Subject:
|
Major breakthrough serial wise.
|
Newsgroups:
|
lugnet.robotics.rcx.pbforth
|
Date:
|
Thu, 1 Jun 2000 12:43:02 GMT
|
Reply-To:
|
sjm@judgementSTOPSPAM.com
|
Highlighted:
|
(details)
|
Viewed:
|
1448 times
|
| |
|
|
I have been fiddling with serial interrupts with little
success. Today I finally figured out that interrupts
are masked in the CCR until you run the word RCX_INIT.
I was blindly assuming that interrupts were generally
enabled and I only had to deal with serial stuff.
Now I have to think about the correct place to clear
that bit. Ralph, do you think it is safe to enable
interrupts earlier? Normally I would initialize the
state of all devices that might generate an interrupt
before clearing a general mask bit but here we don't
want to muck with much device state that might be
depended on by firmware.
I guess that I need to experiment, i.e. clear the mask
bit when I need it and see if everything still works.
We certainly can't have serial interrupts dependent
on RCX_INIT.
I guess I need to look at RCX_INIT to make sure it
doesn't fiddle with serial registers.
This means that today for the first time I can actually
get to my now gutted interrupt handler (it now only
contains debug code.) Now I have to restore my original
code and get back to normal testing. This weekend should
do it. I only have about 30 minutes a day to spare to
this project so it gows slowly; but it goes.
Onward and upward.
Steve
PS. My IR connection stinks. It is seldom that I can
upload more that 5 lines of script without at least one
error. I don't completely understand what is happening.
It won't be just one bad byte. An entire string will
be wrong. My workaround is to only define words
in a script and then download several times till everything
gets defined, jiggling with the positions of RCX and
tower inbetween. Very annoying. What is strange is that
the echo in the window is bad. This is generated by the
tower and probably has nothing to do with the RCX except
as a source of reflections. I don't understand how
reflections can cause problems. It is not like we are
dealing with current speed. IR is moving at the speed of
light (it is light) eand all reflections basically get
there at the same time at 2400 baud speeds. It isn't the
normal signal broadening and overshoot that electrical
reflections cause. As far as I can guess it must be a
signal intensity problem eithor 0 or 1 or both.
|
|
Message has 2 Replies:
 | | Re: Major breakthrough serial wise.
|
| in article 39365A56.86986217@nexen.com, Steve morris at smorris@nexen.com wrote on 6/1/00 5:43 AM: (...) Try posting this on a more general list if the following suggestions don't help: 1) put the RCX and tower underneath a box to block out all (...) (24 years ago, 1-Jun-00, to lugnet.robotics.rcx.pbforth)
| | | RE: Major breakthrough serial wise.
|
| Soirry for the late reply, I've been on a sudden business trip and was unable to get to my email... (...) Great! I would leave the basic serial stuff alone and enable interrupt driven serial IO in a separate set of words, like RCX_INIT and (...) (24 years ago, 3-Jun-00, to lugnet.robotics.rcx.pbforth)
|
4 Messages in This Thread:
 
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
