RoboCalc

[elfprince13]

I just had this idea:::

a hacked combo of an audio cable and a lego motor cable plugged into the sensor port on an RCX would allow a and the I/O port would allow a calc to control the RCX, and even communicate with other calcs using the I/R port.

I can do the RCX part and even create a Basic interface using Omnicalc's getByte/sendByte pair, but if an Asm programmer did the calc side it would be faster and more polished. We could even do a codex style thing and allow Basic to pass parameters to the Asm utility to control it.


@ some mod:

could you add new forum for this?

[DigiTan]

This should be possible. If you want to emulate raw linking, you'll need PindurTI because VTI freaks out and hangs unless you use the normal calc protocol. PTI has a way more-detailed link debugger also. After that, it's a matter of sending the right values to the link port--which is not as complicated as it sounds. New forums are usually handled by the AlienCC, JBirk, or the other founders. Tough I should hint that typically, they are offered rather than requested.

[elfprince13]

the description for this forum said you request a new one....sorry for being overly bold....

anyway the RCX end is fairly easy. monitoring raw sensor values with NQC and adding a certain action for each sent value.

[DigiTan]

I didn't mean to imply you shouldn't request it. But it's best to PM or AIM them and go from there; and get an earlier response that way too.

Anyway, would the Lego's be controled or an on/off basis? If so, you can start with the two I/O lines and possibly expand them beyond two if needed. MC had a news topic on a calc-bot a while ago. Something about a European competition.

[elfprince13]

yeah I heard about that...it wasnt legos though.

so heres how it would work:
the calc connects through a hacked cable to the sensor port on the RCX. the RCX controls the motors and everything but it takes instructions from the TI. this would allow the RCX to utilize the TI's superior math capabilities to make more complex robots. Or youcould use the keypad on the TI to change the direction of the motors, or display sensor info from the RCX on the calc's screen etc. Even make games that would involve driving the robot around. lol

[elfprince13]

Sorry 'bout the double post,

anyway just an update on this project:

I got a broken pair of earbuds from my little brother and extracted the little plug thingy. Then I got a Lego cable that had already had one end snipped off that was left over from a course that my dad taught on robotics. When I get access to a sodder (how the heck do you spell that?) gun at school tomorrow Ill have a prototype cable ready.

One possible set back is with the amount of current the calc uses to poll the sensor cable: it is constantly putting out current, and I dont want to fry my linkport. does anyone know how much current the linkport is designed for? a simple tongue-test seemed to indicate that the RCX is putting out less than 9V, but more then 1.5....actually probably less than 6 volts, cause it takes 4 AA batteries.

[DigiTan]

Don't worry about the post. We allow a multipost now and then; just as long as there's at least a day or two between.

I could only guess at what the current limit is. Joe Clussman says the links are fused, but no one seems to know the limit. As best as I can figure, a TI transfer moves between 0.6mA and 12.6mA, but the diagrams at ticalc end there. You should probably keep it below 12mA to be totally safe.

[elfprince13]

okay, anyway, here are a couple of problems:

one:, I melted the plastic separator in my little headphone stub
two: the sensors use analog input that is then digitally interpreted. This will require an interesting workaround. The RCX will have to monitor for pulse patterns instead of bytes. Anyway perhaps a sound sensor routine could be used to interpret output from OmniCalc's play( function. This is somewhat strange however, since the rotation sensor has 16 increments, the touch sensor has 2, and RAW input is scaled from 0-1023.


so is there anyone here who has experience programming the RCX, other than me? I found a couple of sites with interesting insights to the internal workings of the RCX, but most stuff isn't well explained.

[elfprince13]

does anyone have tips on building an IR cable?
before I start buying materials, id like some advice.

*looks at digitan expectantly*

[DigiTan]

Sami Khawam's IR project should do the job. He used the exact same 38KHz carrier the legos can receive. It also used a Siemens IR receiver that is compatible with the Sharp receiver inside the lego. Actually, the lego can also detect TV remote signals, but the encoding will not match. You'd have to modify one of his IR programs to get the right encoding.

[elfprince13]

I knew the bit about the TV, and Ive seen Sami Khawam's IR site, but it seemed to be sort of dead , looking at the latest news. Does he have drivers for it that will allow you to send any value? or is it just the IR chat program? I dont care about encoding because I can set the RCX to recieve raw bytes and thus create my onw protocol. thnx

[DigiTan]

I think his just works for RC5 and REC80 encoding. Of course, if you get the PCX to read those, that's all the work you have to do.

[elfprince13]

so would I have to write new could for the PIC?

[DigiTan]

Yeah, that or the calc itself (only if you're using 83+ or 84+). You can eliminate the PIC althogether but that makes things real tricky because you still need to create a 38kHz carrier, which eats up most of your calc processing time. Most Z80 calcs aren't fast enough to do that.

[elfprince13]

well, Ive got the 84+SE, but I want to try and do as much as possible in Basic. I guess it is time to start learning how to program PIC chips