Oscar-Home made PICAXE controlled robot arm

How do you rate the arm in term of deign/skill/etc...

  • Excellent

    Votes: 8 88.9%
  • Good

    Votes: 1 11.1%
  • OK

    Votes: 0 0.0%
  • Nothing Special

    Votes: 0 0.0%

  • Total voters
    9

Lliam20789

New Member
Hi everyone,

Just thought some of you may be interested in the robotic arm I have made that you may have herd about.

It uses PICAXE 28X and 5 servo motors and was constructed and designed interlay from scratch.

Please rate the project, in terms of creativity/skill/design/etc...
Also feel free to post comments on how I could improve Oscar in the second build.

http://au.youtube.com/watch?v=U7l1Rn10I1Q

Working!!!

Lliam
 

Attachments

Last edited:

Lliam20789

New Member
Technical Difficulties

Ok no movie yet but still working on it... (silly youtube)
Well basically the picture sums it up, but the movie should work soon.
 

leftyretro

New Member
I reserve poll voting for now. The picture is really not clear or detailed enough to see what you actual built.

Just as in the real business world, good documentation is a requirement if one expects positive feedback on one's effort ;)

Lefty
 

Lliam20789

New Member
Fair enough. Well, last resort, does anyone know why it's going all fuzzy?
or is that just for me??

I hope this works soon...

any suggestions?
 
Last edited:

hippy

Technical Support
Staff member
It's always a hard one to rate someone else's project. I can really appreciate a desire to hear you've done well, to have your back slapped - and you have no matter how the votes go - but polling for a response doesn't always deliver what you'd like to hear. It can narrow feedback down to a technical assessment of the outcome, rather than a measure of achievement in getting there. The only quality questions I believe one can ask are; have I achieved something, how instructional or useful is it to others ?

For myself I feel it better to 'publish and be damned', believe that what I'm publishing will be useful for someone, somewhere, sometime even if the majority of the world thinks it's of no worth what so ever. I can bask in the self-smugness of my own achievement and occassional bouts of sheer-brilliance, and someday someone will come along and express appreciation and understanding of what I've done ( thus confirming my right to self-smugness ) and confirm it's helped them along their way ( confirming its worth ), and if they say otherwise; well they just haven't understood or appreciated what I've done :)

I think you've done a mighty fine job in building a robot arm. In bringing mechanical, electrical and control skills together. I am sure it's inspirational to others, more details on 'how' and it will be very instructional as well. How I'd rate it in your poll ... I have no idea.

As an aside ... You're strugging to post a video to YouTube. I doubt you are the first or will be the last. Documenting how to do that, the mistakes you made, and how you got round the problems would be a useful project in its own right.
 

Lliam20789

New Member
I appreciate your advice hippy, thankyou.
I guess your right about the poll, the main feedback I was after is how to improve?

The movie is in MPEG4 with no sound...

When we get this going I might post an easy to follow "how".

HISTORY:
Oscar was developed as a test of the mechanical, electrical and programming abilities of Glenn Maple and myself.
The name Oscar, (only a recent addition by me) came about by the first sentence ever written by the arm, which was a quote by the Irish-born author of the late nineteenth century, Oscar Wilde. From this point on ‘The arm’ became known as Oscar.

Design:
The design came about after a few hours of research on the Internet. It is mainly taken from a robotic arm called the Lynx 5 Robotic Arm. From this concept I created a simple template for the Perspex.

Materials:
Oscar’s base is made of wood, on top of which sits a second Perspex base. The upper and the lower arm are also made of Perspex. A metal brace is used to support the shoulder. Screws, nuts and bolts are used for joints and support.

Electronics:
The entire arm, consisting of 5 servo motors, and is run using a PICAXE 28X Micro controller. Other electrical materials include, a 6 volt power supply, wire, capacitors, resistors, diodes and connectors.

Time:
The arm was constructed and ran with a basic program in Semester 2, 2006, allowing it to type the alphabet on a computer keyboard. Further improvements to programming were made in Term 2, 2007, including the ability to write words and sentences.

