Full mechanical Reprap

Hello.

I'm trying to build a full mechanical Reprap.
Full means that there is no computer : programming is mechanic too !

My firsts drawings show that the strength of materials might be a problem.
I have to check it : I'll do it first in my computer (Bubuntu).

What "easy to use" free software do you prefer for CAD ?

Thanks.

There was some discussion a while back about something similar to this. There are some printable mechanical stepper motor designs and a small investigation into using water for logic circuits. Should be somewhere on http://www.3dreplicators.com/cgi-bin/cblog/index.php

I think "art of illusion" is the free CAD-like tool of choice around here.

I'd love to see your designs too. Not that I'd wanna program the thing mechanically, but you might come up with some novel solutions to existing mechanical problems while you're at it :-)

I keep finding AoI buggy beyond use, frequently.

I use Blender for all my mechanical designing. You can even setup animations to see how certain sections would work. It does however a high learnung curve, but once you grasp the basics and hotkeys, its a pleasure to work with.

Sounds very "steampunk" have you read the sci fi novel "The Difference Engine"

Methods of distributing "software" .. would the models would be recorded on punch carded, or vinyl

or as a halfway house, how about a reprap that was controlled via radio / analogue computing only

You know the first CNC machines used data stored on paper tape. This is the obvious storage technology for an all mechanical 3D printer.

Frank Davies

So could one make a genuine ultra-low-tech reprap (suitable for life after the twin apocalypse of economic collapse & swine flu)
powered by windmill, heated by firewood..?

McGyver could.

'sid

[www.youtube.com]



See everyone thinks McGyver can do anything with anything

Jaccard designed a loom that could weave a tapestry from instructions on punched cards. A three-dimensional object would probably require a lot more cords.

I actually am interested in something similar. I once looked down the rabbit hole into fluidic logic, which was popular a while ago but digital logic marched all over the idea. Now its only used in fighter jets because its resistant to EMP attacks.

The idea of embedding logic into the form of an object is gorgeous. It also means that it would be possible to actually "print your own circuit".

Post about how your going to go about the mechanical computing aspect!

I've actually had a lot of trouble with CAD applications. I've become proficient in blender, but its a pain to get constraints working properly. So for low tolerance parts, I've had issues with not being able to align things properly. Blender is very intuitive for some things, but it needs a lot of love for good cad work.
This is a link to precision blender modeling tutorial that I wish I had two months ago.

Hi Lawrence,

... for designing and milling complex and high precision parts for micro- and nano-mechanics (accuracies down to 10 nanometers ) i used Lightwave.

Actually you have a good entry with Truespace ( [www.caligari.com] ) - it's free and capable of much more than 'simply constructing'

Microsoft purchased Calligari and give Truespace away for free but it's only short time until they shut it completely down

Viktor

For more on punched card control systems and mechanical replicators see this:
[groups.google.com]


Fluidic logic is used in fighter jets? I've heard of it being used as part of a helicopter control system, and even missiles but not fighter jets. Though there have been fluidic jet engine control devices. Do you have the link to that information?


I've been looking into using fluidics too. Currently I'm designing a fluidic stepping motor that reprap could print and fits all reprap's requirements for a stepper motor.

The only problem I see with fluidics is speed, other fluidic stepping motors don't turn much faster than 20 rpm if I remember correctly. Also, the speed that signals travel through a fluidic circuit is limited by the speed of sound of the media used. To get faster fluidic circuits one could use liquids as the media, but the problem with liquids is that they can't be exhausted to the environment like air.

... very interesting are electro- or magneto-rheological fluids and maybe mixtures of oils and ferrofluids too - imagine magnetic stall- and direction-management or magnetic valves and oszillators for fluids ...

Viktor

There are some links on google about using fluidics as a backup flight control system, but nothing with details. Most of the comments come from a discussion I had a while back with a friend of mine who is an aerospace engineer. He was talking about non-linear dynamics and fluid flows with regards to aircraft control systems.

The link to trueSpace looks interesting. I've honestly not found a cad program that speaks to me yet. I've been using blender because its what I have the most experience with.

About the fluidic stepper motors, very interesting stuff! I wish I could take a look at it, so post your progress when you make some.

I've come to the conclusion that hardware research is difficult enough, and I want to take a side trip to setup some collaboration tools that are more useful to get people interested in projects so they can contribute. Forums and wikis are nice, but you'll notice that most developments happen by individual people who get something that works and then posts progress. Maybe like a sourceforge for hardware. I like things like Thingiverse, but it doesn't have tools specifically for development.
I want to see 3D spinning models, bug reports, change logs, model repository with history, etc... It would be especially useful for larger projects.

Sorry for the side track...

Hello.
The information is stored in screw-cam. The elementary screw-cam are aligned to make a long screw-cam. This gives more information than a unique cam.

The problem of the strength of materials is solved !
The nuts, that support the screw-cam, has its thread very deeply dug.
So we can have nuts near the cam followers !
The axe of the screw-cam will not bend !

