Custom track with wide radius curves

[zephyr1934]

[

full gallery]

As everyone knows, lego curves are very tight, so one has to build short rolling stock to make it around the curves without problems. Unfortunately, even at 6 wide many passenger train cars built to scale should be 50+ studs long, e.g., my Superliner cars were built with accurate length to width proportions and they are 52 studs long. They look HORRIBLE in the curves, with a single car spanning a quarter turn- notice that the ends swing out almost six studs (a full car width) past the adjacent car in the pictures below. These long cars on the short curves also have so much drag that it is hard to find a good speed that the engine can pull through one curve that will not send the train speeding away on the straight track to derail at the next curve.

<<regular lego curves

Over the past few years several folks have promised wide radius curves and I've grown weary of waiting for someone to actually produce a viable curve. So I set out to build my own. I contemplated 3D printing, but it would be way too expensive. After several initial attempts, I settled on using third party ABS rail stock glued to lego tiles (yes, "glue" is a four letter word, but so is "lego," grin... in any event, this exercise is definitely not for the purist). The tricky issues are the fact that the lego gauge does not split well into LDU, and worse, making nice smooth curves.

<<new custom curves

My solution was to make a rail gauge to ensure that the rails have the correct spacing. I am still revising the design, but once I've finalized it, I plan to offer the rail gauges for sale.

For my prototype there were several things I had to decide upon. I personally like the look of one stud ties, but they do not provide a good opportunity for a rail joint on a curve. So I decided to use two stud ties for now, mimicking the look of the old 4.5v rails. I also had to choose the curve radius. I wanted to go wide. Until I prove this works, I want to do only two track segments per quarter turn. So with these constraints and my rail stock the widest I could go was twice the radius of regular lego curves, and so that is what I chose for my prototypes. After I finish one loop of track, I'll experiment with other radii. In fact one could even do variable radius curves, but it would be more work to figure out the transitions, so that is something to play with in the future. The rails are darker than lego dark gray, that shouldn't be a problem if the entire loop is the same color.

The straight track went fairly easy, with a single segment coming in at 76 studs (4 studs short of 5 regular lego track segments).

The curves took a lot more time to get right, but I think I've got the process figured out now (I'll need to make another quarter turn before I know for sure). This shot shows one segment of the new curve next to four segments of lego curve.

My 52 stud long cars work so much better on these wide radius curves. Aesthetically, I think the 52 stud long cars would look better on even larger radii curves (perhaps my next project). Meanwhile, I think the curves look about right for my 42 stud long cars, as shown below.

before

after

More to come, but for now, you can find a few more examples of my progress thus far here (including a transition piece to adapt to conventional PF track- primarily for switches- which I have no intention of making... at least not any time soon).

Oh, and if you scroll back up to the top image in this post, you'll see the middle tie on the curve sticks out one stud. That's because I grabbed a 2x2 plate instead of a 1x2 to connect the rails. Also, my carpet has a little give, so I had to put a plate underneath the joints too to keep the weight of the train from popping the rail joints. I hope that will not be the case when I set it up on a hard floor, and I'll keep you posted when I test it. Worst case is that I'll need to add a layer of plates below, in which case I'll disguise it like ballast and then I can go to single stud ties.

[

full gallery]

[garethjellis]

This is very nice work!

[Pizzareno]

Would you be able to glue a metal rail to the tiles? In essence being able to use it with 9v motors?

[zephyr1934]

@garethjellis

thanks for the kind words

@Pizzareno

No, not using this method. But there are other ways to attach metal rails to lego. The big trouble I've heard about metal rails at this scale is the fact that you will need a rail bender to make curves. So it is doable (and I think there is even an railbricks article by someone who has done it) but it is non-trivial.

[Daedalus304]

Awesome, I am super excited for this. I have been disappointed that all the other custom track methods basically died at the hands of 9v functionality not being great; a real shame for all of us who have embraced PF and would have enjoyed it either way.

My big question is, how does it compare cost wise? I would love to do some custom trackwork for a permanent back-yard railway, but there`s no way I`d leave proper LEGO track out there. Your solution seems to be one with a lot of possibilities.