Programming:
In 2006 the program consisted of a set of motor coordinates and pauses. That has now been developed to have all letters, and some other keys as sub routines. This allows any letter to be called upon in any order.

Faults:
Oscar runs well and can complete sentences without more than two or three mistakes. All mistakes are due to miss-presses where Oscar will:
-Press down before the destination of the key is reached,
-Double press a key,
-Miss a key

Cost:
Servos ($16 each) $80
Picaxe 28X and board $50
Other materials $20
-------
$150
Future improvements:
I have tried programming Oscar with a unique speed control program, but this proved to large and clumsy. I may try re-writing this program to get rid of the fast, jerky movement.
Other improvements include making another arm with more abilities, for example, a gripper.

Credits:

Matt O’Neil, Assistant programmer and technician.
Josh, Assistant programmer.
Brian Fitches, Technical backup.

Glenn Maple, Co-production and Co-initial programming.

Lliam Anstey-Sprigg, Design, Co-production, Co-initial programming technician and new
programming.


FUTURE:
My ideas on future design so far are:
-external pots for position feedback
-2 x Home made pressure sensor on top, bottom ,left and right of each main joint.
(these would be wired in series so that top back and bottom front were paired. This would mean that if the joint was trying to pick up something to heavy, the resistance would lower and be read by a PICAXE. Maybe even calculate the weight!)
-GRIPPER! (defiantly a necessity by struggling to find a good, easy, universal design...

Suggestions are welcome!

Lliam
 

BeanieBots

Moderator
What would be really nice is a technical spec. Something that none of the readily available robot arms give either.
The user of any robotic tool needs to know speed, resolution and repeatability.
How long does it take to write a letter?
How closely will it go to the "same" spot each time?
What is the smallest 'repeatable' shape it can draw?
How straight are the lines if it draws a square?
How wide are the lines if it draws the same square many times?
 

Lliam20789

New Member
BeanieBots:

At this stage it can not draw, I lacked the time to come up with a 'straight line' algorithm...

It can type about 1 letter per 1.8 seconds on average, this can be increased but reliability is lost. The video that will be posted soon (I Really hope) did take 3 shoots due to miss presses etc but over all it is fairly accurate with the repeated position within 6m of the first.

I hope that by using external pots in build 2, I will be able to increase the resolution, in turn increasing accuracy.

Thanks for your reply
 

Lliam20789

New Member
That was one of the next steps! No use, but cool...
I was going to make it so it could interoperate a sentence typed in by the user and re-type it.

Well thankyou all for your feed-back and suggestions! looking forward to many more.

Lliam
 

BeanieBots

Moderator
Something I've often toyed with doing is a robotic arm that can play chess.
Stop the gasps, I don't mean play the game, I just mean pick up the pieces and move them.
Leave the game play to a dedicated chess computer.
With a dedicated chess computer, it should be possible to read the display with a PICAXE and then make the move.
Alternatively, it might be possible to find a chess program that has the option to "send" moves serially. I have seen several that produce printouts of each move. That could be read by a PICAXE and acted on.
Would certainly make a very impressive project.

Maybe an interactive board. Magnetic reeds or similar under the board to detect the human move. Then your robotic arm types the move just made and then does the computer move. If it loses, it could even have a 'paddy' and knock over all the pieces:)

Give it pen and let it play tic-tac-toe.
 

WHITEKNUCKLES

New Member
BeanieBots,
We tried the chess arm some time ago, making and testing a rough arm while we searched for a program that output it's moves to a printer. Unable to find a programme that did this in a manner that could be intercepted without spending large sums and with no assurance that the information was output in a useful form we bought some chess machines. The intention being to use the drive for the leds as information on the machine's move which the arm would make, we would move our own pieces.

The nearest we came was with a couple of Vintage (£££) Chess Master machines from the East German DDR. Beautiful workmanship, the magnetic switches (not reed) giving an audible clunk, moves interlaced with from square indicated then too, but no capture with unfortunate timing and a lack of Russian for the circuit diagrams proved insurmountable to us.

