Nachricht geschrieben von Tomasz Motylewski
>
On Mon, 9 Aug 1999, SteenbeckDigital wrote:
> Tryed the example with a 300MHz Pentium and
> could hear the machines load from the servo.
> Using a professional RC-servo I could
> even see the machines load.
> The jitter is slightly to big for this.
> This shows the limit of RtLinuX.
What jitter did you expect? You will not get below 1 us with any program
running on a PC Intel x86 withiout some special hardware.
If you want to reach 1 us limit, you need to schedule your interrupt about
20-30 us before the deadline and then wait reading the clock counter - I
guess you do the same when you really want to catch the train :-)
Expect 5-10 us jitter with standard RT-Linux.
<
############################
I didn't expect something different.
Just mentioned it to the group so not everybody
runs into trouble relying on RC-servo control.
Would be interested if polling the time could
fix the jitter for this application.
############################
Stuart Warren wrote
I tried this myself and didn't have any trouble. That was on a P123 and
RTL V1.0 For a while there I thought I had a problem with the timing
accuracy but it ended up being my power supply not being able to handle
the load of the servos.
#############################
- For my test I took the power from an extra power supply.
- Where did you get the P123 from???
- What kind of RC-servo did you use??
Only the slow servos didn't show the symthom to me.
The fast one showed the jitter very good.
- Is there a patch to RtLinuX where the jitter has been reduced??
#############################
Jens Michaelsen
--- [rtl] ---
To unsubscribe:
echo "unsubscribe rtl" | mail [EMAIL PROTECTED] OR
echo "unsubscribe rtl <Your_email>" | mail [EMAIL PROTECTED]
----
For more information on Real-Time Linux see:
http://www.rtlinux.org/~rtlinux/