Sorry for my enthusiasm ! I've solved the only problem that blocked me !
I'll continue in French (it's more easy for me) here :
[dev.forums.reprap.org]

I can see acouple problems with using cams for information storage,
1. We have to store a lot of information(especially if you want to store the movements for an entire print), the cams might end up being very very large.
2. Analog control systems need to be made to high tolerances, otherwise they will be inaccurate and produce inaccurate parts. It'd be much better to use something digital for information storage, as you don't need incredibly high tolerances. Paper tape, folding printed hole-cards, or even phonograph disks are ways we might do this.

I believe that if a totally mechanical reprap is going to be built, it's going to be very large and impractical, but then again the "work after nuclear/zombie/grey-goo apocalypse" factor is pretty cool.

You know, I believe that the earliest CNC machines used paper tape to store commands and patterns. I bet they did not have much computational brains, either.

fdavies

You need a method of translating a human interface device into cnc commands. Like a pencil that is attached to three record players, where an analog representation of the pencil in space/time is represented by the change of the grooves up , down or constant with respect to a maximum movement rate and a constant time progression.

You could use three records for the, x/y/z axis where playback would allow a duplicate of the work you performed ideally.

Problem with analog systems is usually the difficulty in developing systems that allow easy in system editing. You might consider the movie projector, where a film maker can 'splice' together a film with discrete images. Where instead of scenes, you use discrete motions strung together to form motion.

Quote
fdavies Wrote:
-------------------------------------------------------
> You know, I believe that the earliest CNC machines
> used paper tape to store commands and patterns. I
> bet they did not have much computational brains,
> either.
>
> fdavies

I wrote an overview on the old paper tape systems recently in the firmware section. it is part of this thread Third post down [ there does not seem to be a way to reference individual entries]

Currently, I am working on a reprap firmware variation (written in asm) which implements the older register style NC controllers of the late 1970s. These were based on counting registers.

More to the topic, the reason I want a reprap is so I can make some working models of the 1830s era mechanical computing devices that Charles Babbage developed. I have been able to locate a reasonable overview of how the analytical engine might have worked by the late Alan Bromley. Tim Robinson has been able to implement some of this using mechano parts. there is a pretty good video showing it working.

While history has not been too kind to Babbage, he was actually a rather important participant in the industrial revolution. A prolific writer on technology, his book on the 'economy of manufactures' has affected how factories are set up and operated efficiently. He is also the inventor of the screw-cutting lathe, which more than anything made the industrial revolution happen. Later, he developed the processes for injection molding. All attempts in the 19th century to build a computer...

-julie

Edited 1 time(s). Last edit at 02/11/2010 12:32AM by sheep.

This would seem relevant to this idea. The Clock of the Long Now.
It's a project to design and build a clock that will keep accurate time for the next 10,000 years.
Their current design uses a fully mechanical binary computer to avoid the accuracy problems associated with analogue states.
Complete CAD files were available, though I can't see them readily on the site now. I have a copy saved to my computer however.

---

--
Peter "Sci" Turpin
London, England

Provider of practical solutions.

(Sometimes stellifying Jupiter IS a practical solution)

In Strasbourg, there is a clock dating from the 1840s designed by Jean-Baptiste Schwilgué.

Schwilgué.was an optimist and is one of my personal heroes. The calendar has digits to run through 9999. "After that one can place the number 1 before the other digits," the designer was reported to have said.

But the real gem of this clock is a gear intended to measure the precession of the equinox, which will take some 28,000 years to complete one revolution. Now that is long term planning.

Some other features of this clock, which was the most famous clock in the world, until the completion of the Westminster clock that rings the Big Ben bell, are:

A cam based lunar calculator, Complex as anything Babbage invented.
An Easter calculator, which is table based and accurate for 1000s of years, Arguably one of the first mechanical computers.
The date calendar, contains a wheel of 365 slots, containing the name of the saint for that day. In leap years, the wheel splits open introducing a 366 slot for the leap day. It is also Y19K, Y2K and Y21K complaint correctly interpreting the 400 year rule.

This clock was copied in the 1940s in Denmark by Jens Olsen. The olsen clock is situated in a self contained climate controlled room, which seems to be a granite vault. Olsen was not quite the optimist of Schwilgué, as he used an arc sector for the 28,000 year gear.

Much of the clock of the long now, seems to be based on Schwilgué and Olsen's work.

-julie

[1] Strasbourg clock wiki entry
[2] Jens Olsen's Copenhagen clock wiki entry

Edited 1 time(s). Last edit at 02/11/2010 02:53PM by sheep.

Hi All,

I started a wiki review article on Actuator Fabrication that you might be interested in:
[www.reprap.org]

Many of you will be familiar with most of the information, but I would wager that even the most grizzled old replicators among you might find something new.

Please visit and add your own contributions!

-Matt