Due to heavy spamming attempts on this forum, automatic registration has been disabled. We will approve registration requests as quickly as possible (unless you're a spammer of course :) )

Main Menu

Understanding Tolerances and Autoconnect

Started by dlanciani, April 10, 2010, 11:05:19 AM

Previous topic - Next topic


I'm trying to understand how these settings work/interact.

First, it appears that in addition to controlling the actual tolerance of a connection, the Tolerance
setting also affects the "capture" range for an autoconnection by (I'm guessing) some multiplier.
For example, with the Distance Tolerance set to its maximum of 25/32" I can drag a piece of track
within slightly more than 2" of another and it autoconnects.  With Distance set to its minimum of
1/32" a piece of track must be positioned within about 1/8" (which can be difficult at lower
magnifications) before it will autoconnect.  Am I understanding this correctly?

The Track: Autoconnect check box is documented to control whether track is automatically
connected when endpoints are close enough, but this does not appear to be its function.
With Track: Autoconnect unchecked track is still autoconnected, but it is not positioned
automatically.  Instead it remains where manually positioned.  Does Track: Autoconnect
actually control whether an unconstrained piece of track is positioned "perfectly" when

With Track: Autoconnect checked is there any gap inserted at a connection point?  I notice
that 4 segments of Lionel 65038 (nominally 8 3/4"/222mm) together are longer than one 65024
(nominally 35"/889mm).

Do Tolerances affect only track connected while they are in effect?  That is, if I increase the
Tolerances to insert the last piece of track that closes a loop how is the "slop" distributed?
Is it divided between the two endpoints of that last track?  I assume that other connections
throughout the layout that could absorb a discrepancy are not altered.

Once an "imperfect" connection is made to close a loop does it remain fixed if the loop is opened
elsewhere?  If so, is there some way to highlight connections with "slop" so the error does not
accumulate as the layout is modified?


You are mostly correct:

When autoconnect is on, the 'distance' of the tolerances settings is used to determine whether or not another endpoint is close enough. If it is close enough, the track is rotated and shifted to make a perfect fit. When that is done, the other endpoints of the track are checked one by one, to see if it should connect to anything else. For that check, bot the 'distance' and 'angle' of the tolerances are used. This could result in connections that are, at least theoretically, not 100% correct.

When autoconnect is off, both 'distance' and 'angle' are used to verify if track fits. If it fits, it's connected without rotating or shifting the track. When done, the other endpoints of the track are checked and connected as described above. This function is useful if you need to span that last gap, just like you would use the 'wiggle room' of the real track.

When connecting track, all the other track in the layout is left as it is. Only the part that you connect is involved, and existing connections remain untouched.

There is no way (yet) to indicate and find the 'sloppy' connections.

Your issue with the Lionel track may be caused by an error on our part. We've bought and measured all the pieces, but of course we only have one of each. We measured the 65038 (and the turnout 65021) to have a length of 224mm instead of the 222mm in the catalogue. There might be differences between batches.

I hope this helps!


David Hoogvorst. Founder and Owner of DRail Software. Creator of AnyRail.


I measured some actual pieces and I agree that the 65038 is 224mm; however, the 65024 is 896mm so it
is (I believe) intended to be exactly the same length as 4 65038s.  I also measured the 65019 at 123mm,
an O27 turnout at 225mm, and an O42 turnout at 347mm.  It's entirely possible I'm off by 1mm in either

Am I understanding correctly that the length shown when I mouse over the track selection (222mm in
the case of the 65038) is not the length used internally by the program (presumably the 224mm you
measured)?  That can be confusing.

It would be nice to have a mode where only "perfect" connections are made automatically but the "capture"
distance is large enough to make placing unconstrained track easy.  Imperfect connections could be made
manually after being shown the x & y offsets and angle.  It might have to ask which connection was to
be made if two or more were within capture distance but only one could be perfect.


Thanks for the measurements. They are identical to what's defined in the software, except for the 65024. We will update that one.

The problem with most track systems is that the dimensions as advertised don't match the real thing. Sometimes it's only a few tenths of a millimeter, sometimes it's more.

Of course, we use the exact measurements in the track definitions, but when the descriptions would not match the ones used in the catalogues, it would cause confusion too!

To keep it simple, we've decided to always use the description from the catalogue, but use the real measurements in the track definitions.

When autoconnect is on, the capture distance is 4x the tolerated distance. When it is off, the capture distance is exactly the tolerated distance.

We try to keep all this away from the user as much as possible, as it's complicated and not crucial in practice. However, maybe connecting the last gap is still somewhat more difficult than it should be.


David Hoogvorst. Founder and Owner of DRail Software. Creator of AnyRail.


Quote from: David on April 11, 2010, 11:55:27 AM
However, maybe connecting the last gap is still somewhat more difficult than it should be.

Thank goodness for flex track! :D
Later,                                                AnyRail Fanatic
Jeff                      and Unofficial Guy Who Knows Almost Everything About It


Funny thing.  I realized I had a genuine Lionel carton of 65024s.  The carton actually says 35.25"
in the description.  Surprisingly it was not written as 35 1/4".

I think the real problem with Lionel track is that the rails are not always well justified.  So at each
connection point one rail will touch and the others will have a little gap.  At the other end of that
track a different rail will touch.  This makes the effective length of the track slightly longer.

Maybe I worry too much about these things.  My main interest in a layout program was to see how
much I'm abusing the geometry. :)