MA2000 Robot arm - bringing back from the death

[Ichan]

I have a chance to play with 4 of MA2000 robot arm, old good made electromechanics for educational purpose. I will post here the progress of giving new life to them, but i think this will not to become a quick moving thread.

Here is the photo of the fourth robot, the most dead one - being mutilated by the killer.




Some cleaning task done, you can start to see the build detail.



The end effector is pneumatic grip, the three end manipulator is driven by rc servo but not as rc servo (the electronic stripped off), the other manipulator is geared dc motor reduced by timing pulley.

More pictures will follow, ask me any particular parts you want to see in detail...

-ichan

[Ichan]

I will take more pictures in following days, just finish the new servo driver design. They are tiny, the rectangular one is 1.8 x 3.6 cm to be mounted inside of rc servo housing while the round one is 3.5 cm in diameter for the round geared dc motors.

The inspiring page is here, and the youtube video.

-ichan

Edit: add forgotten schematic and replace the pcb picture with the last iteration.

[solutions]

Nice.

If you get a chance, I'm interested in details of the mechanical stuff for the arm. How the motors mount, how the arms are built, and what kind of belts and pulleys are on each axis, along with the motor specs.

For instance, two motors at the joints, and it looks like some kind of encoder?

SOME DAY, it'd be fun to build one from scratch...

So, if you're into the guts, anyway, some more detailed pics would be appreciated.

[Ichan]

Sure, i will do taking detail pictures of all the 6 axis. Already take some but the picture quality is not good so i have to retake it, give me a day or two.

-ichan

[solutions]

No hurry.

Getting the pulley ratios and reverse engineering the motor torques & speeds, bearing sizes and types, as well as the weight and counterbalances would help in replicating one some day.

Not asking for much...
:-D

[Ichan]

Sorry Solutions, you still have to wait for the mechanical details, need a special mood to deal with photographs... 

I am working with the controller now, attached is the photo of the old controller, it is powered by a Rockwell R65C02P2 (2MHz) - I am going to replace it with a new controller based on STM32F4.

The design will be based on STM32F4-Discovery + Embest expansion board (add Ethernet+RS232+SDCard+320x240 TFT LCD with touchscreen+OV9655 Camera), the reason is practical - many code example available for it so it will make a good jump start for me  .

Asigning what port for what purpose is always a hard task at the begining of the design work, here i share my excel file of it - it must be very useful for anyone who work with the F4.

It is complete with the list of alternate and additional functions for each port, just choose what you want to use - the cells in yellow is what i choose for my motion controller board, and again my choice is based on what i said before. The original F4 Discovery use 100 pin F407 mcu, while i need more ports then i will use 144 pin mcu - so the table consist both for 100 and 144 pins of STM32F407.

Ups, can not attach .xlxs files? So it is in a .rar

-ichan

[Ichan]

Some photo of the mechanical started with the first axis, the waist. Sorry Solutions, you have to count the gear teeth by yourself  .




The motor is Maxon DC motor, there is a writes with permanent marker on the body "18V 11W" - i do not know who write it. Will try to find the gearbox ratio later.




Partly assembled, in this mechanism the large pulley is stationary, the motor rotates with the axis.




Fully assembled.




Originally it use potentiometer as position feedback, this one will be replaced by magnetic encoder.




And the new dc servo driver will be on each motor, communicate with the main controller via rs485.




-ichan

[solutions]

I hate to bring this up, but....

What are the chances the mag fields from the motors, or even motor wiring, will mess up your position sensor scheme? Ferrous metals brought into proximity?

[Ichan]

I read that doubt before, i think it is where people talk about HobbyKing MI Digital servo series which use magnetic encoder:



But i can not found a single confirmation about that, so i take the risk - (live need it right?)

As the backup plan, the servo drive can also read analog position from the potentiometer - if the magnetic encoder doesn't work then analog pots will be used on the last 3 axis (rc servo) and US Digital E4P optical encoder on the larger 3 servo (dc servo).

That is a good comment, let me know if you find more info about this.

-ichan

[solutions]

LOL - you don't have enough projects, so here's another for you Ichan: http://hackaday.com/2013/10/09/scara-arm-becomes-enormous-3d-printer/

[Ichan]

Interesting, i do have a spare MK7 extruder unit - but thinking about programming 6 axis articulated arm to do the 3d printing job, well i just can't imagine that.. 

Solutions, are you still under tropical sun? I heard about some activity related with coin mining in there by some western people, are you?

-ichan

[solutions]

Nope - back in the land of corruption and autonomous killing machines for the past few months. I'll head back to feed the mosquitoes at the turn of the new year.

One of these days I need to drop in on your country and spoil your women...

[Ichan]

It will be my pleasure to become your driver in here, even on it's bad traffic jam. As for women it will be always depend on your "charm" before them, right? LOL.

-ichan