Use mini-ATX+Phidgets or Bioloid for my robot?

[alexannic]

Hello,
I've been investigating the Bioloid kit for some weeks now, and I realize that the AX-12 Dynamixel servos have many technical advantages over conventional servos.

I am trying to build my first robot which I intend to develop as follows:
1. Initially, be able to control the 4 wheels of an R/C car
2. Add distance sensors and map the terrain of a room
3. Program the robot to act independently based on rules

I'd like to ask you the following question:

Is it better to buy a VIA Mini-atx board and using USB Phidgets control servos & sensors or get the Bioloid basic kit and familiarize myself with the CM5 controller & software?

My only worry with Bioloids is that the Motion Editor and Behaviour software is too limited, and I do not have the skills to program the servos - sensors to work with an independent controller.( I am not an electrical engineer or into electronics so much!).

However, I can easily progam a high-level language (.net) and manipulate the Phidgets using their API. On the other hand, the Dynamixel is superior and cheaper in terms of performance with any other servo I've found on the Internet.

I'd really appreciate some advice!
Thanks,
Alex

[jonkster]

Hi Alex

if you are going to use the CM-5 to directly control the ax-12's (rather than use the motion editor etc) you are going to have to program the CM-5 at a reasonably low level anyway. Problem with the CM-5 in this role however is the lack of access to any of the AT-Mega128 ports meaning not much option for other sensors.

Options then are to either develop a device that communicates with the CM-5 using same protocol over the 3 wire bus (like the AX-S1 but with options to add your own sensors) or to use another board.

I started developing a AX-S1 like replacement but quickly ended up deciding it would give me more control to replace the CM-5 (doesn't mean developing an AX-S1 replacement is not worth trying just I went another way). Think others are pursuing this though.

I am now using a MAVRIC IIB board (ATMega128 based) http://www.bdmicro.com/ and a simple circuit to drive the 3 wire bus (as per the AX-12 manual). I have ditched the C code in the AX-12 manual and have written my own communication routines but the stuff given in the manual does work fine. When I am happier with my stuff may post it here. (I am writing it to run under FreeRTOS http://www.freertos.org which adds a level of complexity.)

Don't know much about the USB Phidgets you mention but don't think you need that much electronic knowledge to interface most high speed processors to the AX-12's. Use the circuit given in the manual (2 simple chips and you could build it on some veroboard - maybe cost a couple of dollars all up).

If you are proficient in programming C# then the actual programming won't be too much of a hurdle. Probably biggest learning curve may be understanding the operation of the microprocessor and it's programming environment. Perhaps here running something like a mini-atx board
may be less of a learning curve as you could run a development environment you are more familiar with.

The one thing I would suggest though is to get a CM-5 anyway. Reasons:

1. you can verify the servo's are working

2. you can experiment with communication protocol

3. you can change baud rates on AX-12's and change flash values (can do this with another processor but you will have to get it talking to AX-12's before you can do this). I am now using 1Mb same as CM-5 but did use lower rates when using a slower processor and when tracking down a communication problem.

4. you can use it to communicate to your development system using the correct protocol - I found this invaluable when getting my device talking properly

hope this helps

cheers

Jon