SCARM Out-and-back Fastrack help?

Instead of the 5", try a 4.5" and a 1.375". Worked for me.

I missed that because SCARM complains. Is that a SCARM issue or a real track issue?

Great  timely post.  My reversing loops on our Christmas loop worked great until last weekend.  The loops are in tunnels and are starting to separate on a regular basis.  Had to disassemble 2 of the 3 levels to get to it and make repairs.

The first loop is 036 and the second loop is 048.  Would we use the same size "short" pieces of track for both loops or does the 048 need something different?

Thanks!

SCARM didn't complain at all for me. I first played with it in RR-Track and came up with a negligible .08" error. However, when I redid it in SCARM, I didn't get any complaint. You did use the O36 11.25 curve extension after the switch, right?

SCARM file attached.

Initially I thought since the library has 15 and 7.5 small pieces for O48, you'd have to use one of those with the O48 switch. However, when I put it together in SCARM, I found that standard O48 curves to the turnout of the switch works with 1.75", 1.75", 4.5", 5.0", 10.0" and 10.0" to the rest of the curve.

SCARM file attached.

Slightly cryptic notation.  Also note that I used the 6" long 90deg crossing to show that the "perfect" length in a couple places would be six inches.  Alongside, I show some approximations to that six inches.

I think the best setup is the third one with the 1/4 turn followed by the 1/2 turn after the turnout. It adds very little extra length, and eliminates the S-turn without too many specialty lengths.

  --Joe

Joe, I totally agree with using a configuration that would eliminate the "S" curve. I suppose I shouldn't have just taken the initial question at face value, so kudos to you for going the extra mile.

Dave R.,

It's not SCARM, it's the track geometry. The 036 switch with 036 curves requires a 6" custom made straight piece (a 10 + 10 + 6) or 12 -1 3/4" pieces and a 5".

No matter how you slice it, there's 26" of straight to complete the reversing loop.

Why won't you make 6" piece? Custom cutting track is not unreasonable.

SCARM file attached.

Santa Fe Va,

When you force the fit FasTrack will separate.

The 048 reverse loop has a 33" straight made up of 2-10", 1-5", 1-4.5", and 2 1.75".

SCARM file attached 

Okay, I forgot I changed the edit tolerance settings from 2/2 to 4/4 for something else, so that's why my example in SCARM using a 10, 10, 4.5 and 1.375 works, sorry.

But, are you saying a FasTrack loop on the floor or a table can't stretch 1/8" and not come apart without being screwed down?

This is my S-curved reverse loop.  It does show a slight gap in SCARM but has worked for me for 5 years now with no screws.

4.5 + 1.375

Ed

Daz,

I've only found that too long creates a separation problem. eddiem has no problem with slightly shorter.

Here's one that has no errors and meets the stated requirements of limited small pieces. As you can see, I used a 10", 10", 5" and 4.5" along the straight side and then added a 1.75" and 1.75" to the loop at the bottom. I would still add a 5" to both sides of the loop to at least minimize the "S" curve (a 10" would be best), but I don't know the space limitations.

Thanks guys!   Everything worked for about a month but lately it has been separating much too frequently.   I need to pick up some smaller pieces and give it a try.

That I know from my HO and N Scale days. I'd just be surprised if 1/8" was "too" long.

Dave,

i don't get how your latest loop fits... You added 3.625" on the top and 3.5" on the bottom...doesn't make sense to me, but it looks ok!

ed

Actually I added 3.5 to the top too. Remember, even though my first attempt appeared to work, it was really still off by .125" because I had changed the SCARM tolerance settings from 2/2 to 4/4. Once I changed the settings back to default, the .125" error popped up. Since the straight section was really 26" and mine is now 29.5", I only added 3.5" to the "actual" length of the space. The 3.625" you came up with was because you used my original length of only 25.875.

At any rate, since there are no standard pieces that fit together to fill a 26" space, I thought maybe increasing that space might let me find pieces I could use to stretch the loop to make it all fit. So, I tried increasing the space to 30" and saw it was just .5" over the 3.5" I could expand the loop using the 1.75" pieces. And there you have it.

Thanks to all who provided input. It's fascinating to see multiple (and widely varied) solutions to the same seemingly simplistic goal.

I loaded the most recent configuration from Dave (DoubleDAZ), and with the default SCARM tolerances, and it doesn't bark at me.

I'm really not that opposed to cutting track (I have the ability and tools), but there's always a waste product, and I'm OCD about that. If forced to cut, I'd grab one of my spare 30" pieces and cut that (since they are spares).

Regarding the "S" curve issue, absolutely yes I will be adding 10" track; a switch on one side for an inside siding (which forces to match the other side obviously). I've got one steam engine that really doesn't like those quick twists.

Huh, that last smiley kinda looks like me! But anyways, thank you folks!!!

Hello all,

Just one note about SCARM default tolerance setting.

It is 2mm (0.079") by default and is intended for smaller scales, such as HO and N. In larger scales, like O gauge, it is a good idea this value to be increased to 0.15" as such misalignment is normal and easily compensated at the joins when the real tracks are connected on the floor or over the benchwork.

