Subject:
|
GBC siding control - status
|
Newsgroups:
|
lugnet.org.us.laflrc
|
Date:
|
Mon, 30 Jan 2006 19:34:05 GMT
|
Viewed:
|
2240 times
|
| |
|
|
I built quite a nice sensor, neatly landscaped into a small rocky roadcut...
and then realized it would require two RCX inputs, which I didn't consider
acceptable. So I thought about it and rebuilt, but didn't "pretty it up" yet.
However, it works. Right now I've got a double-loop set-up, or more correctly a
loop with a siding, but with the siding re-entering the mainline "before" it
exits - see:
http://www.brickshelf.com/gallery/brdavis/GBC/GBCtrains/trialloop.jpg
for a pic. The yellow "cuts" across the small loop are the isolation points,
with the bulk of the small loop being under the control of the RCX while the
outer loop is driven by the controler. The RCX can keep two trains running on
the outer main loop, with the GBC train being redirected into the siding on each
pass (the sensor is on the green baseplate - close-up here:
http://www.brickshelf.com/gallery/brdavis/GBC/GBCtrains/sensorfront2.jpg
And yes, it can be made prettier, I'm thinking either rocky roadcut or
disgrundled passenger station (since the trains never actually stop).
Right now, this set up only uses one input and two outputs from the RCX, but
all the timing on the GBC siding is "blind" - the RCX does not know when the
train has cleared the point on the way into the GBC siding, or that the train is
"parked" at the proper place at the end of the loop, ready to go. It does time
the non-GBC train so that it knows when to inject the GBC train into the main
loop. It should *almost* be suitable to work in John's conception of a
multi-siding system with one RCX controling switching (the siding is driven by
the 2nd RCX output because it's availible). The problem here is unless the
scheduler RCX knows the status of a siding, (i.e.- if a train is still being
serviced), it might direct a 2nd GBC train in resulting in a collision (similar
to what happened at BF'05 when both trains got into one control zone). This is
only a problem if we want more thaqn one GBC train, of course.
My questions: first, is that sensor assembly OK visually? I'm not sure I can
make it smaller and still distinguish GBC trains, but I think we can hide it as
a "natural" rocky feature or a station of static unloading dock. Second, is that
geometry OK (i.e., the GBC siding as a loop)? I did this to compress the layout,
while at the same time making the distance between the sensor and the turn-out
point a maximum: as soon as the RCX detects that the train is a GBC train, it
throws the switch, but throwing the switch takes about a second, and the sensor
only registers as a GBC train after the engine has already gone by and is
rapidly closing on the critical switch. Third, while this set up can handle the
trains moving at different speeds (the bigger the outer loop the easier this
gets, obviously), there are limits: I can't do three trains (yet), and the
longer the trains get the tougher timing everything gets, requiring either more
sensors and/or longer transitions and sidings.
Opinions? Comments? Snide remarks? Yes, I know the inner loop isn't "proper"
geometry, but ironicly I need more curves (!). I'm rather happy just watching
the two trains dance around the layout - they've been running for something like
two hours with no problems... well, none that haven't been directly caused by an
inquisitive two-year-old.
Folder after moderation:
http://www.brickshelf.com/cgi-bin/gallery.cgi?f=166084
PS- lots of flying balls (but a big download!) at:
http://www.brickshelf.com/cgi-bin/gallery.cgi?f=165930
--
Brian Davis
|
|
1 Message in This Thread:
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