[Pizzareno]

Thanks Zephyr, I have seen the article in the Railbricks where the guy cuts the plates and adds the metal rails. I might have a go at this and get some plates on a lugbulk order or something

I guess I just want something to make the rails at the right gauge and that is the major hang up.

[codefox421]

I have been disappointed that all the other custom track methods basically died at the hands of 9v functionality not being great; a real shame for all of us who have embraced PF and would have enjoyed it either way.

I couldn't agree more. I used to be hung up on 9v functionality, but I'm realizing more and more that it makes total sense to switch to PF (unless you have heaps upon piles of 9v track! )

After several initial attempts, I settled on using third party ABS rail stock glued to lego tiles (yes, "glue" is a four letter word, but so is "lego," grin... in any event, this exercise is definitely not for the purist). The tricky issues are the fact that the lego gauge does not split well into LDU, and worse, making nice smooth curves.

How much does it cost you to make a section of track?

Where do you get the ABS rail stock?

My first thought for affixing ABS to ABS would have been acetone; what glue are you using and how does it hold up?

And how do you make those curves so nice and smooth?

By the way, Daedalus...

I would love to do some custom trackwork for a permanent back-yard railway, but there`s no way I`d leave proper LEGO track out there.

...I stumbled upon

a while ago. It offers another promising option for low-cost, outdoor track, though maybe not quite as pretty as an ABS variant.

[zephyr1934]

@Daedalus304

The material cost should be less than that of lego straight track assuming $2.50 per lego straight track segment. I'm using Plastruct, PLS90025 I Beam ABS 1/4" for the rails. It is roughly $5-$8 for 4 rails (or two track sections), where each one is just shy of 5 lego track segments long. At least those are the prices in the US, might be more expensive internationally. Then there are the ties. I'm using 1x8 tiles and for one segment the ties cost about the same as the rails. If everyone starts doing this I would think the tile price would go up, but there are plenty of other alternatives for the ties if that happens.

Beyond that, there is the cost for the rail gauges (no estimate yet, as I'm still refining the design), glue, and glue applicator. Assuming you are building at least one loop, it should still be cheaper than all lego track.

Assembly is more involved than regular lego, you need to build a jig, etc. etc. [edit- you will also have to cut off the excess rail at the ends of the curves, but fortunately when building straight rails the rail stock is an integer number of studs long].

One neat benefit is that with larger the curves, you will need fewer straight rails. This approach also has the potential to realize some of the "flex track" like benefits, allowing you to add wiggles, gentle curves, etc. as you glue the rails down. Though once you've glued the rails, the geometry is permanent.

My current plans are to work out the kinks of this process before I go into the details of how to do it yourself. If there are early adopters who are just itching to play with it and take on some of the trial and error (e.g., coming up with the best rail joint solution), let me know.

As for a backyard railway, why not put down the right-of-way using 2x4's (boards not lego) or similar. Then lay the lego only when you run the trains, and pull the track up at the end of the day? Since these rails are ABS, they will have similar issues with sun exposure that lego would (though from earlier threads here in EB it sounds like it might take months or maybe even years before the sun damage is significant).

@Pizzareno

Yes, there is not much tolerance for the interface between lego train wheels and the track. It has to be precise and consistent throughout. Sure, if you just need gauge bars, I should be able to adapt my some of my designs to other rail stock. Although I have not tried metal rail myself, it does sound like you would need a rail bender for the curves at this size. Find a good hobby store that specializes in larger scales and they should be able to tell you more.

I couldn't agree more. I used to be hung up on 9v functionality, but I'm realizing more and more that it makes total sense to switch to PF (unless you have heaps upon piles of 9v track! )

...

How much does it cost you to make a section of track?

Where do you get the ABS rail stock?

My first thought for affixing ABS to ABS would have been acetone; what glue are you using and how does it hold up?

And how do you make those curves so nice and smooth?

Oh, there is still a place for 9v and wide radius 9v track would be wonderful. However, this approach probably is not the path to 9v.

