RobotArm first project ever

Does your laptop have a parallel port?

What computer language and do you want to use?

--
JC

Thank you for your reply JGCASEY

Computer languages:
Java, C#, Visual Basic 6, VB.NET, ...
I have a little bit basic knowledge of C++ too.
And I wouldn't be affraid to learn some new language as long as they
aren't too hard. (like assembler would be TOO hard for me I think...)

It only has USB. But if it's worth getting a converter for
parallel/serial ports I'll just buy one.

Hi there

Have you looked at the Basic Stamp from Parallax,
http://www.parallax.com . You can program this in PBasic which is pretty
close to VB. If you look on
the site you'll find some handy manufactured boards to do the
controling of your motors for starters.

You can control the Basic stamp from a serial port, I have before. You
even get the cable to talk to your "chip" with depending on what you
buy. The stamp is quite expensive, $50 but it is really eay to program
and nice for beginning and prototyping later. I currently use mine to
test first and then build the final product.

You haven't said how much you have to design/build yourself?

I'm designing my own arm currently, I'll let you know if I run into
anything unforseen.

Hope this help
 Fick

Take a look at these guys. They sell parts that allow you to design
your own robotics without wasting a lot of time on all those tiny
little details.

http://www.lynxmotion.com/Category.aspx?CategoryID '

-howy

bvandenbon@gmail.com wrote:

    --The thing I don't like about the lynxmotion kits is they are
terribly small and wimpy. I suppose once you know what you're doing you
could scale one up, yes?

--
        "Steamboat Ed" Haas         :  Better an early adapter      
        Hacking the Trailing Edge!  :  than an early adopter..
                   http://www.nmpproducts.com/intro.htm
                   ---Decks a-wash in a sea of words---

Hey Burner buddy ;-)

Scaling up is a pretty big problem because when you, for example, double
the height of your bot, you actually create 8 times the mass. Since all
your levers double in length as well, you need to generate *16* times
the power of the original.

A complete redesign is usually a much better approach.

Matthias

---
http://matthiasm.com/vehicles.html  for my BM '00-'03 art cars
http://www.robowerk.com/

    --Agreed. I went for Stamp (cuz I don't have a clue about circuitry
and their tutorial is first rate) and I'm now building big-ass "servos",
bwahaha..

--
        "Steamboat Ed" Haas         :  Better an early adapter      
        Hacking the Trailing Edge!  :  than an early adopter..
                   http://www.nmpproducts.com/intro.htm
                   ---Decks a-wash in a sea of words---

I made this one ( http://www.waycoolgear.com/er1/servogripper.jpg  ) using
the servo brackets from LynxMotion, by just connecting them directly to each
other. I "stole" the idea for using a hobby clamp for the gripper from
Gordon McComb's website:
http://www.budgetrobotics.com/shop/index.php?shop=1&cat7  .

I used a Phidgets 4 servo controller board to control the servos. With
Phidgets, you can write your own code in C++ or VB, on Windows or Linux, and
control them from a USB port.

I just used a couple of cheap L brackets from Home Depot and some scrap mdf
wood to mount the servo to it.

This is not a very good design because all of the servos are the same size
and strength -- this design requires at least one stronger servo to lift the
whole gripper assembly, especially with any payload. It was mostly just an
exercise to see how easy it was using Lynxmotion's "servo erector set". It
was a breeze!

I think 3 servos + brackets + phidgets board would probably have cost me
about $200 if I had bought everything new. Lynxmotion's new aluminum tubing
connectors make some really cool looking robotic arms! See this link for a
great photo of one of those:
http://www.lynxmotion.com/images/jpg/aarm05.jpg

I had this mounted to my ER1 robot in such a way that I didn't need to the
other degree of freedom a 4th servo would have provided:

Hope that helps some ...
JCD

Hi

I have this book:

"Build Your Own Humanoid Robots" by Karl Williams. It has three chapters on
building an arm, including VB 6 interface software, and if you want,
optional voice control.

The only problem is the microprocessor code is in PIC BASIC Pro - C is the
industry standard. A trial (*free* ) C compiler from Hi-Tech www.htsoft.com
will program the PIC16F84A no problems The Hi-Tech compiler is one of the
industry standard. Porting from PIC BASIC Pro to C will not be trivial for a
beginner, but an engineering student should work it out -it depends on how
much time you have.BASIC programming for microcontrollers is not recommended
for professionals or at university engineering level. The PIC BASIC Pro
software is expensive and outdated. As a last resort and you decide to
program in BASIC, buy a or build a BASIC Stamp from components available
from www.parallax.com. The code is similar to PIC BASIC Pro (you will have
to get a PB Pro manual to see which commands are similar to the PBASIC code
used for the BASIC Stamp).

