Printing Turnouts on a 3-D printer

[AndyID]

Hello Andy,

          

What size are you using? I have a RepRapPro Mendal and it came with a Standard nozzle size: 0.5mm However a 0.3mm nozzle is available. From what I understand the resolution also depends on filament size, adjstable feed and temperatures along with a few other variables. These are figures quoted by RepRap for my machine.  Accuracy: 0.1mm, Resolution:0.0125mm. The results also vary depending on the colour of the filament as well.

trustytrev.

 

Hi Trev,

 

It's 0.25 mm (or 10 thou in old money). I made it myself, and I only broke two drills in the process

 

Now that I have something that I think might work, I might go back to a 0.4 mm extruder and see how it looks. 0.4 would print a bit faster.

 

The filament size should have nothing to do with the extruder size - the software should adjust the filament feed rate based on what you input for the extruder diameter. You might have to play with the temperature setting when you change to different clours/suppliers.

 

Cheers!

Andy

[trustytrev]

Hello,

       I would think 3D printers are quite capable of printing running line and crossing timbers that can be cut to length quite succesfully and using moulded chairs could look quite reasonable. The variation of timber depth possible would be really handy for doing exposed timbers one sees in goods yards and other locations.

trustytrev.

[AndyID]

After a considerable number of experiments I've finally settled on a bullhead chair design that might actually work. Here it is.

 

 

Not very impressive is it? Looks like a piece of rail that's been jammed into a lump of gelatin!

 

Don't worry about the lack of cosmetic features. They will be added later. The hard bit was coming up with a design for the interface between the chair and the rail that will hold the rail vertical (no, I don't plan to tilt it ), at the correct gauge, grip the rail sufficiently and be strong enough to resist delamination. This is all made a bit more difficult because the plastic (PLA) is not very flexible.

 

More to follow.

 

Cheers!

A

[AndyID]

Cosmetics added. Views with rather harsh magnification.

 

I think this is about as good as it's going to get. Certainly much better than a blob of solder. Now to paste 'em on to some timbers and attempt to build a common crossing.

 

 

[AndyID]

Still a few issues to resolve, but this might be good enough for a test print. The slicer program says the print time will be 56 minutes 

 

[AndyID]

If the first layer adheres properly, things usually go OK. Looks like this will work.

 

 

The complete print. The chairs are a bit "dry". The printer settings will need a bit of tweaking, but I think this print is usable. The translucent plastic makes it difficult to see the detail. I'll get a better idea after the rails are installed and it gets a coat of paint.

[Coder Tim]

Hi,

 

I'm really enjoying this thread and think it's a great idea.  I wondered if you had considered using commercially available chairs and 3d printing a turnout base with recesses to locate them exactly?  The downside would be more individual components to attach.  The upside is that you could use more detailed chairs, and you wouldn't have to try and thread the crossing V onto chairs that were already fixed in place which I could imagine being problematic.  Anyway great work so far!

 

Best wishes,

 

Tim

[AndyID]

Hi Tim,

 

Thanks for the encouragement! At times I find myself wondering why I ever started this. It's pretty time-consuming, but I think I might get a reasonable return on the time invested, or maybe I'm just trying to justify the cost of the 3-D printer

 

It should be possible to combine this with purchased chairs, but it's probably not something I would want for a couple of reasons;

 

a) I doubt if my old mince pies would be capable of detecting the difference.

 

b ) I'm in the US, and it's a bit of a pain importing bits from the UK.

 

c) I'm too cheap! (it's a Scottish thing)

 

d) I'd still have to glue them on to the timbers. I'd rather have the printer do all the work.

 

I sort of cheated quite a bit with that print. There really ought to be a lot more types of chairs, but it takes quite a lot of effort to produce each design, If it looks like the whole concept will fly, I can always add them to the library later.

 

Cheers!

Andy

[raymw]

Hi Andy,

If you want to stick bits of pla together, I found that dichloromethane was best (see http://www.rmweb.co.uk/community/index.php?/topic/82640-domestic-3d-printing-pla/) for a few tests I made. I apply a few drops with a syringe, to the joint between the pieces and capillary action wicks it between them, making a quick, fast welded joint. I expect it can be used to smooth /polish the surface (in the same way that acetone is used for abs) but I've no need to try that - I want the stringy surface for what I do.

