While I’m debugging ARTv4 to have interesting, pseudo-intelligent behaviors (and not running out of SRAM which causes it to reset randomly…), I’m also working on the next generation electrical design for ART… I’m naming the bundle of software and electronics ARTv5.
The plan is to support more infrared and ultrasonic sensors with an RF transceiver for sending telemetry information. I also want to start producing and using PCB boards to get away from the fragile breadboards.
More specifically, these PCB will take the form of Arduino-shields that I can stack on top of each other.
The plan right now is to create three boards that are stackable:
- One for the ultrasonics sensors (front, rear, left, right);
- One for the infrared sensors (front, rear, left, right);
- One with four connectors to ultrasonics sensors, infrared sensors, motors control in addition to the connector for the RF transceiver.
The last “shield” will also host the CPLD (Xilinx xc9536). All incoming and outgoing signals go through the CPLD to give me the flexibility to do additional processing using digital logic instead of the microcontroller code. In addition, I also have a Texas Instrument CC1101 chipset transceiver (supposedly RFC1100-A but the board they sent me has FT1100A-100) with Power Amplifier that I’m adding so that the Microcontroller can send back telemetry information to the PC…
So there’s the following connections:
- Ultrasonics sensors shield: 10 wires (1 VCC, 1 GND, 4 echos, 4 triggers);
- Infrared sensors shield: 6 wires (1 VCC, 1 GND, 4 signals);
- RF transceiver: plugs into two 9×2 headers (1 VCC, 1 GND and the SPI wires);
- Four control motor wires for the H-bridge (used right now to connect directly to the RF transmitter as a hack).
To do the design, I’m still using the fantastic Fritzing 0.4.3. Although it does crash regularly and has various nitpicks that, if solved, would make this even better there’s nothing really comparable. The workflow feels natural and the results are attractive. I haven’t been able to produce etchable versions of my PCB design due to what seems to be a bug, but the Gerber RS-274X files it generates seem to work fine (if GerberView output is any indication).
I’ve compiled a list bought so far and the cost so far is a bit under 800 RMB. On the basis of that new design I need to add to the wishlist a whole bunch of stackable female headers in addition to figuring out the cost of producing the PCB.
The BoM generated by Fritzing for the ART Shield says:
- 1xGeneric female header – 10 pins;
- 1xGeneric female header – 2 pins;
- 2xGeneric female header – 4 pins;
- 1xGeneric female header – 6 pins.