Jumpy touch pad observation

[Gary C Martin]

On 11 Jul 2008, at 06:06, Richard A. Smith wrote:

> Gary C Martin wrote:
>> Forgive me if I'm about to start an old wives tale, but I've   
>> encountered the jumpy touch pad issues quite a lot on the B4 XO I  
>> have  here for testing. I'm currently running a recent joyride  
>> (2137) and  firmware (Q2D16) and am still encountering issues.
>> Observation: If the XO is fully charged and still plugged in to  
>> mains  supply the jumpiness can be very persistent, however if I  
>> unplug the  power adapter, the XO seems to then correctly re- 
>> calibrate within  seconds.
>
> Its ground plane changes.  According to Alps we are supposed to be  
> issuing a recal on AC insert/removal or and lid open.  And we don't  
> do that yet in the chain of events.
>
> A claim from Alps is that there is so little metal in the XO that  
> the ground plane capacitance has a really large variance based on  
> its environment and whats plugged up to it.  Even after we added  
> more metal in the frame to help.  So your observation falls in line  
> with known data from the mfg.

Thanks, that information make me feel a little more sane for bringing  
this up :-)

> If you are plugged up to mains and its jumpy does a 4 finger recal  
> fix it?

Usually not... Let me clarify. There seem to be multiple[1] causes of  
jumpiness, some of which are fixed just fine by the 4 finger salute.  
The really frustrating sessions are when the 4 finger salute makes no  
difference, for the last month or two I've tried to vary my XO use so  
that I use it both on and off mains power. If the touch pad looses  
calibration and won't re-calibrate, it's is _usually_ when I am still  
plugged into mains and with a full battery charge – unplugging the AC  
and waiting a moment or two usually resolves the persistent failed re- 
calibrations.

With the latest Joyride changes, it seems much clearer/easier to see  
when the driver is trying to re-calibrate as the pointer suddenly  
stops moving for a few seconds. In the persistent case (plugged in &  
charged), leaving the touch pad alone to allow calibration to complete  
usually does not work, and the cursor continues to jump (or float when  
your finger is near but not touching the surface) when you try to use  
it again. Unplug the AC, and the auto re-calibration usually[2]  
succeeds.

> Either way when its jumpy please enable kern.* logging in /etc/ 
> rsyslog.conf, route it to a file, and restart rsyslogd.  Then echo 1  
> > /sys/module/psmouse/parameters/tpdebug
> Then use the mouse jumpy for a while and send me, dilinger, or  
> deepak the log file for us to look at.

OK, will do. Thanks for you feedback.

[1] The common case seems to be touching the pad in two places in  
quick succession, this is easily done when trying to use the left or  
right mouse buttons while moving the pointer as they are placed so  
near the touch surface. I often hit this when using the drag metaphor  
(probably what makes arranging activities on the new random home  
screen layout feel more frustrating than it should be).

[2] I think the cases I've seen where it does not succeed are down to  
damp hands, greasy/wet smear somewhere on pad, perhaps nylon  
underpants (I'm joking!)

--Gary