As for cost, see above. For the rails, google "PLS90025" and you should find it pretty quickly. Most hobby shops that sell the ABS parts should also carry ABS glue (I think at least one brand calls it plasti-weld or something similar). You will also want an pinpoint applicator. It is the same stuff that lego uses to glue the large models together for parks, etc.. The glue bond is actually stronger than the plastic. As for durability, so far so good, but I have yet to do extreme stress tests (hence my earlier warning about trial and error).

The two keys to the curves are the rail gauges and a good jig built out of a lot of lego. Both designs are still in flux, which is why I haven't posted photos yet.

[greenmtvince]

Ahhh! The Kragle!

Did you happen to try 2x8 plates and removing the studs in this process like the old metal flex-track technique? I feel like you might be able to place a 2x4 and 2x 2x1 plates or bricks in there temprorarily (maybe you still need a styrene spacer) to help line up the rails and eliminate the need for a track gauge. It also might lend itself better to traditional ballast techniques.

I may have to play with this.

Also thanks for introducing us to those I beams. You saved me from hacking up old 4.5v/12v rails to make switch tracks in a street.

[Dan-147]

I did a quick check in Plastruct's catalog http://www.plastruct.com/picat/STRCTURAL_SHAPES_FTGS_9.pdf

I was wondering if you had considered the 1/4" Te (T-8 on page 2) rather than the 1/4" I-beam (B-8 on page 3). A quick draw-up in AutoCad shows that it is a better match to both 9-volt/PF track and 4,5/12-volt track. Plus, the Te has more surface area that is glued to the tile so it should have a better bond.

Hope this helps.

Dan-147

[greenmtvince]

Great work, Dan.

Makes me wonder if Plastruct couldn't just offer a custom extrusion that was exactly what we needed.

[Aaron]

Thank you for this information. Most of my rolling stock is 80+ studs long, so this is definitely something I need to experiment with. I've always wanted wide radius track, and it's incredibly silly that the people who initially produced it gave up because of the lack of 9V functionality. To me, power functions trains are superior in every way due to the flexibility, and the fact that they function more like real trains.

I seriously wish I would have bought some ME Rails while they were still in production, but I think this could be a viable alternative. I'm tempted to just ditch all of my normal LEGO track and do it entirely using this method, so that I don't even have to worry about conversion pieces. I'm going to experiment with single tiles, and try to find a way to make them work.

Edited March 14, 2014 by Aaron

[zephyr1934]

Ahhh! The Kragle!

Did you happen to try 2x8 plates and removing the studs in this process like the old metal flex-track technique? I feel like you might be able to place a 2x4 and 2x 2x1 plates or bricks in there temprorarily (maybe you still need a styrene spacer) to help line up the rails and eliminate the need for a track gauge. It also might lend itself better to traditional ballast techniques.

I may have to play with this.

Also thanks for introducing us to those I beams. You saved me from hacking up old 4.5v/12v rails to make switch tracks in a street.

Yes the Kragle, sometimes drastic measures are necessary (grin, but as I said, not for the purists).

The problem with lego track is that as far as I can tell, except for train wheels, there is NOTHING else in the lego universe that matches the lego rail spacing. I actually envisioned being able to go further than traditional ballast techniques, e.g., glue the rails to gray 1x1 or 1x2 tiles, and put those on brown or black plates to emulate tie plates on top of ties. While potentially looking great, such a build would be more expensive and might not be as sturdy. I say "potentially" because I've never sketched it out. The combined height of the rails and tie plates might be too tall.

Yes, look through the Plastruct catalog, there are other options too if you want to build switches.

I did a quick check in Plastruct's catalog http://www.plastruct...APES_FTGS_9.pdf

I was wondering if you had considered the 1/4" Te (T-8 on page 2) rather than the 1/4" I-beam (B-8 on page 3). A quick draw-up in AutoCad shows that it is a better match to both 9-volt/PF track and 4,5/12-volt track. Plus, the Te has more surface area that is glued to the tile so it should have a better bond.

...

Hope this helps.

