Experiments in making a 3d printed replacement for an aluminum extrusion frame

Hi all

This is a write up of experiments to replace a frame made of aluminum extrusions with a 3d printed one.

Having some time off at Christmas I wondered if I could use the non 3d printed parts of a ReprapPro Prusa i3 to make a Smartrapcore Alu (a core x-y printer [smartfriendz.com]). However I found it very hard to find the correct aluminum extrusions without long shipping time, so I wondered if it was possible to create an equivalent using PLA. I'm putting this post in this section because I can see it may be of use to other printers as well, I'm just using a specific example.

Here are the main advantages and disadvantages that I can see of having a 3d printed frame:

Advantages:

• Make the printer more self replicating (I think if it was possible it would be one of the highest percentage self replicating printers)
• Reduce the number materials needed to create a kit for people having trouble sourcing the parts like me
• Potentially reduce cost (although maybe not by a lot)
• Make it easier for people to experiment with different designs for the frame
• You can have it in what ever colour you want

Disadvantages:

• Much longer print time
• Potentially weaker frame

What I'm showing is where I have got to when I ran out of time, I would really love it if others could pick up where I've left off. I wont have access to my printer for the next few months but very happy to help in other ways. I've added everything to Github here [github.com]. The major thing I'm missing is I don't understand what tolerances are needed for a 3D printed frame and how to test if I have reached the requirements.

The current design was done in Tinkercad and is quite simple, there are three different parts:

• Internal rods of 100mm in length

external sheathes (100m length) which come in two versions which can be shortened to create different length frame sides.

• end pieces which will screw into the existing corner pieces
• middle sheathes which are just a long tube.

There were several previous iterations that were simply too complicated, I think my current approach has a few strengths:

• The end results seem very straight
• Uses PLA which means you don't need a heated bed
• The parts are easy to print and can be printed at a higher temperature to ensure good layer bonding
• The parts are printed with the rods lying down and the sheathes on their ends so printed layers run in different directions meaning the parts are strong in all directions.
• Simple construction, you only need a wooden mallet or some other tool to encourage the parts together
• Strong joins between parts, the inner rods join in the middle of the outer sheathes
• The longest dimension of parts is 100mm meaning it can be printed on a small print volume printer.

Printing notes:
I printed everything seen with Hatchbox PLA using a Printrbot Simple Metal with a E3D V6 head with 0.3mm nozzle. I used a 0.6 outer shell, 35% infill, 2mm layer heights and printed at 217 degrees C with a cooling fan.

The internal rods are printed at around 95% size so they fit snuggly inside the outer sheathes. I printed shorter test pieces to find the right scale however shorter pieces have slightly less friction when pushing them together, I found 96% worked fine with the test pieces but 95% worked much better with the full size pieces. It also helps a lot in getting the parts to fit well to scrape the corners of the internal rods a bit with a sharp knife.



Issues

• The time it takes to print the pieces is quite long
• Tolerance for print size of internal rods is very small, however its easy to make shorter lengths to check tolerances
• The direction for the join for the internal rods is the sample as the direction for the direction of the print layers in the outer sheath meaning this could be a weak point but I think this could be fixed with glue.
• You have to scrape the edges off the internal rods so that the size is correct for the sides of it to grip the sides of the outer sheath.
• Once the parts are put together its not possible to take them apart again
• Whilst this maintains the width of the Aluminum you're not able to screw the other printed parts into the rods so would have to adapt the or make much wider rods to accommodate screw holes, here is a test piece.
  

Ideas for future versions

• use a glue to join parts together, perhaps using a very rigid glue e.g superglue or araldite to increase rigidity of the frame. This could help with both the potential weak point in the middle of each section and make it easier to push the parts together because it would reduce friction when inserting the rods
• Round the edges of the internal rods
• Create some sort of mark on the internal rods so it is clear where the middle of the internal rods are
• Use transparent material to see when the parts are fitting together correctly

Any help or suggestions would be very much appreciated, again here is the github repository of the files [github.com] , as I mentioned at the start the major thing I'm missing is I don't understand what tolerances areneeded for a 3D printed frame and how to test if I have reached the requirements. Here is a photo of how much the full 400mm length bends when one end is held down and the other end pulled up.



Thanks

John

Edited 3 time(s). Last edit at 01/02/2016 08:37AM by John Cummings.

Nice work. If printed in ABS you could bond the parts together with acetone. Have you estimated the cost/meter of printing these? It's an intersting idea to print frame parts, but the price/performance of aluminium extrusion is tough to beat.

Thanks, I don't have a heated be so ABS is out of my reach for now however I really like the acetone bonding approach. Is there any information on what would be the best readily available and preferably as non toxic glue to stick PLA together?

I'm not thinking about it from a price point of view really, more of a self replicating and reducing different materials needed.

After looking at this further it seems as though a simpler solution would be to print standard alu v-slot extrusion shape [www.thingiverse.com] and use a similar rod solution to my above post. I think that either the 20x20mm or 40x20mm could work but each have their own issues.

• 20x20mm would fit into existing designs and could be joined by creating rods that fitted in the external v slot groves, however this would make it awkward to attach parts (you'd have to leave gaps in specific places that couldn't be moved) and it would stop you using them as rails if this was something wanted. I don't think its realistic to print a cylindrical rod to fit in the 5mm hole that would be strong enough but I'd be very happy to be corrected.
• 40x20mm could have rods running through the internal cavity which wouldn't interfere with the external structure at all however it would increase the amount of material needed and also the time needed to print.

I also think that there will be pros and cons in deciding which way to print it.

• Printing vertically (like a tower) would mean that the shape is very easy to print but would make it unusable if you wanted to use it as a slide (not needed in the design I'm trying to make it for).
• Printing horizontally may allow you to keep it as a slide but would be much harder to print because of the overhangs.

Thanks

John

Edited 1 time(s). Last edit at 01/02/2016 07:15PM by John Cummings.

I've created some simple rods to fit in the v-slot grooves and put them on Github, I assume if they are printed at 95% or so they should fit well, however I'm not sure how strong they will be as they are only 4mm tall at their highest point, printing them nearly solid should help [github.com]

I have just done a quick calculation (UK prices) for the amount of 20x20mm vslot rail needed for the Smartrapcore Alu:

4 x 400mm
4 x 340mm
4 x 300m

Ooznest will charge £23.40 for the aluminum rail cut to size + £4.36 First Class recorded post

4160mm of 100% infill printed will cost around £23 at £18 per kg (obviously you have buy 2 rolls) and this doesn't take into account failed prints or electricity costs.

So it would cost about the same

Edited 1 time(s). Last edit at 01/02/2016 08:31PM by John Cummings.

That's pretty good! I was expecting it to be multiples of the cost, and in fact it's only a little more than I pay for 1" square Al tube, which I thought was a bargain ($22 CAD in 8' lengths).

Off the top of my head:
You could probably make use of tensegrity structures to stabilize the frame. That is an idea I am currently thinking over for a super lightweight and simple frame.
Even diagonal bracing similar to the ones found on the Darwin, done with tensioned Dyneema should result in more stiffness.

Or, if you can sacrifice some modularity, you could also replace the aluminium extrusions with bog standard square profile ones from the home improvement store. Some even come with center marks on the sides for drilling. They should be commonly available to about everyone, a lot cheaper and about as stiff as the t-slot ones.

The picture links are dead

Any advances on this? I have a fairly tall delta printer and no easy source of aluminum extrusions.. This would be particularly interesting to me in that certain parts could be directly integrated into the build, rather than bolting them to extrusions.