Monday, June 27, 2011

Inspiration comes

My first attempt to a system frame was a flat based polyhedron, made out of M8 threaded rod.  The Z-axis was made from plywood glued to M8 nuts on the threaded rod of the frame, configured as a Sarrus Linkage.  Problems were plentiful:
  • The M8 rod was not rigid enough to handle the sideways forces applied to it by the Z-axis.  This flexibility created play in the linkage, and the resulting platform was not stable enough.
  • The M8 nuts had too much play on the threads of the rods.
  • Mounting the nuts were also problematic, and my temporary solution turned out to be very temporary indeed.  Metal and wood have different expansion curves, and the glue simply broke off pieces of the plywood.
It was simply put, nuts.

I have since explored electronics options, and decided to go with the RAMPS board for now.  My intention is to get a working Rep Strap as soon as possible in order to start working on Morgan 1 - a potential Gada prize candidate.

Next up however is the X-Y axis implementation.  The RepStrap should be a close approximation of the completed RepRap, since it will probably replace parts of its self until it is transformed into Morgan 1.

Morgan will return to the Darwin-like top mounted x and y axis format.  The main reason for this is for print quality.  The enemy of fast controlled movement is momentum.  You want to reduce the weight of the X and Y axis as much as possible.  Mendel did this by removing the need for the Y axis to move around the heavy stepper motor for the X-axis, but as build sizes increase, this advantage will diminish.

The plan is to implement the X and Y axis using two stepper motors mounted off axis, while keeping expensive rods and endless belts out of the equation.

Total cost target for this section: R800.00 ( US$ 115.00 )

More info to follow soon...

Saturday, June 18, 2011

Framework halted

So, plan B was crazy - and will not work for the same reasons plan A did not work.

In the end, what we need is living hinges, and that will be the first thing I shall work on when I have some system I can print with.  Chicken and Egg.

At the moment I am focussing on the electronics.  Arduino is working, and I am looking into the shields for making the motors turn.  Something that intrigues me is the power of the Arduino Mega, and how we are underusing it's capabilities.  It should be possible to build a shield that can control the stepper motors with lower cost electronics than the pololu stepper drivers.

Looking into ways to modify either the "orginal" 5D code or the Teacup firmware to make this possible.  The standard arduino stepper code available at the moment is "locking" code, preventing the system from doing anything else while the stepper is running... not ideal.  It will have to be interrupt driven.

At the end of this experiment I would love to have a new shield, at lower cost - and single sided that will have the same functionality at the current RAMPS 1.3  board.  Watch this space...

Monday, June 13, 2011

Science for Dummies


So, what have I learnt so far?

  1. Sarrus' linkage looks extremely good on paper and You Tube
  2. Stability depends on the play in the hinge and the stiffness of the support plates.
  3. There has to be a more efficient way to do this, material and space wise.
What I attempted was to have the linkage extend from under the base, and attach to the frame of the unit, using the support beams as attachment points.  The hinges are made from m8 nuts, spaced on the m8 support beam.  This provides stability in that the play can be eliminated by off setting every second nut.  It pushed the thread closer on both sides.

The problem came in the arrangement of hinges.  In order to save space, I opted for under platform hinges.  This limits the placement of the hinges because they cannot occupy the same space.  Additionally the support beams are not rigid enough by any stretch of the imagination, being designed to resist pulling forces, but not worth much in terms of lateral support...

It will be extremely hard to have the stability you need over the whole platform without placing the plates on the outside and on all sides, making it a bit ungainly.

Looking for plan B... and I think I have an idea.  A crazy one perhaps

Friday, June 10, 2011


The guys at are FAST!

Fired it up, working out the box with Ubuntu 11.04.  Absolutely painless.
My Arduino "Hello World" was changing the steady blink on the pin 13 LED to a fade-in, fade-out blink.  It is almost too easy.  My Eprom eraser and programmer will miss me...

Thursday, June 9, 2011

Vitamins are good for you - in moderation. Elements are always good.