I am explaining more about these issues here: Tracks Do Not Match/Connect As Shown in a Given Layout Plan.

Mixy

Thanks Mixy, I've often wondered about this. I have my tolerance in RR-Track set to .20", the max allowed. I don't know if FasTrack and RealTrax are as forgiving as regular track without roadbed attached, but I do think they would forgive 1/8" (.125"), so changing the 2 to 15 seems reasonable to me. I don't know if 20 is too much, but I'm not too worried because I won't be using tight reversing loops and most straight sections will be flextrack. It would be nice to hear from anyone who has used the higher setting and ran into problems connecting their track.

The "Proximity" setting in RR-Track Lite (MTH RealTrax) allows up to 1/4 of the track gauge as a tolerance.  That translates to a 5/16" (.3125) allowable gap for joining O gauge.  If you disallow proximity joining in RR-Track Lite, some of the example RealTrax layouts that they provide have gaps of .125" that the user would not normally  be aware of.

From RR-Track Lite Help:

If you have enabled proximity joining (this is the default), RR-Track LIte will automatically mark as joined (but not move) unjoined compatible track ends that lie within 1/4 track width of each other and are misaligned by no more than 2.5°.

Steve

Steve, your comment stinks and I mean that in a friendly way. Because of it, I now believe that I've totally misunderstood what the "join error" setting in RR-Track means. I thought this setting allowed a gap of up to .20" for joints, but what it actually seems to do is lengthen each track section by up to .20" and that doesn't seem to be good when trying to design a layout. So, it appears that RR-Track doesn't actually have a variable "gap" setting like SCARM does, "proximity joining" is either On or Off and the setting is up to 5/16" gap and 2.5° of separation. Is that your understanding?

Also, I want to change the distance tolerance in SCARM from 2 to 15, but I know many people probably won't change the default and this will lead to problems/discussions when sharing SCARM files. I already hate having to use those small tracks that all manufacturers offer and if increasing the tolerance to 15 will minimize that, I think that would be a good thing. Maybe I should start a new thread to discuss the pros and cons.

Joe, after sleeping on it (I do my best work after "sleeping on it" ), I tend to agree that the 3rd layout is probably the best based on less specialized track, and the inherit elimination of the "S" curve by it's simple design.

I'm going to plug it into SCARM and play with siding options, as I tend to favor the more (rounded) square-like layouts, but by adding a siding to the turn off the switch, I can possibly achieve the square-like look I prefer, while maintaining the integrity of your design.

"You want the truth? You can't handle the truth! Why didn't you post a SCARM file???"

(---> A joke people...stolen from Jack Nicholson in "Few Good Men" <---)

Thank you again!!!!!!!!!!!!!!!!

I thought this setting allowed a gap of up to .20" for joints, but what it actually seems to do is lengthen each track section by up to .20" and that doesn't seem to be good when trying to design a layout. So, it appears that RR-Track doesn't actually have a variable "gap" setting like SCARM does, "proximity joining" is either On or Off and the setting is up to 5/16" gap and 2.5° of separation. Is that your understanding?

Dave,

I don't think it lengthens the track sections but just operates on the gap.  At least that is what I would hope.  Lengthening the track would be bad as you stated.  I don't really use RR-Trackl that much, mainly xtrkcad and SCARM.

Mixy has a good discussion on misalignment and for O gauge layouts I think it is worthwhile resetting the defaults to something tolerable like you suggest. 

Steve

While attempting the third layout in SCARM, I'm getting an error. What did I miss?

Originally Posted by Dave_R:

While attempting the third layout in SCARM, I'm getting an error. What did I miss?

Quote from RR-Track Help file:

One of the difficulties with comparing layouts designed with the aid of a PC and those that are actually on a layout is that the PC assumes perfect track joining (it always joins them with no gap error and no angular error) while the reality of the typical layout is that track pins don't always fit, you get tired of cutting your hands trying to shove track together, joints are misaligned, etc. All of these lead to mismatches between what is on the PC and what is on the layout. To simulate joining errors, RR-Track lets you set the average junction error at each joint by using the radio buttons on the Joining Error tab.

To set a non-zero (or a zero) joining error, choose the appropriate value. You can set average joining errors of up to 0.2 inches or 5mm. Every track junction created after you set the joining error will be gapped by this amount, effectively lengthening the track by the joining error. The default joining error is, of course, zero.

Unquote:

I could be wrong, but this seems to be telling me that if I set the value to 20, each joint made after that is 1/8" wide. If you zoom in on a set of tracks, you can actually see and measure this gap. The attached photo shows 3 pieces of 10" track joined with the join error set to 0 and 10. You can see that the first joint is completely closed whereas the 2nd has a .20 gap and what is supposed to be a 30" section actually measures out to 30.20".

Since I had it set to 20 while developing my current design, I now have to reset it to 0 and go back through my design to re-join every piece of track and contend with those that now won't rejoin. Most will probably be within the error margin, so I don't think there will be any real problem with the design. Fortunately, I'm not too far along.

