Subject:
|
Rotation Sensor Accuracy [was Re: Newbie needs Help (diff sensor)]
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Thu, 8 Jun 2006 04:02:24 GMT
|
Reply-To:
|
<dickswan@sbcglobal.netSTOPSPAMMERS>
|
Viewed:
|
4073 times
|
| |
|
|
Steve wrote on Tuesday, June 06, 2006 5:59 PM
<<snip>>
> If you are going to use a rotation sensor to measure the slippage then
> you need to realize that the sensor is notoriously poor at very low
> speeds.
<<snip>>
I think what is really meant is that most RCX firmware implementations
are poor at filtering hardware transient errors. And the Rotation sensor
is very "good" at generating transients at a typical rate of about 0.5%
errors. At very slow speeds, the transient error rate increases
dramatically.
Usually, you get a single transient error. However at very slow speeds,
it's quite possible to get several consecutive faulty transient readings
in a row -- I've seen up to four in a row at very low (about 1 complete
rotation per minute) speeds.
The standard Lego firmware provides no transient filtering. BrickOS
filters single transients but always fails when there are two or more
consecutive transients. The RCX replacement firmware used in Robolab 2.9
is the only RCX firmware that I know that properly handles consecutive
transients.
In the interests of full disclosure, I should mention that I wrote the
Robolab 2.9 RCX firmware.
|
|
Message is in Reply To:
 | | Re: Newbie needs Help (diff sensor)
|
| (...) I think you need another pair of gears on one of the wheels to reverse its direction. If you are going to use a rotation sensor to measure the slippage then you need to realise that the sensor is notoriously poor at very low speeds. You'll at (...) (18 years ago, 6-Jun-06, to lugnet.robotics)
|
39 Messages in This Thread:
 
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
