Subject:
|
USB / x86 open sys based RCX replacement??
|
Newsgroups:
|
lugnet.robotics.rcx
|
Date:
|
Mon, 9 Oct 2000 00:58:53 GMT
|
Viewed:
|
1833 times
|
| |
|
|
After playing with my legos, RCX, NQC, stuff for awhile, I find a more
powerful RCX with proviions for more inputs and outputs and lots more
memory, CPU power, options, and interconnections, seems like it would
be desireable. A very common reaction I think (but maybe I am all wet
here).
I have seen lots of great work on homebrew sensors including sensor
multipler/duplicator efforts. But what I found a little surprising was the
limited posts/sites/etc on alternatives to the RCX itself. MIT has its
crickets (tiny RCX) which are similar in some ways to Mark Fabiny's
C505Brick work (http://home.flash.net/~mfabiny/C505brick.htm), but not
much else.
It doesn't seem like it would be that difficult (though I admit its been a
long time since I designed let alone built any hw) to make an RCX sized
PC (I'll call it an RCU), using one of the recent low end PC-on-a-chip
chips that had lots more memory (a few, if not a few 10s of, MB of flash
and dram), perhaps a PCMCIA slot (gps, wierless networking, 1"
microdrives, etc), 3 or 4 downlink USB ports, one uplink USB port, and
batteries of course. USB provides signals and power. A sensor/motor
pod could offer RCX compatible inputs & outputs (perhpas more battery
capacity as well) and connect to the RCU via a USB port. Multiple RCUs
could be chained together via their uplnk ports as well as be connected
directly to a PC/Mac/etc (such could not only be used for programming,
but I think, to charge the batteries, remember USB is powered).
I have no doubt such would cost significantly more, but since from what I
have seen about half of the robolab/RCX/mindstorm customers are
adults, not kids, I'm not sure this would be a serous problem given the
power of such a unit/system. Obviously there is a limit, no one is going
to pay $1,000 for such a beast, but a premium of a $100 or $200 over the
RCX cost doesn't seem out of the quesiton.
DRAM is about $1/MB and National's Geode line of PCs on a chip are
<$100 (and is a full x86 200mhz core, with PCI, DRAM control, video
in/out, audio in/out 3 USB ports all using 0.8W at 1.6 volts internal), so it
doesn't seem impossible though the profit margin is definetly tight, so
good thing we're not in it for the money eh? :-) Having video (or maybe
even two half sized video inputs) would allow really cheap cames to be
used so that would save money over the USB camera (which could also
be attached directly to a RCU USB port).
So does that seem just out of the question, silly, not woth the effort, I'd
never buy such a thing, why bother or..? ..or are there others that would
like to build more sophisticated robots? Or has this all been hashed out
many times before and I just didn't search the archives or web well
enough (sorry if thats the case).
|
|
1 Message in This Thread:
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