The hex code is also supplied with the book, so you could use that 'out of
the box' if despearate with the Microchip MPLAB assemble/IDE at
www.microchip.com.

Also consider the book "123 PIC Microcontroller Experiments for the Evil
Genius" by Myke Predko -an excellent coverage of Hi-Tech C programming and
programming in Assembly ( the chip in this book is different to that used in
the book "Build Your Own Humanoid Robots" by Karl Williams, but with help
from others you should get around this).

If you live in the USA, Microchip will send you some FREE samples, and with
the Hi-TEch compiler, you will have free development tools. You can build
your own PIC programmer from plans off the Internet, or from tyhe book:
"programming Robot Controllers" by Myke Predko.

This beats paying USD $250 for the PIC BASIC Pro compiler, and the outdated
PIC Programmer in the book "Build Your Own Humanoid Robots" by Karl
Williams.



Cheers

|-]

Dale

I guess the "selling point" of PIC Basic is that it is BASIC Stamp
compatible, so if you come from that environment it's viable option.

Just to clarify a point: The standard Hi-Tech C compiler is not free.
They have a PICC Lite compiler that is free, and though for the 84A
there are no limitations, it's mostly because this is an enemic chip.
Otherwise the software has a 1K or 2K limit, and supports only a subset
of processors. Not sure I'd recommend the Lite version for professionals
or university labs either (the full version is fine), and the full
version is close to a thousand dollars. Anyway, just to put this in
perspective. If the lite version is enough for a project, then that's
great, but it isn't always the case. Many a project I've started where I
realized the lite/demo version was insufficient, and I had to fork over
money to complete it.

Replying to the OP through your post: I would seriously reconsider your
points 1 and 2 if you've not done this before. It's rather like: "1.
Take some art classes. 2. Paint the Mona Lisa." Of course, you could be
a mechanical engineer by trade, but you didn't mention it, so I'll just
point out that the physical construction of the robot should not be an
after-thought. A good base will go a long way to building a reliable
arm.

Another approach: Buy off-the-shelf parts if you can. Some systems are
more flexible than others. Though I haven't personally tried them,
Lynxmotion sells some generic brackets that you can assemble to make
various types of servo-based robots. If you don't like your first design
you can take it apart and rebuild it. Most standard Futaba and Hitec
servos will work with the brackets. You can also fashion your own
brackets out of aluminum channel, and build your arm out of those. The
benefit of using component parts is that you can change your mind
without having to rebuild from scratch.

-- Gordon


Dale T Stewart wrote:

Hello,
you can see my project of a robot arm:
http://www.geocities.com/gidesay/roboticarm.html
It seems to me very similar to your ideas.
I builded all parts myself (except servo motors
and chips of course! :-)
I use plywood in different thickness. There is one
of 3 mm for airplanes models that is as strong and rigid
as aluminium!
In the site there are details of mechanical construction,
electronic card assembly to control the servos from
a serial port, source programs for a Pic 16F628 microcontroller,
and finally a Delphi program to command arm from the pc.
Have fun

Gidesa

Hello,
you can see my project of a robot arm:
http://www.geocities.com/gidesay/roboticarm.html
It seems to me very similar to your ideas.
I builded all parts myself (except servo motors
and chips of course! :-)
I use plywood in different thickness. There is one
of 3 mm for airplanes models that is as strong and rigid
as aluminium!
In the site there are details of mechanical construction,
electronic card assembly to control the servos from
a serial port, source programs for a Pic 16F628 microcontroller,
and finally a Delphi program to command arm from the pc.
Have fun

Gidesa

great job on the wood arm!

Rich

First of all I would like to thank all of you!

The lynxmotion parts look very nice, but too expensive for me.

I am not sure what kind of material I will use for the arm. I think I
might go for wood after all. Light, strong, easy to get my hands on,
... It's something I can easely do on my own. I guess this will save me
a lot of money.

And I figured it's probably better to do the electronics first. Either
way I will need to know the sizes of the servos before designing the
construction parts of the arm.

