Magnetic ball joints on our printer

We are proud to show some photo's of our working solution for making very simple ball joints for our printer. A few months ago Kees came up with the idea of using magnets to pull ball joints in a ball seat. We have done some dry runs before but did not print because of the lack of an extruder. Because of us both being busy with our work we did not make as fast progress as we would have liked.

The big plus about this joint is the very limited number of parts that are used for this. You need a base (carriage or effector plate) for the seat, a steel ball (ball from a bearing for example),a strong rare earth magnet and a carbon fibre rod.

Furthermore there is no play because of the magnet pulling the ball in the seat, always in the same direction. Another advantage is the big angle that can be reached.

The making:
The rods were made by drilling a hole in the balls (no accuracy needed for this) and gluing 2 balls on the ends of the carbon fibre rod. The important thing is that all the balls on the rods have the have the same center dimension. By making a simple rig for fixing the balls while the glue sets this is relatively easily done.

The effector plate and the carriages that were made for this printer are made with an SLS machine because we did not have a printer ourselves at the time. I'm positive though that this can be made with the normal FDM printing techniques. That's the next goal for us... Making another effector plate and carriages because one printer is not enough for 2 people.

Below some pictures, enjoy!

ball joint with steel ball and magnet

ball joint with steel ball and magnet


It's alive!

This week I've spent my evenings getting the printer to run. The problem is that I want to print above all other stuff that needs to be done before you should be printing. Things like getting the homing to work (no end switches), a heated bed (we don't have a heated bed), leveling the bottom with respect to the x-y movement. The way to quick results is done in buurman-en-buurman style.

Yesterday I tried to use a big round mirror because I though the multiplex base was not level enough. Also with the mirror I could not get the settings right. When moving in the x-y plane the height from the nozzle should stay at the same height. I measured a difference of 3 mm between the middle (0,0) and the end of the working area... Hmmmmmm... I even thought that the mirror was not level. So I went to the hardware store to get some 15x15 cm mirror tiles... They are level. No doubt about that. But the error was still there. (then I double checked the round mirror, and lo and behold. It's as straight as an arrow).

After some measuring and searching I found that the dimensions of the machine that are needed for making the calculations were off. We used a slightly wrong dimension. After updating the firmware a 7 cm movement of the machine was actually 7cm. Further more all was (reasonable) level. Below some visual proof of the result:

Printing an internal spur gear in layers of 0.10 mm

20130526-000108.jpgThe result when I stopped the printer because the rest would take another hour:


A movie of the printer in progress. The ticking you hear is the filament coil bouncing against a column:

The first filament!

Last thursday I worked with Kees on our 3D printer. It was long overdue because I have been busy working... Until wednesday we had a machine that was moving. We could give some commands, and the machine would move to the designated coordinate.

A few weeks back I had ordered an extruder because making one ourselves would take too long. Thursday we connected it to the electronics and after some modifications in the firmware we could see the extruder nozzle heating up, we could see the temperature, and the stepper motor driving the extruder moved. We only lacked filament.

A friend of Kees had an extruder and 3mm PLA material. So with this we could try to extrude. See below the result... Next week we'll get the steps per mm filament correct and I hope we'll be able to print something then more sophisticated. We're nearly there.


SVG to G-code... "hello world" !

Today I wanted to print more with doing less work. Making the previous program took quite some time and concentration for making some straight lines. Nothing Fancy about that.

I wanted to have a program where you can print (plot) fancy work. In a way that drawings made by my kids (when they will do so in the future) can be plotted. I searched and found an Inkscape extension (I frequently use Inkscape). It didn't work for me.

What I actually wanted was to get all the points from a line and convert them to coordinates in G-code. Basically it means instructing a machine to go to point X... Y...

Here is how I came from this:

to this:

I needed to investigate the .svg file to see if I could easily change the coordinates. I opened Inkscape and first drew a simple line consisting of a few points. I found out that (0,0) is the upper left corner of the canvas. I drew the line in millimeters but the .svg file gives coordinates in 'px' (pixels). After some calculating I decided to scale the line in Inkscape so my coordinates would be in mm. (90 dpi (pixels per inch) gives 28,222 pixels per mm so the line needed to be scaled to 28,2222 percent of the original size)

What is x and y in Inkscape translates to y and x in the .svg file... good to know.

Now the file needs to be opened in a text editor. I highly recommend Sublime Text 2 for this since it is a very powerful editor (I'll come to that later on).

Here are the coordinates:

Delete the rest of the file since we don't need it. Also in my case the first coordinate was written as 0,-1.484375e-5. I changed it to 0,0

Now here's what is so shizzle about Sublime Text... Select one space, and with a shortcut you can select ALL spaces. If you then press "backspace" and after that "enter" and type "G1 Y" you have changed all spaces in "G1 Y"!!! Timesaver!!!

Select all spaces:

Change all spaces:

The same goes for the comma, select "," and change all to " X"

Finish the textfile (add other G-code for putting the pen on paper etc) and it is ready to use for printing...

First 3D printer output

Today Kees brought the 3D printer. After setting the dimensions to the correct sizes and recompiling the firmware it moved! "It's ALIVE!"

Accuracy needs to be improved, I do not have homing switches mounted yet so I guestimated the zero positions.

The first question I got from my kids after seeing the head move op 10 cm (G1 X0 Y0 Z100 F9000) was: "What else does it do?"

I took my bloc note, made a sketch and manually programmed the G-codes (with 2 hyper boys next to me testing the limits of my concentration) to show them what this contraption can do more.

Below is the result. Programmed in 15 minutes (with in-between-tests), printed in approx. 4 seconds...

3D printer besturing

Kees en ik zijn een 3D printer aan het bouwen en hebben de mechanische delen bijna klaar (geleiding, frame, aandrijving).

Ik ben afgelopen zaterdag aan de gang gegaan met de besturing van de 3D printer. Het doel voor mij was om beweging uit één stappenmotor te krijgen. Het is zaterdag wel wat laat geworden.

  • Ik ben begonnen met connectors pakken, draadjes knippen, soldeerbout aan, een stukkie solderen... Voor mekaar!!!
  • Daarna moest de firmware worden gecompileerd (de brains van het kaartje). Dit was met wat zoeken op goed te doen...
  • Ook het uitzoeken van software die 3D modellen in plakjes hakt en vertaald naar G-codes (instructietaal zoals ook CNC draai- en freesautomaten worden aangestuurd) was best snel geregeld...
  • Voeding uit een oude adapter was in een mum van tijd gefixed...

Het lukte alleen niet om de motor in beweging te krijgen. Laat op de avond (eigenlijk heel vroeg de volgende morgen) kwam ik erachter dat de voeding die ik gebruikte niet voldoende was. Ik had iets nodig dat 12V kon leveren met een minimale stroom van 5A... Ga dan 's avonds laat nog maar eens in de zak met oude adapters en transformatoren lopen trekken.

Toen ik Pug ging uitlaten zag ik de acculader... effe kijken en.... 5 Ampére... Ajetoo!.

Zondag was het toen binnen 15 minuten geregeld. Alles werkt. en achteraf gezien viel het me 100% mee.

Volgende stap is de 3 assen aansluiten en "kijken wat ie doet..."

3D printer brains

3D printer stepper motor