On 4/9/2010 12:04 PM, Jon Elson wrote: > The problem with the homing report is the destruction of data. When the > home position is found, the machine position count is set to zero (or > whatever the HOME_OFFSET value is). So, the old home position is > destroyed by that operation. To stay with the canonical encoder > definition as it is, you would have to store the position JUST before > the home operation completed. I think you could make a custom HAL > component that recorded the machine position every servo cycle. > Whenever it saw ENCODER_INDEX transition from true to false, it would > save the value recorded from the previous cycle. This would be the > position one millisecond (at the default SERVO_THREAD rate) before the > home position was found. I think this would give you a pretty > consistent way of checking for drift of the home position. On the first > home whe EMC is started, it would show the distance traveled from where > the machine was to the home position. On later homing operations, the > value should be very close to zero, assuming the approach to the home > switch is slow. (Maybe watching ENCODER_INDEX only works where the > index pulse is being used, how does the position count get set for > stepgen, for instance? I'm only familiar with servo-like interfaces.....) > > I think this will work if there's a signal that tells you the home > operation is happening NOW, and doesn't require any change to EMC2 > itself. You would monitor this with HalMeter or the show hal signals menu. > > Jon > > > Hi Jon,
Slavko is running steppers so I believe he has a base thread running also which should be plenty fast enough to catch the step count just before hitting the home switch. :-) And if not, then I would think that a base thread could be added to increase the accuracy. Just has to be worked out. Dave ------------------------------------------------------------------------------ Download Intel® Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance. See why Intel Parallel Studio got high marks during beta. http://p.sf.net/sfu/intel-sw-dev _______________________________________________ Emc-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-users
