possible next gen reprap design

[repositorium.sdum.uminho.pt]
looking at this article it seems like welding robot type reprap would be much more useful:

it can print objects larger than itself ergo an itself in full eliminating some or all parts of manual construction process
LOTS of work area, enough room for holding different toolheads, drill bits etc for automated changing
you can turn the toolhead in different positions, it doesnt have to be vertical
because of abundance or space you could have multiple robot hands working simultaneously - for example one could hold and turn the object while other deposits material

compared to a cartesian bot a welding robot is simply uber
so how about investigating that build possibility for next gen reprap(or the generation after that)

Edited 1 time(s). Last edit at 04/17/2008 01:09PM by r2kordmaa.

I read the paper. It's certainly a fascinating idea. I only see two problems.

1) the cost of the MIG wire to make something

2) the electric bill you'd get at the end of the month.

[20:41] er.. i think i failed to get the point across
[20:41] point one
[20:41] the robot arm would hold whatever toolhead you need
[20:41] plastic extruder or whatever

[edit]

point 2
the robot arm doesnt have to be the same as the one in the article one would build one according to the needs, smaller/bigger whatever so im not telling you have to have something like that on your desktop
the idea is to design a reprap machine like a robot arm

Edited 1 time(s). Last edit at 04/17/2008 01:46PM by r2kordmaa.

I'm not worried about the cost of electricity for the robot arm. You can always design a purpose built positioning system that would do that job much more energetically efficiently.

Even without changing the arm, though, the electricity that that would take woulb be nothing compared to what it would cost to run that MIG welder.

look here... the idea is not to build a multi kw consuming robot arm. the arm would be much smaller and driven by motors alike to ones used in darwin and a mig welder is just a toolhead you would use if you need to print in steel, you can put a plastic extruder on it or milling/drilling head or whatever even a pencil or a cat toy for all i care. idea is to achive reprap design capable to build objects bigger than itself, ability to access objects from side etc

"look here... the idea is not to build a multi kw consuming robot arm."

I never said it was. Mind, if we want to take self-replication to the point of having a reprap assemble other repraps we are most definitely going to have to look into making a general purpose pick and place machine. I'm already looking into the laser radar part of that design problem.

Well the other problem is that the the design now is fairly basic and because of that it is very simple to calibrate and adjust...where as a robot arm. If the first motor is off then the entire arm is off. Also the motor not only have to move the weight at the end of the head be it the extruder or a mig welder, they also have to move the weight of the arm and motor. So for now i think we should stick to a cartiesin bot (may be add rotation)


Now that being said I can see the advantage of a robot arm And something we can may be look in to AFTER Mendel( which it self is not discussed much). Those advantages being more complex design and the ability to make larger parts than it self. Right now even if the Darwin could "print" steel it could not do the length of one of the cross beams. So for now I think we should just keep it on our ever expanding To Do List.

"Right now even if the Darwin could "print" steel it could not do the length of one of the cross beams."

Sure it can. You merely have to print on a diagonal of a square printing table. That will let you print an element up to 40% longer than a cross beam.

Yeah, and as we pointed out before in other threads, the bot doesn't have to be a monolithic structure composed of a single piece. We can build it up out of many pieces to make a larger whole.

Also, the robotic arm has the disadvantage of loosing accuracy the further out it reaches so you have differing tolerances and accuracy zones for an object that could use the whole 'print volume' of a robotic arm.

Plus, my understanding is that controlling a robotic arm is much more complicated than controlling a simple cartesian positioning system. I could be wrong though. Viktor? Got anything on that score?

Demented

well of course the robotic arm is more complex to control and build therefore the next gen idea, im not saying we should go over to that design tomorrow. something like that is way in the future. the welding robots are quite widely used tho, i think its best to not to reinvent the bicycle we can always take over designs and software from the industrial robots. i bet its not too difficult to achieve accuracy tho, its all about calibration

... here [forums.reprap.org] (data in the link, images 4 and 5 in the paper) and in the previous posts we discussed the accuracy of a parallelized scara-robot in respect to a sequential setup as in a robot-arm or the scalability of some other systems (or search for "tripod" in the forum).

Viktor

its indeed highly accurate but its building height is limited. anywho im planning to do some r&d in the robot arm design direction and see what results i can come up with. first i need to get my darwin doneand functioning

This is an awesome paper! Thanks for finding it. Love the technology.

r2kordmaa Wrote:
-------------------------------------------------------
> well of course the robotic arm is more complex to
> control and build therefore the next gen idea, im
> not saying we should go over to that design
> tomorrow. something like that is way in the
> future. the welding robots are quite widely used
> tho, i think its best to not to reinvent the
> bicycle we can always take over designs and
> software from the industrial robots. i bet its not
> too difficult to achieve accuracy tho, its all
> about calibration

Robot arms are significantly harder to control and to keep stable. Think about it... you have a cantilevered load out a significant distance from the motor that's controlling it. In a cartesian robot, the load is usually supported in two or three points (on rails) that make sure it is aligned correctly. Cartesian robots are also usually cheaper because you don't need as powerful of motors or as heavy-duty drive systems.

Of course, that's not to say it can't or shouldn't be done! Building an arm-bot is on my (admittedly impossibly large) personal to-do list for after I get my RepRap done as well. That should make fabricating the necessary bits quite simple. And I think when you get five or six DOFs on an arm you begin to see some really cool possibilities - self assembly to complement self replication, for instance.

More stuff to look at!

I saw this [lasa.epfl.ch]
on hackaday.

It is a robot who can draw portraits of people. i think that the idea behind the humanoid arm doing the drawing is what we were after in this discussion. (and yes, robots are people too)

---

Jay