- connecting past and future -

Archive for the ‘fablab’ Category

Hexapod in the press

without comments

IEEE Pervasive Computing magazine recently had an article featuring the hexapod..

Written by admin

February 18th, 2012 at 2:36 pm

TechFest 2012

without comments

Together with PLANETART and FullFormFreedom invited to show some of the robots at the TechFest 2012 festival in Mumbai, India.


Written by admin

January 22nd, 2012 at 12:31 am

MakerBot Thing-O-Matic

without comments

Ok, one could argue that it is cheeting, but before the PrinterPrinter is ready, I thought of taking a shortcut by building a makerbot first. The Thing-O-Matic works right out of the box (and about 200 m3x16 socket screws later):

Written by admin

March 21st, 2011 at 4:05 pm

PrinterPrinter @ Refab session

with one comment

During Fab6, the international fablab conference, the printerprinter has been presented during a session on ‘refab’ at the fablab Amersfoort. Results of the workshop, presentation slides etc. can be found on the wiki. The workshop resulted in a set of ‘house rules’ for refab in fablabs (but they equally apply to work in whatever lab you’re making stuff):

  • bring your own stuff, take your own waste
  • be efficient with your design and materials (limit your waste)
  • be efficient with your tools (hand tools are handy too)
  • make it modular, upgradable, repairable, recyclable!
  • make sure you know the material you are working with
  • contribute your errors and mistakes (document!)
  • mark what your thing’s made of (document!)
  • reuse what you can find!
  • build stuff to last

Written by admin

September 13th, 2010 at 2:33 pm

Posted in fablab,fabrication

Tagged with ,

Repstrap using HP840c’s

with one comment

With the StarLC10 matrix repstrap ending up as pizza plotter, a start has been made building a new 3D machine using only the parts of 3 HP deskjet 840c printers.


The printers contain many useful parts, the casing is made almost entirely out of ABS (printing material, hurray!)


Especially the sliders come in nice, rectangular frame’s which make mounting them easy. For the first tests a frame has been assembled using standard aluminum X-beams (boikon).


Using the cleaner bin (which normally brushes and seals the cartridge) a Z-axis has been made which has at least 8cm of travel:

r0013947 r0013958 r0013959

Arduino-based control of a DC motor using incremental encoder and PID is not much more difficult than controlling a stepper motor. See the wiki for full documentation.

[sourcecode language="c"]
void setup(){
int oldposition;
attachInterrupt(0, encoder, RISING);
while(position!=oldposition) {
void loop(){
float setpoint = 1000+400*sin(n/200)+200*sin(n/50); // 1890 positions
float error = setpoint – position;
setMotor((int)(limit((Kp*error + Kd*(error-oldError) + limit((Ki*sumError),antiWindup)),255)));
oldError = error;
void encoder() {// encoder service routine
if (digitalRead(4)>0) position++;
else position–;
void setMotor(int value){ // set PWM of motor, value may range from -255 to 255
if (value>=0) {
if (value<0) {

next up: including the extruder head, interfacing, making connections, shredding abs, making a scrap-2-3mm-wire extruder system, etc.. etc…

Written by admin

August 22nd, 2010 at 9:28 pm

Posted in fablab,fabrication

Tagged with ,

18-servo walker

with 3 comments

New A4 flat-kit design: a 18-servo 6-legged hexapod. It does not walk (yet) but servo control works and the mechanics look good:


All parts can be cut from a single A4 sheet of acrylic glass. Other parts include 18 miniature servo’s (selling for 5EUR as TopLine mini-servo ES-05 JR (at Conrad Electronic)) and an ottantotto board


The servo’s are controlled using a simple soft-servo library (for the timebeing) which can be found here.

Written by admin

August 22nd, 2010 at 8:53 pm

Posted in Arduino,fablab,fabrication,Robotics

Tagged with

Flatpack Walker with Sensor

without comments

As presented for the first time on FAB6, using a Sharp gp2y0a02 sensor the flatpack-walker just got a little bit smarter. As long as the distance sensor is triggered, the robot walks backwards.


