Raildig Store Z Scale Snowshed
Raildig Scratchbuilders Calculator
Ztrains Blog

The Figure-8 Plan Resolved

| August 28, 2013 | 9 Comments More

In our previous article, Inaccuracy In Track Planning?, I looked at why a simple Figure-8 that looked good on the back of a Rokuhan track package couldn’t be replicated to a 100% degree of accuracy using AnyRail track planning software. As I mentioned in that article, I’d write to both Rokuhan and AnyRail and include screenshots to illustrate the “misaligned” situation.

I heard back from both companies today, David from AnyRail said he’d have a look at their pieces for accuracy and immediately after David’s email, I heard from Rokuhan. Excerpts from the Rokuhan email:

You will see the attached image (top image), which explains the case of R270 30° Figure-8. You will see the 55mm track is too long to connect with the 90° crossing and 25mm track is too short. However, both can be used to make the Figure-8. With this understanding, we’ve suggested using the 55mm section in the instruction manual, which will be better for the customers (according to our thoughts).

In the case of a R245 30° Figure-8, the situation is similar as there are two choices: to connect a 25mm section of track or connect nothing. Our conclusion, we’ve suggested to use the 25mm track section to connect them.

John, please understand there are several factors to generate such “inaccuracy” in Z scale track such as the ratio (we use) of the circumference of a circle to its diameter, the nature of the injection material we use, etc.

We selected, by chance, to develop 25mm and 55mm straight track. If you ask for real “accuracy” on Z scale, really PRECISE dimensions would be necessary, such as 18.678mm (just as an example). However, we think pieces like this would be impractical in the current Z scale world.

I’d like using the 25mm sections of straight track in these figures rather than the 55mm sections, but this is just my personal preference. I came away from this email and sketch sent to me from Rokuhan feeling good. I appreciate a manufacturer responding quickly and thoroughly to a concern. I should also point out that I’ve had a Figure-8 set up for a while now and use it for breaking-in locos. Had I not seen the “inaccuracy” on the computer screen, likely I would never have noticed any issue at all.

I can definitely be a bit picky, and in this case I’m glad I was as I now have a better understanding of the wiggle room that’s sometimes required to make things work in the real world, away from the computer screen.

I’ll also contact David from AnyRail with link to this article so he’ll also know all is right in Z scale world tonight.

Tags: ,

Category: Blog

Raildig Newsletter
  • ViperBugloss

    Recently, I have been considering the same problem. If you draw a circle of radius 220 mm and construct two tangents at right angles then the length of these tangents will be 220 mm. Since the R025 Crossing has a length of 25 mm, the straitght piece to the crossing should be (220 – 25/2) mm, i.e 207.5 mm. The R025 Crossing 90 degree instruction sheet must assume that there is sufficient play in the track to ignore this difference, whereas AnyRail will not join up the track. However, by breaking the 220 mm curved section diagonally opposite the R025 crossing and including a R030 25 mm straight section will allow the track to join up with AnyRail. Another solution would be to make up a 207.5 mm straight using a R001 110 mm and a R031 adjustable track, although this is probably an unnecessary complication and expense.

    The R025 Crossing works well with the R120 curves as this requires a 107.5 mm straight, so using a 110 mm straight just requires 2.5 mm of “fiddle”.

    • Ztrains

      It’s funny but this whole issue came about as I wanted to post up some track plans. I assumed they’d still work with the necessary fudging but the fact that I had the broken track joins made the effort look a bit sloppy onscreen though in practice, the layouts were fine.

      As a very practical test, I’d guess that if you asked 10 people to build a 3′ long oval of track and were somehow manage to overlay all 10 on to each other, no two would be dead-on the same. The wiggle and shifting with sectional track is a much a part of the track as the track itself.

      What I normally do in practice is to relax the “snap” distance in the software a bit. I do like the idea you put forward here and I’m going to give it a try!


  • ViperBugloss

    I agree. AnyRail can look very sloppy whereas in practice the fiddling can take be extended much greater distances. I checked the AnyRail tolerance that I was using and it was 10 mm which is a little excessive but again doable in practice. Since the fit is good using the 120 mm radius curve, this means that adding 25 mm to the curve radius can be accommodated by adding 25 mm to the straight section. However, a figure-of-8 built using a curve radius of 270 mm would require a straight section of one 110 mm piece and six 25 mm pieces, so in total that would mean 24 R030 25 mm pieces, resulting in a global shortage of R030s.

    • Ztrains

      Global shortage of R030s… LOL.

  • David

    Thanks for the research and your clear blog post.

    Perhaps it’s good to know that in Anyrail it’s possible to simulate
    the ‘wiggle room’ required in many cases when working with rigid track
    instead of flex track.

    To do so, switch off ‘Autoconnect’ in the ‘Settings’ tab.

    Now, if the tolerances in the same ‘Settings’ tab are met, track
    will connect as you position it instead of jumping into the 100%
    correct position. Often it helps to connect track at a slight angle of
    1 or 2 degrees to make suggested designs fit.

    Unfortunately the side effect of a software program is that any
    time something doesn’t fit in theory, this shows immediately.

    • Ztrains

      Good tip David, thanks. I just played a bit with this suggestion and it gave a very good result. I’ve now learned there were good reasons why the initial Figure-8 appeared the way it did.

      As long as I know there’s a good reason for something, what that reason is, and what effect it may (or may not) have… I’m happy with that.

  • Minuteman

    Interesting input, especially as I have just come away from the the “golden spike ceremony” on my new 2′x4’2″ layout which incorporates a continuous run which is circular for 180 degrees at one end and elliptical at the other. The idea was to break the evident tailchaser (aided further by hidden track), Now here’s the gag: it worked out perfectly in AnyRail, and it worked out perfectly 1:1 with Rokuhan track – except I used one piece of R030 25mm that wasn’t needed according to my 1:1 scale printout…

    I see the point with tolerances, and in my view the important element is that Rokuhan offers us the flexibility of having 25mm and 55mm track pieces, plus flextrack and the adjustable track. For me, it turned out to be a solve everything solution, and trains are running fine.

    • Ztrains

      Thanks for sharing this. It does make the point that as accurate as AnyRail may be, there’s always going to be some wiggle room with sectional track.

      I do that that 25mm piece in particular is a great length of track to make things work.

  • Pierre

    I am now totally confused. I bought the track you originally suggested using two 110mm to connect the curves to the crossing track. I have used it to break in a Marklin 88036 NYC steam locomotive, a Marklin 88591 Re 4/4 II SBB and a Marklin 88592 Re 4/4 II SBB Cargo and I had no problems. I ran all three for 30 minutes and I even added a few cars for a little fun. No problems.
    Now when it comes to my AZL PRR USR Light Mikado I could not get this beautiful locomotive to tun the curves. It ran fine on the straights but immediately derailed once in the curve. Have I caused the problem by using two110s without a 55mm? I hope that the fix is as easy as that. Thanks, Pierre

Raildig Store