The Flatcars not Flat

I introduced my project to model the shipment of the Dominion of Canada from Exporail in Montreal to the National Railway Museum in York England here.  Obviously, a key part of this project, is the heavy load TTX flatcar that the locomotive was loaded on for shipment from Montreal to Halifax and the ship to take her back across the Atlantic Ocean.  No Ready-to-Run model from a major manufacturer exists of this type of car, and I couldn’t find any evidence that any resin or small run manufacturer has done the car, so I took it upon myself to design it as a 3D printed car.  For the most part, after some struggles with how to design the span bolsters to support the trucks, getting the dimensions right so the car would ride at the right height, and the couplers would be in the correct heights, I thought I had gotten close.  The one big issue was how to get weight on the car.

To accommodate weight, I designed the centre section of the car to have a cavity between the frames where a combination of 1/4 ounce and 1/2 ounce lead weights could be hidden.

131344RenderShapeways Render of QTTX131344 showing location of cavity for weight.

The second opportunity for weight is to be hidden inside the wire frame A4 locomotive that will be riding on top of the car. The one concern with weight here is that it will make the car top-heavy and tip prone as the weight will be well off the rails.

When I test ran it on a friends layout, it had some tracking issues as all the weight is in the centre of the car, which is prototypical, but in 1/87th scale, doesn’t necessarily work so well.  There is some limited room in the span bolsters to put some moldable lead weight into them without impeding the space needed for the wheels to move, which I will do whenever i get my hands on some moldable lead weight!

I thought I had sorted the weight issue, but then I took the car out last week to start weathering and rusting the deck, and I noticed a major problem, the little sag I had notice some months ago when I last was working on the project, had become a major sag that would no longer be acceptable.  I suspect, that this is a result of a design flaw in my 3D print, on a few different grounds.  I had to cut out some of the supports in the original print so the trucks had enough room to move, in doing so, I likely hurt the strength of the material.  I’ve found that Shapeways FUD material can be drilled for putting in details, but doesn’t seem to like being cut much in terms of the material retaining strength.  A second problem, is that the deck of the flatcar is only just at the minimum thickness to print properly, and be close to the scale thickness of the deck.  The third problem, is I suspect I ignored a lesson of those who originally designed flatcars for the railway, that they need a strong centre beam below the deck to prevent sagging and maintain structural rigidity.

IMGP2289RawConv3D Print of QTTX Flatcar 131344.  The sag in the middle isn’t as obvious in a broad shot, but it is there.
IMGP2296RawConvWhen the deck is set upside down the bowing is much more noticeable.
IMGP2291RawConvThe tip up is very obvious in a closeup of the span bolster area.

The good thing is, I haven’t put the car for sale through my Shapeways Store, so I haven’t sold anyone a car with a flawed design.  As well, once i determine the right fix for the deck, ordering a new one and painting/finishing it is easy enough to do.  The bad news is, I’m not 100% certain what the right approach to a fix is.  There are a couple of different options.  My recollection is that the car was perfectly flat when it was received, which means its warped over time from a combination of washing with warm water and soap to clean the support material, painting, the natural tendencies of the material, and potentially exposure to environmental factors (i.e. the heat in the apartment).  Not knowing which is the root cause, or the percent to ascribe to each potential issue, is going to make a fix tough.

The one universal change I am going to make is to thicken the deck between the frames, to make it as thick as the thinnest part of the frame.  I’m trying to strike the balance between cost and benefit.

Beyond this, to try to tackle the major issue of the bowing, the two most straightforward options I can see, are to re-print the car as it was and immediately inset a brace, or to re-print with increased internal supports.

For the first option, instead of the weights I used, i envision take a piece of flat steel bar, cut to fit precisely between the two bulkheads under the car. This would provide both weight, and hopefully, structural rigidity.

For the second option, as rendered below, there would still be channels on either side of a new centre beam to place weight

131344Render2Potential re-design to add a centre beam in the weight cavity.

I’m no engineer, but I have a basic understanding of structural design.  I think the above will make a stronger car in that having the integral bar will help eliminate the bowing (or hopefully reduce it).  Anyone with more experiene in structural design’s thoughts would be appreciated.  Once I settle on something and eventually order it, I’ll report back, but that will be some time down the road.

6 thoughts on “The Flatcars not Flat

  1. Hi Stephen,

    I think you made a nice try with a very difficult car to design for stiffness.