Ok, that's all I'm going to say on the subject in this thread, I've hijacked it enough. Sorry, Dave R.

My gosh, this is getting frustrating. Setting it to as little as .081 works, but surely we shouldn't have to reset things every time a SCARM file is posted or every time we try to replicate something someone says will work. I'd set mine to the .15 that Mixy says will work for O-Gauge, but then anyone trying to replicate one of my examples may have problems if theirs is set differently.

Well, may be I should update SCARM code to write this setting directly in the SCARM project file, so when the project is reopened, SCARM will use the tolerance, written in the project. However, this will be effective for the newer versions of the program...

Mixy

I suspect opinions will be divided. I'd be content with a default based on the gauge of track that is chosen while keeping the option to change it if desired. You can see how many versions of this relatively simple loop have been posted and not been able to be replicated because of the different setting being used. You said you use 2mm for HO gauge, so maybe use 3.8mm for O gauge. The problem will be not arbitrarily changing a user setting. That said, I can't imagine anyone deviating from the 3.8 for O gauge if that setting truly works when actually laying the track.

Yep... this will be more complicated than I expected. May be I will make the default setting in SCARM to be gauge dependent and will not write the value in the project files, but before doing any changes in the program code, I will reconsider all this tolerance-related stuff and will carefully re-read this topic again

Mixy

Frankly, I cannot agree more (as the thread originator). IMO, we cannot allow for minor "gaps" or "twisting" of track to fit a layout. Either it matches to 0%, or it's a no-go (or the software needs updating).

In the real world, track sections will have their own tolerance vary; but if we purchase specific track pieces based on flawed software design, then we'll surely have alignment issues later.

I'll concede that the error values are minimal at this level of design, but one has to assume that these track configurations will be expanded to a greater level at some point in the future, and the errors introduced now may be problematic in the future.

Is it possible that the answer to my post is: "It can't be done"?

If the track pieces in the real world were able to make a perfect joins without any gaps or misalignments then there would be no need of tolerances in the software at all. But the real world is not perfect just like the track geometries of the different manufacturers.

Note that most track planning software packages are using a solid circles to indicate track joins - this is not by accident. The circles are used for smart hiding of imperfection in the track geometries. This is not the case with SCARM - it will always show how the tracks are connected - correct, with gap or with overlap. The tolerances are intended to show the "acceptable" error, but they are not a total solution for any misalignment and should be used with care.

And the answer to your very first question for that reverse loop is correct - it cannot be done in that way without forcing of the tracks as the gap is too big, no matter of the tolerance setting.

Hope that my explanation is of help

Dave, those loops should fit. The standard deviation I assume you are using is 2mm and the last one posted needs to be 2.1 to not show an error. Mixy says the max is 3.8, so it should be well within tolerances. But then I don't use FasTrack, so I don't know how flexible it is. I also don't know what setting Rail Reader is using and he didn't post the SCARM file so we could check.

As far as the answer to your post, I suspect the answer is "No, it can't be done" if you want a 0 tolerance. However, I think 0 is expecting too much. We're talking tubular track with plastic roadbed and even solid Atlas won't meet a 0 tolerance mandate. FWIW, people have been using the same or similar tolerance settings in other software for a long time designing layouts much more complicated than a loop. Until this topic came up, I didn't even know what RR-Track's setting was or how the join error I was using was affecting my designs. I'm confident the design for my layout is okay in spite of my gaff, but I'll find out for sure as soon as I fix things.

Originally Posted by DoubleDAZ:

I could be wrong, but this seems to be telling me that if I set the value to 20, each joint made after that is 1/8" wide.

Considering the default is 0, I'm the one who messed up by not checking what the setting was all about before I changed it. I thought it was a tolerance setting, but clearly it's not, at least not directly. I honestly have no idea now why I'd ever change it unless I had reason to believe, say, that a given brand of track was off by .10" or something when connected.

Folded dogbone layout with double reversing loops

This design was made using SCARM.  It uses RealTrax. 

I also have designed this one using Atlas O.

That is correct. SCARM does not extend or shorten the tracks. It only marks the tracks as connected if the distances between the ends are within the tolerance setting.

Mixy

I like to take a shot at summarizing this.

When completing a loop of track the join that is made is controlled by the following:

xtrkcad - Preferences->Connection Distance and Angle (user defined)

SCARM - Settings->Edit->Tolerances Distance and Angle (user defined)

RR-Track - Proximity setting (software defined and can only be turned off or turned on)

In addition, RR-Track has a setting under Options->Options->Join Error that allows you to specify the gap in the track joint while laying the track.  (user defined presets).

All of these settings take into account that in the real world, with only a limited number of manufacturer supplied predefined lengths of straight and curved track and not considering cutting track, there may be some mismatch that has to be handled within reason.  And the software allows you to do that.

I think the software does a pretty good job of eliminating wasted purchases of track pieces that cannot be used.  Is the software foolproof...no but it is an extremely huge benefit.

Steve