The code is still pretty simple:

[sourcecode language="c"]
// avoid obstacles using Sharp G2P sensor on Analog 0
#include <Servo.h>
Servo frontservo,backservo;
char walkpattern[] = {60,100,100,100,100,60,60,60};
void setup(){
void loop(){
for(int n=0;n<4;n++){

Written by admin

August 22nd, 2010 at 7:40 pm

Posted in Arduino,fablab,Robotics

Tagged with

Spark Creativity? Don’t get out of the box. Design your own box and think from within!

without comments

For years I have dismantled old printers, clocks, radio’s, disk drives, got all the screws out and sorted them in neat little boxes, taking the plastic gears, bearings, etc.. and storing them all separately. This all resulted in a workshop with a large amount of boxes, drawers, cabinets, all filled with a wealth of electronic parts, gears, motors. Strangely enough, coming to think of it, I have hardly ever used most of these parts later on. They always -just-don’t-fit-entirely- but ‘will do as placeholder for now’.

I’m currently in a luxury position that the projects I build are being paid by other people. In order to save time, I order parts from online suppliers and hardly touch what’s left on my shelves. It is quite risky to use salvaged parts in production, you can never be sure that you can order replacement parts within a day. (Instead you have to go out on a Ebay-hunt for that one specific type of plotter you remember taking these nice gears out)

As engineer your problem solving space is mostly defined by performance on the one hand and time + money on the other hand. Technically and creatively, virtually no boundaries apply. Just get the best parts you can find for sum X and build your machine in time Y. Interesting observation is that these constraints do not go well with hobby work (which most of the time is lacking both time and money) Perhaps a good definition for distinction between hobby and professional: lack of deadlines?

Recently I started giving physical programming courses in a Creative Technology bachelor track. The assignments that spark most creativity are the assignments with on the one hand a clear goal (make something that does this or that) and on the other hand a severe limitation (but use only these parts). One nice example that yielded most creative solutions this year was designing a drawing application using Processing/Arduino. The only input that could be used for this drawing application was one light-sensor (LDR) connected to the Arduino. No other inputs could be used. At first students were complaining. It could not be done. It was way to limited. What? was I really serious? There must have been a mistake in the assignment-description… and so on. After a while they accepted the challenge and were getting into it. I have never seen so many different outcomes of one (simple) assignment that work all differently. Not that they were all good or usable, but all solutions were different, working, creative and most important of all, fun. This assignment worked by far better than ‘make a nice animation using shapes and images’ – which leaves far more room for creativity (you would think so) … but does not spark imagination at all, apparently.

picture-2 picture-3 picture-5

Some time ago the spark of Personal Fabrication, or desktop fabrication caught me too. I need a 3D printer! Yes. I wants one! I started out searching and found most of the designs on internet that are neatly listed in Make Volume 21. Then I immediately hit the engineering boundaries of optimization. If you want a really useful machine, you need the fastest, most accurate and preferably not too expensive machine you can get. If you want CNC on your desktop, you need a system with backlash-free trapezoid spindle rods with ball-nuts. Nice linear sliders.. Oh. and standard Nema23 size stepper motors with micro step control. The easiest way is to order a kit, or get quality parts at a distributor – and after some 500$ to 1000$ you are the proud owner of a machine that will work for you..

I had three old dot-matrix printers lying around, which contains sliding rods, timing belts, stepper motors.. the basic ingredients for a Cartesian (XYZ) robot. My first approach was the classical one. Dismantle the printers, get the sliding rods out, salvage the bearings and stepper motor, start designing a base plate with nice brackets to mount the rods, get correct gears, get better stepper motors. I started reading on internet again. How somebody started out with printer parts, but regretted this, because he should have bought linear bearings from the start. How someone tried to use threaded m5 rod as spindle drive and failed miserably. No-one was even THINKING of using the crappy 50-step low resolution stepper motors that normally inhabit old StarLC-10′s, but immediately got themselves a decent set of some 400-step hybrid motors.