Dan-147

Now that you mention it, I can see where the "T" beams have more of a retro lego feel to them. Back in the sketching phase I had contemplated the "T" but I did not like the wide base (simply a matter of personal taste). I had originally set out to try to find a glueless method and in that context the I beam is superior. It is 3.2 mm wide and 6.4 mm tall, while a lego plate is 3.2 mm tall. So in theory you can slip the I beam between the studs (though I have since found that the width varies, so it is more like 3.2 +/-0.15 mm, so the clutch varies along the beam). I wanted to use a 2x8 plate and just wedge the rails between the studs, but the resulting rail spacing is about 1.2 mm too tight for the train wheels. So I also bought some of the C-channel (C-8). My thought was that I could glue that to the plates and wedge the I beam on the other side and get the correct rail spacing. What would be ideal for this context would be to remove one of the flanges from the I beam (I tried to do so with horrible results). That was when I settled into the glued method at the start of this thread. In this glued context, I THINK the I beams are easier to work with because the rail gauges guide both the top and the bottom of the rail. Though I suspect the T beams could also be made to work.

The fact that I beams are wider than rail head on lego brand track is less noticeable than the color change between the two track styles. Another thing that I like about the I beams is that they kind of look like actual rails from the side. For my uses I am not too concerned about the appearance of the interface with regular lego track. I will have my two loops of lego track in my lego room and I will use the custom track for temporary layouts in... say... the living room (but don't tell my wife...). While I have some long-range ideas for switches, I do not plan on doing anything in that direction in the near future. So being able to convert to lego track is very helpful to incorporate switches.

As for the bonding area, the I beams have sufficient surface. I tried pulling a tile off of a rail and wound up peeling off the lower part of the rail with it rather than breaking the bond, the plastiweld is strong stuff. But unlike crazy glue, it will not stick to you (though you want to prevent getting it on yourself as it is still a nasty chemical).

BTW, great drawings.

Great work, Dan.

Makes me wonder if Plastruct couldn't just offer a custom extrusion that was exactly what we needed.

I believe the plastruct ABS is molded rather than extruded (while the styrene is extruded). I had contemplated approaching Plastruct and asking them what it would take to do an I beam with only three flanges. Still might in the future, but I would imagine they would want dozens if not hundreds of folks to express an interest before they make a new mold. Or maybe they simply speak in $$$. Both are possibilities, so keep the critical mass growing...

Thank you for this information. Most of my rolling stock is 80+ studs long, so this is definitely something I need to experiment with. I've always wanted wide radius track, and it's incredibly silly that the people who initially produced it gave up because of the lack of 9V functionality. To me, power functions trains are superior in every way due to the flexibility, and the fact that they function more like real trains.

I seriously wish I would have bought some ME Rails while they were still in production, but I think this could be a viable alternative. I'm tempted to just ditch all of my normal LEGO track and do it entirely using this method, so that I don't even have to worry about conversion pieces. I'm going to experiment with single tiles, and try to find a way to make them work.

From what I've heard, the problems with the earlier designs was not the lack of 9v functionality, rather, that the original approach to get 9v functionality lead to other problems that ultimately made the design unworkable. Even my approach would not work as a commercial product. The only way it works is with the user exerting the sweat equity to build the rails.

As I said earlier, if any of you are so anxious that you want to help with the trial and error part of the design, I could polish up my rail gauges and start sharing those too. Just like lego, there is no single correct answer.

With 80+ stud long cars, you'll probably want 4x standard lego radii curves (2x of what I have right now). I'd prefer to get the rail gauges working for one radius working before I start making other radii, but if there is a lot of noise for a wider radius from someone who wants to do the development I could probably be talked into making the gauge.

[Bricktrix]

I got all excited about this until I saw it was plastic rails

I've had a few discussions over the last few years (with a mate of mine) and he has offered a VERY real possibility to put into production through one of his businesses ( http://www.justliketherealthing.co.uk/ ) plastic sleepers with combined ties to fit LEGO dimensions and also whatever metal gauge rails we / I wanted to go with. He even gave us some rails to experiment with, BUT I've never pursued it due to the many legal complications I can foresee with a potential production run with stuff like this.