    The problem is not at the middle of the car. When I see the picture of the car up-side-down on the wood parallel, the bow gap is more or less constant along the “middle” of the car. I believe that the bending starts where the side sills angle up to a thinner height/depth. In the first 3D picture showing the underside structure in this area, the stiffening ribs are not very high/tall, probably to gain clearance for the wheels/trucks as you mention. This looks like the weakest area; there isn’t much to provide stiffness against the bending moment being encountered.

    I would try to make those inner ribs to maximum possible height, spaced closer together to hopefully stay clear of the wheels/trucks (like on a typical prototype). But that is likely not enough. Here’s a crazy idea: leave small square holes through the cross-beams at the bottom of the floor so that a piece of code 83 rail could be inserted & CA’d length-wise inside the side sills. The two rails would be inserted in the same orientation as it sits on the ties, for maximum stiffness. Something like that might be enough to stiffen up an area that needs to be kept with a low profile.

    Perhaps a piece of rectangular tube/solid brass running the centreline of the car from inner truck to inner truck (inner trucks mounted to this tube). Space for this would be left as a pocket in the underside of the car. Epoxy/CA fixing?

    Good Luck

  2. Rick,

    Thanks for that, i think that’s a really good suggestion. The next time i see you I’ll bring the car so you can see what it is doing in person. A re-design to add a channel down each side to take a rail or brass rod (I have some 3.2mm/0.1in brass square tube at home) isn’t a very long task, I’ve already updated it to see if Shapeways system finds any non-compliant areas from creating openings to slide a bar down. I think i will also design a channel along the centre of the car to take a metal bar of some sort which would go between the two square openings which are the mounting points for a styrene block and the span bolsters. A trip to an art store at lunch today may provide some different options in sizes of solid brass strip to use in adjusting the 3D model.

    I’m not planning on rushing into another print and a fix, so i can take my time and examine options, I’d rather get it done right than done fast!



  3. This is something you’ll never get away from in any process where heat is involved. It happens in 3D printing, injection moulding, welding… literally, everywhere.

    The difference is that in injection moulding, it can be offset by longer than normal cooling time, shot size variation, etc. In welding, small tack welds and jigs can be used to hold the metal straight before the final welds are made, but it’s still challenging and often machining is required.

    With 3D printed parts- FUD/ FXD acrylic, in particular, it is easy to straighten parts by holding them in hot water (but not so hot you can’t keep your hands in it) or using a blow hair dryer to heat the area, then hold it straight.

    Once straightened, it will not warp again. Your warp may have increased due to the adhesive used on the weight, or possibly it was inadvertently warped more during handling.

    The best thing to do is to warn your customers that this WILL happen (particularly with the shapes found in flatcars or gondolas), and tell them how to straighten the part.

    Unless you’re going to personally straighten every part, this is a fact of life. Any customer who has built resin kits will be used to this, and to date I haven’t found a single customer who, when forewarned, minded the couple of minutes required to straighten warped parts.

    By the way, don’t despair because your model is painted. I’ve straightened painted models before, where I missed seeing a warp.

    One minor note- prototype flatcars are generally designed with an upward arch, so they settle to a true flat or even a slight sag when fully loaded. If this really bothers you, you could consider arching the car slightly so its “warp” straightens it.

    Good luck!

    Steve Hunter

  4. Steve,

    Thanks for your thoughts, i appreciate them as i know you’ve got lots of experience dealing with Shapeways materials as well. I will try the heat method to see if i can effectively temper the car and clamp it back to straight with something i know is level.

    I have also been working on a re-design to try and make accomodations to add a centre stiffener and side bars as Rick suggested as a belt and suspenders approach.

    I agree that anyone likely to buy cars like this from Shapeways is likely experienced with resin kits and fixing warped parts. I just don’t want to put things out there if i’m not sure how they are going to react or what to warn people about. This car was causing me all kinds of concerns with the weight/tracking, and the bowing is just adding into that. Nothing that can’t be resolved with some careful thought, so onwards we go with experimenting!



  5. Hi Stephen,
    Steve does make valid points (I am in the injection mold and part design field).
    I assumed that the car was flat on arrival from the shop, and bent after running trials.
    The tapered section to the side sills could very well be a place for bending while the resin cures.
    It may also be a weak spot for weight.
    I will be at Supper on the 29th, if you wish to bring your flat car there.
    In any case, I look forward to seeing you there.
    Enjoying your blog!

  6. Thanks Rick,

    I will definitely bring the car to a future diner, but i unfortunately won’t be there next week because of some family commitments.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s