Subject:
|
Re: Rotation Sensors - slightly dissapointed...
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Sun, 27 Feb 2000 02:34:06 GMT
|
Viewed:
|
867 times
|
| |
|
|
In lugnet.robotics, Ralph Hempel writes:
> > If you check the legOS group you can see my posts with some extensive • analysis
> > of the sensor values for the rotation sensor. There is a patch to make • legOS
> > less sensitive to voltage fluctuations and one to debounce the input (which • is
> > what probably causes missed counts on the standard firmware as well).
>
> I have to go back over the changes, but does legOS sample at the same rate
> as the standard firmware? The standard rate is about 333hz, so debouncing • would
> likely reduce the effective rate to 166hz, which limits the max rate at which
> the sensor can keep up.
I believe it is. I tried setting the sampling rate to high speed and it just
didn't work at all. Yes, debouncing the input halves the effective clock
frequency, but the problem is that single-clock transitions at slow speeds are
always transitory and result in counting errors. The sensor itself is very
"clean" -- it will never sit halfway between two values, but it can move
between two discrete values during the A/D conversion, and in one transition
case this results in a false reading for another state.
--Ben
|
|
Message is in Reply To:
 | | RE: Rotation Sensors - slightly dissapointed...
|
| (...) Ben, I have to go back over the changes, but does legOS sample at the same rate as the standard firmware? The standard rate is about 333hz, so debouncing would likely reduce the effective rate to 166hz, which limits the max rate at which the (...) (25 years ago, 25-Feb-00, to lugnet.robotics)
|
16 Messages in This Thread:
 
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