Ben, if this of interest to you, PM me with your email, etc and we can talk further.....I think legally its a non starter, but ya never know.

(I still need to get some of your custom rods off you also, when I finally get around to revisiting some of my models!)

[Hrw-Amen]

Looks good, I really need those wider curves for the same reasons everyone else is saying, long rolling stock. Most of my passenger coaches are in the 40+ stud range so any way to make wider curves. I will have to keep an eye open.

[JGW3000]

I've been following this topic with interest, since I will likely get trains sometime in the next year... Quick question - are the flexible tracks really that bad that you need to make custom alternatives?

[zephyr1934]

@Bricktrix

Yes, metal rails would be fantastic, but that would be more involved than this project. I really like the idea of custom sleepers to interface between lego bricks and metal track. I'll chat with you off line about thoughts and ideas.

@Hrw-Amen

Indeed, lego trains need wide radius curves, hopefully we'll have a good process worked out here soon.

@JGW3000

Most pure lego train layouts would use straight and curve segments (roughly 16 studs long) for most of the track. The flex tracks are very handy filler, but horrible in quantity compared to the regular track. Normal lego curves limit most trains cars to 28 studs long (or perhaps push it out to 32 studs). There are tricks to go longer, but I would not recommend it for your first build (if you keep reading the train tech forum, you should see lots of ideas come up in the various MOCs). So the wide radius curves are for the folks who have hit the limit with the regular curves. Most folks don't need to worry about these limits imposed by regular lego curves, but there are some pieces of equipment that just need to be built long.

[ronenson]

What I don't understand is that in these discussions on custom curves almost no-one seems to mention that you can mix flextrack with straights alternating each other. You get a nice long curve without any modding required. They even put the 2 together in one set.

[zephyr1934]

Oh yes, long before the flex track, there were folks who were alternating straight and curves for a similar effect. At least as I've heard, the flex track was originally intended to make wide radius curves in that manner. I suspect lego ultimately added in the guardrails (not found in the prototypes they shared with AFOLs) to protect against the kid who alternates left-right-left-right with a sequence of flex track. Between the look of the guardrails and the extra bumpiness of the implementation, many AFOLs have not embraced the flex track (though some AFOLs like them). The flex-straight-flex-straight idea arose at the same time as the "slightly misaligned" pure-straight-track-built wide radius curves. So far the latter has proven to be more popular among AFOL clubs, but I've seen examples of the former too. These two straight track methods are probably the best pure lego solution. The one clear problem with them is the fact that they will not work with 9v trains (although the two commercial wide radius curve efforts sought to address the 9v factor, mine will not).

There is definitely no single best solution. My specific motivation to deviate from a pure lego solution was two fold. First, I wanted to avoid the high cost of lego straight track (though using enlighten rails might still be cheaper than my solution) and the pure straight track wide radius curves still have a lot of drag at the end of each straight track segment. For example, my superliner train set shown in the top of the first post stalled out on the PennLUG curves at Brickworld after only one or two loops, but I can run it until the battery dies on conventional lego curves. I doubt it would be a problem for XL motors, but I'm only using two pf train motors on this train with a single v1 IR receiver (I still haven't tried the v2 IR receivers yet). I also had an original hope to improve on the aesthetics of the track design, though I don't think I've achieved that yet.

[Hrw-Amen]

This is my answer so far to wider radius. It can be viewed on my flicr account http://www.flickr.com/photos/hrw-amen/

Edited March 17, 2014 by Hrw-Amen

[zephyr1934]

Nice Hrw-Amen

[Hrw-Amen]

It simply comes down to how acurate you are with cutting and how good you are with super glue. But you have to have a fairly sharp knife blade or a fine saw. I tried both and the saw was by far the easiest although you loose a little in lenght but not a great deal. Using one straight and one curved normal track per section does give you quite a large radius in the turn. The first one I did admittedly looked a bit messy but after a few it became better. I think the only down side is that if you do not line up all the segments really well the trains tend to hop around abit and not run smooth, but once you have it worked out it is good. A bit brittle if you handle it roughly, but a lot of that is down toi the glue. I used superglue as I wanted a quick result.