Actually that was caused by a bug
The KRG-3 refresh rate is very fast!
here you can see the DC value (no rotation) at about 280. Clockwise rotation (when looking at the KRG3 from the front, goes up to about 750 and anticlockwise goes down to 0. Which makes sense because on that Japanese sheet that comes with the KRG-2 it says that the DC is 1.35V so since our Pepper Board V1 runs at 5V with 10 bits of ADC resolution, (1.35/5)*1024 = 276
we (user rustyarmor + me) took it a step further and let the Bioloid swing while measuring the gyro..
perfect decaying sinusoidal
we could actually measure the friction on the hands.
Here's a phone clip of happy engineers at 3am.
<object classid="clsid:D27CDB6E-AE6D-11cf-96B8-444553540000" codebase="http://download.macromedia.com/pub/shockwave/cabs/flash/swflash.cab#version=6,0,29,0" width="128" height="96">
<PARAM name="PLAY" value="true">
<embed src="http://robosavvy.com/site/Builders/limor/bioloid_swinging_krg2.SWF" width="128" height="96" play="true" loop="false" showcontrols="true" quality="high" pluginspage="http://www.macromedia.com/go/getflashplayer" type="application/x-shockwave-flash"></embed>
And here he is swinging in sync with his 3D avatar in Simulink
Here is a rar
file with the CM5 code and the low-level serial device driver S-function that is needed in order to allow Simulink/Matlab Real-Time-Windows-Target to talk with the CM5 while XP is running in "close to real-time" mode.