Subject:
|
Angles and Diagonal Lines (was: Re: Angle Sensor)
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Fri, 5 Feb 1999 19:56:55 GMT
|
Viewed:
|
1236 times
|
| |
|
|
"Carlson, Doug R (MN65)" wrote:
>
> > > The LEGO sensor can keep up with a motor directly connected to
> > > it.
> >
> > Has this been everyone's experience? I have not had the chance to
> > experiment with my angle sensors very much since I got them, but by
> > hooking one up to the RCX and VIEWing the sensor reading (sensor set to
> > angle via Command Center), it appeared to me that the shaft had to be
> > rotating pretty slowly to keep the RCX from missing counts. How slowly?
> > About as slow as the micro motor, maybe a little slower, as I recall.
> > Perhaps the display of the RCX doesn't update as quickly as the internal
> > counters..?
> >
> > Regards,
> > Issac.
>
> I am using two angle sensors in my synchro drive bot and both are rotating
> at
> mini motor full speed. There have been times where it seemed to drop a few
> counts ( approx. 1% )
> but I have yet to verify this. I only noticed this because after running my
> bot through several square
I run a plotter with two lightly loaded 'old' 9v motors. At power level
4,
the angle sensors dropped readings. I geared the sensors down 3:1, and
they
seem to work fine now. The newer 9v motors are slower - your results
may
vary.
(BTW: the total gear train is: Motor at 1:3 to sensor; Motor at 1:24 to
an 8-tooth gear driving a rack for x and y positioning. This is a
pretty
good speed/accuracy trade-off. It gives 16 angle 'counts' per tooth of
the rack, which are about 10 to the inch. I'll take 160 dpi from a Lego
plotter... Mind you, mechanical looseness degrades this a lot.)
Another factor depends on your program: The RCX runs tasks
"simultaneously"
by executing instructions from each in turn. From what I can tell,
running
many tasks decreases the frequency at which your task performs an "if
angle
> x then stop" type of action. So, even if you *tell* the RCX to stop at
a certain position, it may overshoot unless moving slowly. By designing
with fewer parallel tasks, you can improve performance.
My plotter manages about 0.5mm resolution over extended drawings by
using
few tasks, 3:1 geared-down sensors, and by ramping power down as it
nears
the target coordinate. The only thing it has trouble with now are
slanted
lines.
Has anyone come up with a good algorithm for this? I currently do a
line-
slope approximation and try to connect-the dots at about 0.1mm
resolution
as I draw diagonals. This follows the line pretty closely, but it's
kinda jiggly. The springiness in the pen and mechanism, the rapid motor
starts and stops, and the processing speed of the RCX seem to militate
against a perfect solution. My orthogonal x and y straight lines are
great, but my diagonals and circles look a little drunken...
Jeff
|
|
Message is in Reply To:
 | | Re: Angle Sensor; 1999 Catalog
|
| (...) I am using two angle sensors in my synchro drive bot and both are rotating at mini motor full speed. There have been times where it seemed to drop a few counts ( approx. 1% ) but I have yet to verify this. I only noticed this because after (...) (26 years ago, 5-Feb-99, to lugnet.robotics)
|
2 Messages in This Thread:
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
Active threads in Robotics
|
|
|
|
|
|