This project was going nowhere. I constantly found myself optimizing the design, trying to think of better parts, more accuracy, order good parts, order a kit. Perhaps abandon the printer junk completely. Hours on the project were mostly spent surfing internet, finding other CNC DIY designs on Instructables and maker’s blogs, browsing through kit distributors’ sites and manufacturers of spindle rods. All bouncing back and forth between the engineering criteria of performance vs. time and money.

Then I chose to severely limit my options. I was going to make a repstrap (3D fabricator) only from printer parts and salvaged computer Junk. No spindle rods, expensive bearings or high qualtiy stepper motors. All of a sudden the goals and the problem-solving space were very clear, and the project took off like a rocket.

The initial steps were simple. I took three printer head mechanisms and bolted them together to form a two-axis machine. I cut the printer-driver boards to leave only the stepper motor drivers, used the printer’s power supply and took out an Arduino board to control the steppers. Key ingredient here was that I choose to leave intact as much of the original mechanisms as possible, saving me from the tedious work of making high-precision bearing mounts etc.

The initial tests were very disappointing. The resolution of the print-head was way too crude, and, above all, the stepper motors were by far not powerful enough to reliably power the construction. I was on the verge of abandoning the project again. Then creativity really kicked in. Using the gears that were normally used for driving the paper rolls, I made an extra gear reduction in the printer-head mechanism. With the rods available from various clamp rolls, I made an axle synchronizing the movement left and right. The optical interruption switches were retrofitted as end-stop detectors.

r0013736 r0013686 r0013691

So.. the project is still underway. Conclusions till so far: store (and use) mechanisms, not just parts. Try to use as much as possible from original mechanisms as possible. This will save time, money – and also important – saves waist!

Conclusion number two: if you want to spark your creativity, severely restrict your problem-solving space. It sounds counter-intuitive, but it works!

Written by admin

May 26th, 2010 at 8:53 pm

nylon extruder #2

without comments

nylon1 nylon2 nylon3

After the promising results with the first nylon extruder, we (woutzweers and I) made a new nozzle following reprap extruder size (so a 3.5 mm feed hole and a 0.5 mm extrusion hole). The extruder head consist of a small piece of brass with two holes placed at 45degree angle. The soldering iron has been mounted in the clamp of a milling machine, everything in the following movie is hand-operated.

Written by admin

February 13th, 2010 at 7:45 pm

Posted in fablab,fabrication

Tagged with , ,

Getting a HP7585B plotter to work

with 2 comments

r0013747 I recently got an abandoned HP7585B plotter. Probably over 30 years old. However, perhaps fast and accurate enough to do some serious fabrication? Why not add a cutter wheel and cut foil, or fabric?

Anyway, first hurdle to take is to get the think working. Searching for drivers on internet resulted in two options: 1: winline ( or 2: autocad/autodesk built in..

I tried to connect it using a normal serial 1:1 cable. No result. Self test of the plotter worked. Searching HP’s sites, some old re-converted pages came up showing a different cable mark-up. Why didn’t anyone just used one type of serial cable in those days? Oh wel.. HP gave a pinout on their site Unfortunately, the conversion to modern web-site instead of 15years-ago-ascii in courier had crippled the ascii cable pinout.. Browsing on a different site showed why the pipes were included (this is how it looked: (unfortunately the spaces get crippled)


Now after this connection has been made, the printer can be controlled using HPGL. On the HP site the following string is used to test the plotter connection:


which will draw a diagonal line from 0,0 to 1000,1000. After making a connection using hyperterm (9600bd, 8 bits, 1 stopbit, no parity, hardware handshaking) and copying this string the machine worked! Next up: hpgl control from processing.

Written by admin

February 12th, 2010 at 10:44 pm

Posted in fablab,fabrication

Tagged with