It is good to get my mind working on a project again.

While trying to create a stucture that will be possible to easily recreate by building it in the repstrap, I am having a bit of a mind shift.  While the idea is to get rid of all "vitamins" (non reprapable components) there is no way to get rid of it all.  Now, given that some material are easilly available, and cheaply, at exceptional quality all over the world... Some vitamins are not cost effective to print.  If it is more expensive to print an inferior replacement for a freely available part (like a steel nut) then it is just a waste, no matter how you approach the matter.  Therefore these will be called "Elements" in the Morgan project henceforth.

element (n): a basic building block, freely available in some form (standards) word-wide at very low cost.

So, back to building business.

Morgan will incorporate as little as possible vitamins, but will make use of elements in order to create the hard wearing components.  Structural components will be built, but interfaces can be placed.

Just did some maths for the Z-axis mechanics.  Elements sure are going to make things work out very nicely... on paper anyway.

Wednesday, June 8, 2011

Collecting building materials

Africa is a tough country... especially if you are looking for speciality items like the robotics we need for a RepRap.

First Hurdle: Stepper motors.  I have a couple, but all of them have very low step resolution. All but two, and one has an inconvenient sized and pitched timer belt pulley forged onto the shaft...

In the end I had to go to down-town Johannesburg (Mantech electronics) to get decent steppers.  Not even the mainstream electronics suppliers in Midrand had anything useful to me.

Second Hurdle: Open ended timing belt.  For Morgan's POC I need a single timingbelt of around 3 meter length. Almost impossible to find apparently.  Finally found a company in Kempton park that could cut me a piece.

Third Hurdle: Electronics!  Even though I could make all the boards I need here, the cost of building it in small scale would be more just in materials than importing manufactured boards from Europe or USA.

Going to start with Arduino Mega that I managed to source from the local Arduino agent, and prototype electronics.  Now I am just waiting for delivery.
I am terribly excited about this platform.  I programmed since an early age, and used to program for DOS and microprocessors in C and C++.  Cannot wait to be able to dust off my C knowledge again.  Morgan depends on it.

Tomorrow morning I will start building the frame, and hopefully I'll get some work in progress pictures as I go along.

Sunday, June 5, 2011


Today I started to play with OpenSCAD, to get a feel of what can be done, and how much time you need.  Initially you have to get your head around the parametric stuff, but it turned out to be quite powerful.

Here are the initial candidates for the Morgan extremities.
I am going to try to model each of the parts I have to "strap" for Morgan Zero in order to make the machine easier to systematically replace with built parts when I get it operational.  It also helps me to focus on the functionality of the parts to ensure that I have to hand-make as few as possible pieces to get to the working system I need.  These parts will not be the final parts in any way, and will not be published until I start building in plastic...

More news later - I need some materials

Saturday, June 4, 2011

It starts

It starts...
Sent via my BlackBerry

Evolution of the Species

Evolution happens when you keep mating specimens, and keep removing the most unsuitable offspring from the gene pool.  Survival of the fittest.

Mutation helps to keep the pool diverse enough for the off chance that something else might be the answer to the problem.

What I am going to explore is the possibility that a cross-over between different successful rapid prototypers, with a bit of mutation (fuelled by my ignorance) could deliver a next generation of self replicating rapid prototype modellers.

Adding to my adventure, is that my birth county and country domicile do not cater for what I am about to attempt.  We do not have the choice of material from filament right though to "vitamins" that others have.  What makes this exiting is that this project will most likely be forced towards the Gada grand prize specification exactly because of this.

This is the uneasy birth of Morgan.

It will be a RepRap system, with build envelope of at least 300 x 300 x 200mm and will comprise of 80 to 90% printed components eventually.  Electronics will be simple, and as affordable as possible.  Should have at least 3 different build materials, with automatic material selection - and one should be an easily removable support material.

It should be able to build all its plastic parts completely unattended, and possibly its electronics as well.

Let's say its a 5 year plan...