onsdag den 22. maj 2013

Tensile test of RepRapped parts

This my first time on a blog, so bear with me.. :)

I have now build 2 repraps. The first one was technically  seen a RepStrap, on which I printed the second one. But on the second one I have been having quite some issues with delamination and warping of the parts.

 After having used a lot of time tewaking parameters, I came to the conclusion that I needed a objective method of testing my parts so that I could determine which parameters influenced the strength and in which way.

Some of the parameters is:
Temperature
Too low and the layers don't stick to each other, too high and you end up with a molten blob.

Width over thickness
The wider you print, the more the layers gets smashed together, but on the expense of details, as wider paths have bigger corner radius. But the narrower you print, the more you stretch the filament.

Consistency of filament supply
A bad quality hobbed bolt doesn't give consistent extrusion, as with filament with an uneven diameter.

Build environment
Draft can cause uneven or too fast cooling of the printed object.

So inspired by a post on the RepRap forum, suggesting testing parts with a bucket you slowly filled up with water, I made a test part, which was printed at different temperatures, with the sole purpose of destroying them again..

This was my setup:
  

With this setup the repeatability was surprisingly high, however filling stones in a bucket is rather time consuming, so i went in the workshop and made this spring loaded weight:



With the help of a digital caliper, I could measure the travel length up to the point where the part broke, and compare strength.
This was sort of ok, and I found out that parts printed at 230 degrees C, had approximately twice the strength than parts printed at 210 degrees C.

But it wasn't particular precise, and if anyone should attempt to repeat the experiment, it would be difficult to compare the results, of course you could translate the travel to a weight but the lack of precision makes it hard to compare smaller changes in the printing parameters.

So I went off to eBay and bought a load cell, which I hooked up to a PIC 12F675, which has a build in 10 bit A/D converter.
After getting dirty welding and grinding in the garage I got this setup:
 

For getting some nice output I made a small VB program, which draws a nice graph and displays current and max load.


























I haven't really done much testing yet on this new setup, but I have managed to establish a reference, which I can work from and compare, when I change parameters
My test part looks like this:



With these settings:
; generated by Slic3r 0.9.9-dev on 2013-05-22 at 00:23:46

; layer_height = 0.25
; perimeters = 2
; top_solid_layers = 3
; bottom_solid_layers = 3
; fill_density = 0.1
; perimeter_speed = 40
; infill_speed = 50
; travel_speed = 80
; nozzle_diameter = 0.35
; filament_diameter = 1.28379
; extrusion_multiplier = 1
; perimeters extrusion width = 0.50mm
; infill extrusion width = 0.50mm
; solid infill extrusion width = 0.50mm
; top infill extrusion width = 0.50mm
 
My parts break at around 48Kg, most of them on the middle of the part, where it has a 10x10mm cross section.
Of 12 parts tested so far, 1 broke at 38.4kg, 3 at  42-44 kg, the rest was in the range 48-52Kg, and a single high score at 55.7 kg.

Things to try from now on is to lover the E-steps, acetone vapor treatment, different temperatures, different speeds, different width over thickness. And hopefully I will end up with knowing why and when my printed parts fail.