[hayfield]

Andy

 

This does look like an important step in turnout availability in the future and something far better than we could have hoped for not very long ago. There are some who don't mind about having the correct types of chair and those who do require a copy of the full size item, the next problems to overcome is the variety of sizes and geometry. But just to have something superior to the RTR H0 offerings is a major step in the right direction

 

What would be the cost of printing these items please

[AndyID]

Andy

 

This does look like an important step in turnout availability in the future and something far better than we could have hoped for not very long ago. There are some who don't mind about having the correct types of chair and those who do require a copy of the full size item, the next problems to overcome is the variety of sizes and geometry. But just to have something superior to the RTR H0 offerings is a major step in the right direction

 

What would be the cost of printing these items please

 

Hi John,

 

It's probably premature to declare victory. The plastic material (PLA) has limitations. I think I'll have to resort to some subterfuge in the vicinity of the check rails and frog to make the print sufficiently robust. On the other hand, you can do things with a 3-D print that would be extremely difficult with injection moulding, and I can take advantage of that.

 

It would not be a problem to create models of all the correct chairs and print them. Unlike injection moulding, there is no tooling cost - the only cost is the time to design the chairs in CAD, but that is nothing compared to the cost of developing injection tooling.

 

Different scales and turnout geometries are not a problem. It only takes a few minutes to convert the timbers from a Templot turnout into a 3-D version in CAD. Then you drag and drop the chairs on to the timbers. That is a bit time-consuming, but there are some methods that could speed it up a lot. (Once you have a reasonable model of a chair, it can easily be scaled to any arbitrary size.)

 

The state-of-the-art is probably limited to OO/H0 scale. I'm not sure this would work for N, but it might be possible. Going up to O would simplify things a lot, although it might require a larger 3-D printer.

 

Regarding cost, if you were to buy a 3-D printer and print the bits yourself, the material cost (the plastic itself) is peanuts. The cost is all to do with the CAD design effort, plus the "machine minder" time. That crossing took one hour to print, and prints don't always go according to plan. There are printers with multiple extruder heads that could probably shorten the print time considerably.

 

Cheers!

Andy

[AndyID]

For those who enjoy watching paint dry

 

I think I've solved the aforementioned concern regarding the check rails. First the "after" pix.

 

 

And now the "before" pix (before painting)

 

 

 

The solution was to add the reinforcing bar in the gap between the rails. I don't think it will be visible after ballasting.

 

You can see from these shots how much difference a coat of paint makes. Next step is to do something similar in the vicinity of the frog.

[trustytrev]

Hello Andy,

              If you used brown filament would  painting not be easier?

Regards,

trustytrev.

[AndyID]

Hello Andy,

              If you used brown filament would  painting not be easier?

Regards,

trustytrev.

 

Hi Trev,

 

Likely wouldn't make much difference. It would probably still need to be painted anyway. In the sample above I just gave it one coat of matte black from a spray can.

 

Cheers!

Andy

[Joseph_Pestell]

Not that I have yet managed to master Templot, but the idea of being able to transfer data direct to files for printing makes this a very interesting project.

 

I know that 3D printers come in a huge range, from basic and cheap to eyewateringly expensive. What sort of machine are you using? Does it come at a price which would be reasonable to an individual modeller with a large layout project to build turnouts for?

[AndyID]

Not that I have yet managed to master Templot, but the idea of being able to transfer data direct to files for printing makes this a very interesting project.

 

I know that 3D printers come in a huge range, from basic and cheap to eyewateringly expensive. What sort of machine are you using? Does it come at a price which would be reasonable to an individual modeller with a large layout project to build turnouts for?

 

Hi Joseph,

 

It's a Printrbot Simple. It was the cheapest one I could find at the time (a couple of years ago). Came as a self-assembly kit. An improved version is available in the US for $350 at the moment. It's the sort of thing members in a club might share, or an individual might buy one and sell it on Ebay if they no longer have use for it.

 

Templot is not hard. You should be able to print a paper template for a turnout after a couple of hours. Double slips will take a bit longer!

 

You also need to have some 3-D CAD software to create the 3-D model for the printer. I use TurboCAD - mainly because I've used it for many years.

 

Cheers!

Andy

[martin_wynne]

Templot is not hard. You should be able to print a paper template for a turnout after a couple of hours.

 

Hi Andy,

 

I think you meant a couple of minutes.

