How do you build robots?

Some people asked me on methods to design a robot beforehand (sort of drawing table) and then just building the real model in one run.

For truth to tell, I don't know about an efficient way to do this (let alone a way to do so that makes fun also - one could use LDraw, of course, but "Ldrawing" without a real model on your table is likely to turn out a complete drudgery):
as for me, I commonly start with a rough idea only and then build away, frequently changing the design while the robot comes gradually into being.

How about you?
What is your procedure in building NXT robots?

Matthias Paul


Anonymous said…
I usually generate a 3D model of the planned robot in my brain when I'm trying to get to sleep. :P

But yeah, the only pre-building plans I do are in my head. I mostly just experiment around with the pieces until I hit on a good mechanism. Occasionally I make a very rough sketch of the planned robot, and design it from there.

I usually see something out-and-about and want to duplicate it or incorporate it into a bot.

I have two extremes - (1) I sketch and plan the shape and concept of the robot and write out a rough flowchart for how the bot will be programmed, or (2) I just start snapping pieces together and try to make it look good.

Method 2 has come back to bite me in the past where I designed something that wasn't really easy to program. For example, when I use the Sound sensor I usually have to tweak the volume level because I place the sensor too close to the motors.

Right now, I've got something I'm working on - I want to post it on the NXTLOG when it starts up. This design was done using Method 1. I described what I wanted the bot to do, some sketches of its shape came next, along with a written description of how the program will possibly work.

Anonymous said…
I usually plan the basics in my head, then let it evolve as I go. Frequently with input from my son (who is almost 5).

Sometimes I program the bots, sometimes I just take some pictures and tear it down and build a new one.
Andy said…
Checkout my new blog!
I have not posted any thing yet but it will come more soon.
Here`s the link:
Anonymous said…
I just start building from an idea, just tinkering with a pile of Lego is more fun than working with Ldraw.
Anonymous said…
No NXT yet, but the aproach I use for RCX bots is preaty simple:
-Set requirments.
-Start building with the requirments in mind
-make changes in order to achieve requirments

Programing preaty much follows the same patern, but with the word "program" put in wherever "build" build appears.


Nice site.
Anonymous said…
I find that I almost always have to build my thing twice to get it perfect. The first time, when I'm trying to find the right way to make something work, the module comes out fragile and doesn't look nice. So then I do it a second time, this time making the module stronger and nicer looking, since I don't have to spend my time figuring out how to make it work.

Laurence said…
I don't do any planning before building, except in my head. When I build, I constantly refine what I'm building. For anything complex I'll typically rebuild several components many times. Sometimes I'll even start over from scratch. I've found that my creations tend to go through an arc of complexity: they'll start out simple, but not work as well as I'd like, then I'll add complexity until they behave the way I'd like them to, and then I'll find a way to simplify them while maintaining the correct behavior.
Anonymous said…
I usually have only vague ideas of what I want to achieve; and I do most of the building and testing with actual Lego pieces - whether it be robotic (not that much, actually) or anything else.

The only 3D plans I'd have beforehand would just (blurry) images in my mind. Most of the time I wouldn't bother to create plans afterwards either.

I know people who'd create their whole model in LDraw beforehand; but that's definitely not something for me
Anonymous said…
I think CAD (LDraw) is nice, if you're going to build something, and you don't want to mess up, like drilling holes in the wrong place.

However, building with actual LEGO pieces is as easy to do in the real world, as drawing in the virtual world.

I often make LDraw drawings of my robots, just so I can rebuild them later, if I want.

Unknown said…
As for me, I consider the process of the LEGO model coming into being while experimenting with real tenable parts extremely pleasant - this kind of playing with things makes "LEGO time" so satisfying, adressing multiple senses (seeing, touching, feeling, sometimes even tasting ;-) ).
Hence I'd not do "pre-LDrawing" for the world - that would take away much of the fun of free improvisation and flow of imagination.

There are other constraints, too (though of minor relevance): my models tend to change extremely from initial idea to the end result which would make LDrawing pretty costly; furthermore often difficulties, impossibilities or need for optimization are hardly to be foreseen with a design model and pop up with the built robot yet.

Yet, I have taken up the habit of LDrawing my robots for documentation reasons and possibility of rebuilding them (for the museums, once I'm famous...).

After all, that's just the way my creative process works (and not with LEGO models only) - might be completely different for other people.

Matthias Paul
Andy said…
Now I have updated my blog so if anyone wants to check it out just go to:

BTW it was this blog that inspired me to make my own blog!
Thank you very much The NXT Step team!!

Anonymous said…
I start off with a general idea of what I want to build and how it is supposed to work. I then start building and after inevitably running into trouble, I change my design step by step. Ultimately I end up with something completely different than what I had planned.

I do have an excuse for my relative incompetence - I grew up with studded lego. I'm new to this new type of studless construction, but I'm learning.
Brian Davis said…
I usually define the problem and think up a solution, but I don't draw it out ahead of time. I build; as others have said, LEGO is so prototype-friendly that it's natural to try first for me. I almost always try to focus on what I think is the toughest aspect, mechnically, of a project first. Once I have a handle on that, I can start thinking of the other aspects. Redesigns happen at all stages, with the "completed" first attempt almost always stripped down and completely rebuilt once I have it working "the first time".

Oh, and when do I mentally design? Almost always, either long car rides or mowing the lawn :-).

Brian Davis
Ethan Steckmann said…
Hi all when I am going to biuld a robot I all was look at brickshelf first and get a basic idea than i dream about it all night and form a plan of action and divide and concer.

just my ideas
Ethan MindstormMaster1
Anonymous said…
I just recently got my NXT. . . one of the motors was cracked, so Lego is shipping a replacement, so I'm not to be bothered.

Anyway, I've got a question for all you guys that make LDraw files for your robots AFTER you build the robots in real life.

How do you do it? I mean, that seems really difficult with complex robots.

Anyway, I always just come up with a general idea before I build my bots. Usually, I'll have a plan for how the mechanics will work out, but I just fine tune the structure and everything as I go.

Whenever I get any good inventions made, I'll post them on my website at
cmsnxt said…
Well I normaly have a concept of what I want to build but sometimes I try building some thing and end up building something completley different some how better than what I was thinking of .
Anonymous said…
I just think of somthing I want to build and get a basic concept of what I want and need it to look like and I go from there.
although I do get dissapointed when I somtimes find out I dont have enough pieces to build what I want.
One bot in particular that I have rebuilt at least 20 times is a navigator robot. each time strengthing it and refining it.

Popular Posts