If you could direct us to a suitable programme we would be most grateful.

But we do have a small, interesting selection from the "no hope" machines to the "nearly there" machines that range from ChessLight to Academy and Conchess to Corona's.

Dave
 

BeanieBots

Moderator
Haven't actually tried it myself yet. I would have thought that a program that outputs to printer should be easy enough to capture and a 'chess computer' with LED readout should be possible to interface to without too much difficulty.
 

hippy

Technical Support
Staff member
A chess Program doesn't necessarily have to win. It should be fairly easy to knock up a very basic program with some simplistic strategy which is good enough to play short games. The core algorithm can be improved later, the amazed audience facing side of things will remain the same.
 

boriz

Senior Member
I had a small portable chess computer. Can’t find it now. But interfacing to it would have been easy. Each square had a pressure switch and an LED. When the computer moved, the ‘from’ LED would light until you press the piece, then the ‘to’ LED would light until you press the piece on the ‘to’ square.
 

WHITEKNUCKLES

New Member
Lacking the ability to programme a PC this was to be a Picaxe project using an arm similar in design to that on PDF page 19 of this site which is new to me.

We had got to the point where the drive to the LEDs could be intercepted and the arm could be driven with the required power and accuracy to pick the piece and enter the move.

Although large at 430 x 350mm the Chess Master does not need great accuracy as the powerful magnetic pull of the switches centres the pieces.

The Picaxe boards are shown connected to the ChessLight, here the 08M listened for the move ready tone.

Dave

http://www.dcs.gla.ac.uk/~rod/legolab/ESEteamReport04-05.pdf
 

Attachments

Last edited:

WHITEKNUCKLES

New Member

Rickharris

Senior Member
1st - A great project - although I agree with Hippy's comments re publish & be dammed.

2nd - If not too sensitive how old are you? We have had students make robot arms at school using both servos and home made liner actuators using a threaded rod and a nut. with mixed success - much depends on lightness without compromising stiffness.

3rd. - Use the arm to move small objects round to make a clock. - people are always fascinated by a machine using building bricks to make towers - - The Towers of Hanoi are another regular task that isn't too difficult to programme the logic (lots of web support) http://pauls-pc-repair.com/sudoku/Tower.html - moving a web cam around to show different angles/views of something was popular at an open evening - Completing a jig saw, OK the bits have to be in a preset position but no one else knows that! - Playing a musical instrument xylophone may be the easiest - Operating another machine turning on and off various switches knobs etc kind of a recursive art installation if the other machine affect the movement of the robot arm say by putting lghts on that need to be turned off.
 

Lliam20789

New Member
Great suggestions, Thanks!
I'm sorry I do not have the time to reply to everyone but i'll try answer all questions etc.

I'm 18.

I think chess is a bit advanced at the moment, needing a very good computer program. But I will experiment with naughts and crosses or something small like that, does anyone know how I could read drawn circles or crosses or maybe plastic ones?

Moving a camera sounds good.
Building something was a goal, like a miniature production line.

A while ago I had the idea of making it automatically adjust settings on a sound mixer desk, volume from various channels as the inputs...

Does anyone have a design for a gripper?

thanks,
Lliam
 

BeanieBots

Moderator
Several people have already hinted that to be impressive, you not need to do the task in exactly the correct way. That is, for chess, it does not need to win, simply make a few moves that look like it knows how to play.
For naughts and crosses, don't try to detect that what the other player has drawn. Just detect that they have had their turn. That could be as simple as a pressure sensor under the paper. Nine push-buttons being the simplest option.
 

hippy

Technical Support
Staff member
Magnets in the pieces ( be they chess pieces, draughts, noughts or crosses ) and reed relays under the board is one way. Any way you do it you'll have to scan the board. Comparing which reeds ( buttons, whatever ) are different to the last scan will show where a piece was removed from or placed.
 
Top