Subject:
|
Re: RCX Tower Internals
|
Newsgroups:
|
lugnet.robotics.rcx
|
Date:
|
Thu, 10 Jun 1999 18:00:03 GMT
|
Viewed:
|
1637 times
|
| |
|
|
Ralph Hempel <rhempel@bmts.com> wrote:
> This might be a really simpleminded circuit. As far as I can tell there
> are a pair of 132's (simple HC logic chips) - one for each direction.
> They implement direct echo back, which implies that the channel is really
> only half duplex - you can't send a packet to the tower while the RCX
> is sending info back to you.
I'm pretty sure the echo is not direct, it goes out through the emitters
and back into the receiver. This is demonstrated by varying the lighting
conditions and/or the baud rate and noting the response. If the echo were
direct, you'd receive it all the time.
But you are right that the connection is not full duplex. Transmissions
interfere with any incoming signals.
Having played with the tower a bit more a few weeks ago, I noticed that
once the connection warms up, it remains relaible as long as it remains
powered. I then modified my send routines to take advantage of this, but
didn't update the distributed code yet. Basically you can call a routine
"warmup tower" that checks if the link is warmed up and if necessary waits
(with timeout) until the tower is warmed up before proceeding.
A related fact is that once you turn on the RCX, it warms up fairly
quickly, and therefore for the most part its reception is always reliable.
This explains why sometimes send.c reports failure even when the RCX
received, accepted, and executed the command. The old version checked the
echo but failed because the tower reception never warmed up in time; the
new version makes sure the tower is warmed up first before sending
anything, and therefore doesn't demonstrate this behavior anywhere near as
often.
-Kekoa
|
|
Message is in Reply To:
 | | RE: RCX Tower Internals
|
| Kekoa wrote... (...) This might be a really simpleminded circuit. As far as I can tell there are a pair of 132's (simple HC logic chips) - one for each direction. They implement direct echo back, which implies that the channel is really only half (...) (25 years ago, 10-Jun-99, to lugnet.robotics.rcx)
|
5 Messages in This Thread:
 
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
