• Date JoinedJun 4, 2007
  • Total Views113,631
  • Featured %75%


binaryg4 years ago
DB9 connectors would be way to big and their is no need for a master slave situation if you are going for an open source mindstorm type thing. you need a master brick using a microcontroller with as much as possible built in to the chip some thing like the atmega32u4 which has usb onboard which removes the need for the ftdi usb chip.
all you really need is blocks with motors and sensors in that connect to the main block which is easy the hard part is the interface to program the microcontroller which would be easy to use and understand although the arduino bootloader is a lot easier to use than C/C++ but i believe it is still to much for a lot of young people. i believe the mindstorm uses a flowchart type programming interface so you can just drag and drop commands and then upload the program now that is the challenge but could be done using processing along with the arduino bootloader. As for cost you could easy make the master brick for under 5-10GBP depending on quantity made and sensors+motor bricks for 0.50-10GBP.
you could easy add slave devices using SPI or I2C but i could only see the need for it if you wanted to expand the amount of ports because the atmega32u4 has pretty much all you would need on chip for all but the most advance projects and mindstorm is only ment to be an introduction to electronics and programming.
Anyway that my babble on i could go on for hours
all the best

i have thought a
paulware4 years ago
Hey, I have been thinking about this quite awhile and just yesterday I believe I stumbled upon a workable solution.

The idea is to start with a "master" brick and then connect to all subordinates using db9 connectors. Typically db9 is used for serial communication but only 3 wires are needed out of the 9. If we used another 3 for power, and IIC (2 wires). That would leave us 3 spares.

Each brick (master or slave) would receive commands on 1 db 9 and broadcast commands on the other db9s. So a slave brick could also be a master to other bricks.

Part of this hinges on a hunch that I believe a master brick can "broadcast" to many slaves over RS232 by tying the rx lines of each slave port together. I believe this should work in theory, but I haven't confirmed it. This would make the protocol simple.

The goal would be to keep the cost of each brick below $10.00. This would compete favorably with another similiar product that uses cubes and is about $20 a cube.

What do you think?

Have I babbled too much? Perhaps a picture is in order.

I like what you have done by the way. My son loves legos, and cheap electronic legos would be huge.