Subject:
|
Re: RCX & RIS, a fading glory?
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Fri, 7 Feb 2003 03:55:50 GMT
|
Original-From:
|
Rob Limbaugh <rlimbaugh@SAYNOTOSPAMgreenfieldgroup.com>
|
Viewed:
|
725 times
|
| |
|
|
Here are some of my thoughts (some are new, some are old...):
* Use a wire-frame model to design a very basic conceptual overview of the
bot and where different sensors, actuators, and chassis segments are.
(Think shapes like a basic wooden block set, or even tinkertoys, with
different colors). Each of the devices used by the processor could be
labeled or named (even if the bot hasn't been built yet). This way, someone
could write a program. When time to compile and download, the software does
the "wiggle check" and the user clicks on the device that became active.
* Rather than a "wiggle check" (since everyone is talking about using a
serial bus of some type), why not have a command code that can be issued to
a device to replace "wiggle" with an indicator light? Sensors can't wiggle
anyway. What if the gear train is such that you can't tell what "wiggled"?
A light would be more helpful... could use fiber optics to bring it up to a
hole in the framework, if necessary.
* Rather than just a "bus", I'd like to see specific ports on the new
pbrick that could be intialized as sensor OR actuator. I'd also like the
bus. Say there were 8 ports... perhaps I want to use 3 of the ports for
sensors, the other 5 for motors, or 1 sensor and 7 motors. If I need more,
then I could use the bus and worry about the 12x8 connector block if I need
to. Also, legacy-wise, this would fit with the existing system.
* Two RS-232 ports. One dedicated for pbrick-pc/pbrick communication (like
the IR link now), but open so I can either add my own custom communication
device (RS-232 compatible) or do a direct link to a PC. The other RS-232
port would be for whatever I wanted... maybe a dot-matrix display, maybe a
weather station. These ports (and the other bus) I would hope to see as a
ribbon cable connector.
* Put a power connector back on the pbrick, but also make the battery
compartment removable--actually repositionable. This way, the battery pack
could provide power, but also double as a useable counterweight. Use
multiple packs in parallel to increase run time. A charger would go well
with this...
* To be really picky, I'd prefer that the "bus" be Ethernet (and then ditch
the 2nd RS-232 port). A tiny webserver already exists. Rather than 100bT,
go for 10bT which has less stringent wiring requirements allowing all 8
wires to be used. Power could come from a connection to one of those
battery packs mentioned above. You now have an isolated power and control
system, with addressability and a standard protocol that is tried and true.
Technically, you wouldn't even need an RCX (as you could use a wireless
bridge to a hub and then to the devices).
Anyway, my 2 cents...
- Rob
|
|
Message is in Reply To:
 | | Re: RCX & RIS, a fading glory?
|
| (...) It doesn't need to be huge! You can easily fit 6 inputs and 6 outputs into a space the size of the current RCX. But don't forget that the pads don't have to be spread out on TOP of the RCX, you could easily place them along the sides, or along (...) (21 years ago, 7-Feb-03, to lugnet.robotics)
|
17 Messages in This Thread:
 
 
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
