with the numbers I'm choosing and may in fact choose the target number at .210" instead of .212".  That combined with a wider (wideish) tread should be a pretty forgiving general design... I think. 

The thinking is that catching points is a common trouble spot and it would be affected by the backspacing on the wheels and the flange ways in the track work. And then the tread width becomes an issue as the backspacing becomes narrower or flange ways become wider. By making the backspacing narrower it could handle a loose switch and flange way, not be prone to catching the point when running in that direction, but then it would have a propensity to drop wheels in the frog or bind between the flangeways if taken too far.  That's where a wider tread width comes in and it's those relationships that I'm trying to understand and then address in the parts I'll be making.  Striving for a good balance between function, appearance, and reliability all while accommodating the track work that may not be perfect.  Another way of thinking of that would be not putting all of the burden for tolerance on the part(s) that is actually the most susceptible to tolerance variation and stack-up, the track work.  And while I'll be testing these and HAVE tested pretty extensively already, my sampling of switches will never be in the sufficient quantities or randomness that I would like to qualify as a real and complete test. 

I'll keep working on it and would welcome any remarks from those whose experience is broader and more varied than my own.   

[nkalanaga]

I use a brass gauge I bought years ago from Republic Locomotive Works, and some wooden ones I made myself.  The wood works a lot better for soldering.

For flangeways I usually use my N scale NMRA gauge, because I do have a few narrow gauge cars with regauged N wheels.  That seems to work fine with the Z stuff, probably because the Marklin standards are so loose.

For straight track Z will run on 0.250 inch gauge, so for the little truly straight track I have I used a piece of 1/4 inch stripwood for a gauge.

[narrowminded]

Have you ever had occasion to run the re-gauged N wheels through standard manufactured Z switches?  That would be an indicator as the N flanges, if made to RP-25 standards, are .006" wider than Z, or combined, .012".  That's where I experienced extreme sensitivity in MT or Rokuhan switches, running re-gauged N wheels.  There was no derailing but anything outside of a very narrow window (+/-.002") there was a slight "bump" as it passed through the flange ways accompanied with a visible slowing, most evident when creeping really slow.  Worst case it would jack up pretty good and finally, when bad enough, would stall the thing.  By playing with wheel spacing you could find that sweet spot that ran smoothly.  Anything outside that exhibited some level of the binding described  and that would not be acceptable. 

But with the correct Z flanges and the extra .012" (combined) that that affords it should run fine and in fact, when you look at the specs for those parts it becomes evident that with the dimensions outlined, the actions experienced make perfect sense and go a long way (at least for me) to understanding what's happening, underscored by that unintentional but useful test.  Arriving at the same end but from a totally different direction. It affords another level of confidence in the set of measured dimensions that will work. 

The tread width is the only one left that I'll digest some more but I do think I'll be adding a little bit of width as that will hide a host of track issues.  And finally, if THAT'S not enough, I can, with a very clear conscience, suggest that the track needs some fixin.    And it also makes the NMRA specs understandable in their practical application.  Except for, why the difference from Nn3 and Z?  The fact that Z has a target dimension as well as looser tolerances doesn't REQUIRE that someone use those tolerances.  A self imposed tolerance within a window, seeking a quality above the norm, is always OK without making it another whole set of standards.  In fact, I'm doing just that with my dimension choices.  IMHO it goes overboard and until I went through this exercise it added a whole level of unnecessary confusion.  But better that than nothing!

[nkalanaga]

They run fine through Marklin turnouts, but those are the only commercial Z turnouts I have.