Be advised that the circuit on this page is in the "design phase" and has not yet been built or tested. Build at your own risk! My electronics skills are very, very limited. Be that as it may, I felt compelled to share my current thoughts...
We all know that the programmable bricks use IR communication. For most applications, this is quite adequate. However, consider the following project ideas:
Design considerations:
The general idea is to place a small brick with an IR detector and IR Emmitter in front of a programmable brick or in-between multiple programmable bricks. This would allow a PC to communicate with the master brick and still allow master-slave brick IR communication. The idea is NOT to cover the entire front face! This would also allow for each RCX to receive new programs from the PC when developing without having to hold the bot at weird angles--just make sure the brick you wish to program is on and the other is off. The PC could also monitor communication between bricks, if desired or communicate with them.
For the PC side, the front face of the IR tower where the built-in IR Detector and Emitters are would be covered. This is to prevent signal echo in the event a programmable brick is within range of the tower IR communication.
The RF stages in the schematic below shouldn't be critical to adhere to. Essentially, any transmitter/receiver pair capable of 2400bps or better should be fine. The 250 Ohm resistors off of the 4001 were chosen because of the rise/fall times of the TX units I plan to use.
With higher speed RF devices, it may be practical to use Encoder/Decoder IC's to create channels within a specific frequency. The TX/RX components here are for 4800bps (or less), so I won't even attempt to use encoders/decoders at this point. Doing so would mean a significant jump in data that needs to be sent in the same amount of time as "non-encoded" data.
Theory of Operation:
In a nutshell, we want to take incoming IR and send it out over RF. Then, we want an receive that transmission and convert it back to IR. Sounds good, but not that easy.
First, IR detectors are logic-high when no IR is being received. We'll invert that so "no signal" is logic-low. If we leave it logic-high, then the transmitter would always transmit.
Our next hurdle is making a 38kHz IR signal. While it's great that we have a convenient decoder package, there isn't really an "encoder" package. So, we know we'll need to use at least one chip to create that. We'll use a 555 timer IC to follow the "no programmable devices" rule.
Since all the RF devices use the same frequency, we need to control which way data is going. The Quad NAND performs the logic functions. Rather than disabling data paths, the logic has been worked out to only send data to appropriate paths if certain conditions are met. See the truth tables below for a better understanding of what the chip is supposed to do.
NOTE: I'm still unsure about things, but this schematic does look much more impressive than the Rev. 1a did! |
That's the theory, anyway. For each device that is going to participate in RF communication, build another WIRRL. Again, I'd like to remind everyone that I haven't built this. I don't even know if my additions (surrounding the 4001) are correct. If anyone sees glaring errors or actually gets this working, please post to the list.
-->First major change is that the RF units have been replaced with Linx LC TX/RX devices. They are cheaper and should be more efficient. They are also much smaller than the TWS/RWS units.
After thinking about the mode switching a little more, I worked out a truth table. I found out that instead of a 4001 Quad NOR and switching/inverting transistors, a 74HC00A Quad NAND could be used. The NAND handles the logic for inverting the IR Detector output (when no data, it outputs high). The next 3 NAND gates are used to determine if data should be sent to the oscillator circuit or the RF TX circuit. Priority is given to incoming RF data.
...noticed at least one glaring error with the new schematic!
Say, for argument, an RCX sends "1001", which, in IR, would be "0110". The IR detector of WIRRL would do nothing for the first bit ("0"), but the next bit ("1") should be sent over the RFTX, and, according to the current logic/schematic, it is. However, I never disabled the line between the RFRX and the Oscillator, so the RFRX will receive the transmission from the RFTX (because all are same frequency). This would make IR = 1 and RF = 1. According to the truth tables, priority will be given to RFRX data and the third bit of IR ("1") would never be sent.
Apparently, while creating the truth tables, I gave consideration to this... I just never included it in the schematic.
The schematic will be updated shortly.
Pictures are worth a thousand words, so here are links to images that explain some communication options:
Two devices communicating over two carrier frequencies.Results may vary. Use this information at your own risk. Information is for instructional/educational purposes only.
All pictures and content Copyright ©2002, all rights reserved.