 

Hi Joseph, here's a Templot video in under 2 minutes, and it's quicker now (the common gauges are now in the gauge menu):

 

 http://templot.com/martweb/videos/flash/startup/startup_em.html

 

regards,

 

Martin.

[AndyID]

I'm a slow learner, and then there's the time-zone difference

[AndyID]

Still something of a "work in progress", but this should give a general idea of what it will look like.

 

 

[mightbe]

Looks good. Are the two parts of the vee attached to each other electrically somehow? You said soldering was out due to melting plastic..

 

Quentin

[hayfield]

Andy

 

That does look very good. It will be interesting how you fitted the bent bits, as well as how you intend to solder the dropper wires/vees etc. Would low melt solder be one answer?.

 

Full marks for proving the process, this could/is be a major step forward for those who want better quality turnouts and crossings and I would think the next step forward in the hobby

[Joseph_Pestell]

This is surely a gamechanger with 3D printers available so cheaply.

 

I'm not convinced still that I could master the necessary software (I have tried CAD in the past without success) but there is surely an opportunity here for someone to produce bespoke kits.

[AndyID]

Hi Quentin and John,

 

Good questions! As it's getting a bit late here, I'll try to reply in depth tomorrow.

 

This was all a bit of an experiment to see if it might work. There are still some problems to solve, but I don't see any fundamental reasons why it can't be made to work. Getting everything set up just right does take a fair amount of time, but the possibility of converting Templot templates into actual turnouts on a 3-D printer is a very interesting notion.

 

Best regards,

Andy

[eldavo]

That looks pretty tidy Andy. I wonder whether it would be possible to extract turnout data from Templot electronically then generate an assembly script for something like OpenSCAD to create the complete turnout base as an STL file? Chairs and things could be created as common components then just reused to assemble the desired configuration. All very interesting.

 

Cheers

Dave

[AndyID]

Dear All,

 

As promised, some more information:

 

The vee is not soldered (yet). If I'm feeling brave, I might try hitting it with a flux and a very hot iron. Maybe some heatsink clamps would help too. However, the wing rails would still have to be connected somehow. I could embed phosphor bronze wires into the timbers and rely on a pressure connection, but that might fail over time due to oxidation.

 

An alternative would be small sections of fishplate with wires pre-attached that are slid on to the rail as it is inserted into the chairs, but that would detract from the appearance.

 

I could have pre-assembled the entire frog and glued it in place, but I was interested to see if the printed base would simplify the construction. It does seem to do that. I didn't have to tweak any of the rail positions - all the rails were simply slid into place.

 

The check rails were installed first.To insert a checkrail, I pre-flared one end then dragged the straight end through the chairs with pliers. There is enough flex in the rail to allow that. Then I flared the previously straight end.

 

The wing rails are pre-notched at the bend point to form a sort of hinge. They were dragged into place from the switch end. The ends were flared after installation. I was quite surprised that I was able to drag the rail into position without busting any chairs.

 

The vee actually sits in a low vee shaped pocket that also maintains the flangeway gap. That makes it almost impossible to misalign the vee and wing positions.

 

I was really pleased with the way it all went together. No gauges required. It would be interesting to see if other people find it simple enough.

 

My printer settings still need work. In the sample above, for strength, I had to increased the slice thickness. Unfortunately, the slot in the chairs has to be designed around the slice thickness, so I ended up having to "ream" the rail slots. This was done by applying pressure to the rail and touching the end of the rail with a soldering iron to help it slide into the chair! That should not be necessary in future, although it may be necessary to push a knife into the slots to remove some material that inadvertently bridges the gap.

 

The chairs in the vee/frog area are parallel with the timbers. That means the slots are not perpendicular to the chairs. To get the slots at the correct angle, I made the chairs solid (no rail slot) in the CAD model then I "drilled" the slots in CAD by subtracting slots from the chairs using a long extrusion with the desired profile.

 

Templot does have a 3-D export selection, but I just exported the timbers and rail position in 2-D, snapped rectangles onto the timbers in CAD and extruded them to the desired height. It didn't take long. Positioning the chairs etc takes a lot longer, although I could speed that up a lot with practice.

 

BTW, Martin probably guessed, but the sample isn't exactly OO. The gauge and clearances are the same as OO-SF, but the timbers are scaled at 1:82 (mixed units tend to bring me out in a rash, so I'm not going to say how many mm/foot that is  )

 

Cheers!

Andy