[chbot] Lego NXT question

[Richard Jones]

Thanks Charles,

With regard to the move block appearing to run very fast after 5 sec or so
even at 50% power setting.
I solved the problem by replacing the move control with two motor control
blocks with feedback enabled. Seems a clunky solution but works fine. 
In this configuration it appears to me that the NXT power control actually
configures speed.
I compared the behaviour of my kit with another one of similar age and
identical firmware version (1.05 I think) running the same program.
Not sure why the move block runs at constant speed on one nxt kit and
starts at a similar speed on a different kit but then runs fast after a few
seconds. Maybe the amount of grease in the gearbox offering a different
load when it gets warm?
Seems like an improvement opportunity for the GUI, a drop down box to
select speed or power control would do the trick.

Cheers

Richard 
(Wearing black shoes just now, but pink is cool! For those who have watched
the video.) 

On Mon, 5 Jan 2009 11:34:47 +1300, Charles Manning
<manningc2 at actrix.gen.nz> wrote:
> On Saturday 03 January 2009 01:12:12 Richard Jones wrote:
> 
>> Now I have a little problem that an NXT expert may be able to advise on.
>> When I program the dual motor control to turn fully at 50% power
> unlimited
>> time, the motors turn at half rate for a few seconds and then run up to
>> about full speed. The robot still turns on the spot so I suspect that
> the
>> speed is still accurately controlled, just at the wrong rate. Anyone
> have
>> any ideas? I've simplified the program to a single motor control block
> and
>> changed my NiMh cells for alkaline cells. My nxt kit is less than 2
> months
>> old so I'm not keen to void the warranty yet by taking it apart. When I
>> load the same program into a different NXT with different motors, the
>> motors turn indefinitely at the slow rate as expected. Any suggestions
> much
>> appreciated.
> 
> Hi Richard
> 
> I assume that by 50% power you mean 50% speed.
> 
> If it starts at the correct speed then slowly ramps up, it sounds like
PID
> loop integrator wind up. The software thinks the motor is turning slower
> than
> the target speed and is increasing the power to compensate but for some
> reason is never reaching  the target speed.
> 
> I would expect that this could be a result of either loss of counting
> feedback
> or a software issue.
> 
> You can easily check that the counting feedback is working by checking
> that
> the motor accurately performs a precise operation such as a 90degree
> movement. If it does that accurately then the motor feedback is working
> and
> it is most likely not a hardware fault.
> 
> Are you using the latest firmware & software? There might be some bugs
> that
> have been fixed.
> 
> Also try turning off Bluetooth etc in case the extra CPU loading is
> messing up
> the counting.
> 
> Also try using, say, 20% or 25%. Does the same happen?
> 
> Still, this does seem vaguely familiar. On the Science Alive session,
> Matthew
> tried running one motor at 50% and the other at 100% and there was
> insufficient difference to make the robot turn properly. He had to change
> the
> mechanism to stopping one when to make enough difference.
> 
> -- CHarles
> 
> _______________________________________________
> Chchrobotics mailing list, web site: http://kiwibots.org
> Chchrobotics at lists.linuxnut.co.nz
> http://lists.ourshack.com/mailman/listinfo/chchrobotics