I had a lot of nice replies of all of you. But the most helpful one of
all was the one of Gidesa.
(http://www.geocities.com/gidesay/roboticarm.html ) I must say your
website is really impressive. But I was stuck with a lot of small
questions on your diagram of the electronics.

Yesterday I took all my courage together and took your diagram to one
of my profs of "digital electronics and communication technology". He
explained me a lot. Turns out most of my questions like "why use a 18pF
component?, why the 240 ohm?, why use a polarized capacitor?". My prof
explained me that most of these details just come straight from the
manual. He showed me the manuals of the MAX232, PIC16F628 and LM317.
And indeed all my questions seemed to be answered in there.

So the most important lesson I learned here is: digital electronics is
not about small details like in analog electronics: capacitors, diodes,
etc... . But it's about making a selection of the chips and then
looking up the required smaller components that are necessary without
worrying too much about them.

My prof also told me I could use the programmers that are available at
school. He also warned me that I should order my chips before the
diodes and capacitors, because sometimes electronic shops sell
different types of MAX232 chips that might require different kind of
capacitors.

So what I am going to do now is compare prices and look for a board to
mount all my parts on. I checked some websites so far and this really
doesn't look that expensive at all. I guess total costs of the
electronics components will be 15€ at most.
PIC16F628: 5,35€
MAX232: 1,94€
LM317: 2x0,38€
a cheap LED: 0,119€

Does anybody know a good website for circuit boards for soldering my
components on? And maybe somebody knows a good website for servos?

Thank you for all the nice replies so far.

Bvanden,
thanks for your preference for my site! :-)
Indeed I completely agree with your professor.
I am only a novice in electronic, but I find that when using
IC many of the circuital details are just "written" in
datasheet. Of course in industrial projects the chips are choosen
with great attention, but in our simplified projects the main
rule is: I choose the chip more cheap and more easy to find
and use! :-)
In my circuit schemes many parts are really standard
sub-circuits taken from 16F628 or MAX232 or Lm317
datasheets.
The main problems I had in testing the circuit after
the solders. Every time sometimes was bad, but was
also difficult to discover what was bad!
Anyway I built these circuits with very simple materials,
a solder, a board and wire. And they work, no matter
of my really not perfect solders.
Here an hints: Maxim and also Microchip send samples
of their IC at no cost! Go on their website and sign
with a valid email (no google, yahoo, ecc.) and postal address.
And in few weeks you will receive some IC, also Pic 16F628!
Of course be moderate, do not request too many pieces!
two of every types is sufficient.
For servos I know an italian site where I bought someones.

gidesa

Thank you for the hint ;-)

Ordered samples from the PIC16F628 and some kind of MAX232 chip.

Oh, now that I think about it:
Do you have any idea what kind of chip I would need if I wanted to use
USB instead of the serial connector?
I figured that would be handy since USB also delivers some power. So,
maybe I wouldn't need the battery connectors that way.
Would the power supplied by the pc (through the USB cable) be enough to
power all chips and the servos?

Would be pretty cool I think :).

The power from USB could be enough for the electronic part,
it's just good to maintain the "brain" power distinct from "muscle"
power.
But absolutely not for servos! Maybe at max 1 servo, but I have doubts.
When I use my adapter circuit from 12 v to 6 v with Lm317 to power
the 6 servos, it becomes soon very hot! Remember to screw a metal hot
discharger
to the Lm317.
I think all 6 servos can consume around 1,5-2 ampere, for a total power
at 6 v of 9 -12 watt.
The most hot servo is the one in the "elbow", it must balance
continuosly
the load at the gripper, plus the 3 wrist-gripper servos.
The serial connection is very simple electrical, only 3 wires (ground,
tx, rx),
and as a protocol.
I think that USB is much more complicated as a protocol, but I have not
direct experience on it.
Regards

gidesa

I found a great link to make my own PIC compiler.

The site says it works for the PIC16F628 (and the PIC16F84A).
It seems to be a very simple circuit.

The say I only need:
1x 15kO
1x 10KO
1x 100uF capacitor
1x BC547B transistor
1x 5,1V zenerdiode
1x 18 pin's ic socket
1x DB9 connector (RS232)

It says total costs should be about 1,5 €.
What a contrast to the commercial versions that cost about 100 $.

I hope this helps other people that were concidering to buy a
programmer. :-).

Greetings
BramGo

Forgot to mention the URL.

Link: http://www.circuitsonline.net/circuits/view/91