The encoders have a resolution of 2um, so I don't think losing a few counts should hurt me too badly. I've done a fair amount of testing to verify repeatability and it looks very good. I used a custom "sample" component yesterday to capture the two encoder counts on rising and falling edges of the "in position" signal. I'm aiming to have a modular motor with indefinite stroke length, simply by lining up more 6" coil sections and adding more encoder heads at 3" intervals. For now I'm driving all coils all the time, but after I get this encoder transfer working I'll work on shutting off coils that aren't actively driving the carriage.
My efforts so far involve mux2's and offsets, as the offset between the two encoder heads is fixed and measurable with my sample component. I'll have to think some more about Andy's idea of keeping track of all the counts inside a component and deciding which value to deliver. It may not actually be much different than what I'm doing with mux's now. Thanks for the ideas and suggestions! -- Ralph ________________________________________ From: Chris Radek [ch...@timeguy.com] Sent: Tuesday, July 10, 2012 10:10 AM To: Enhanced Machine Controller (EMC) Subject: Re: [Emc-users] ideas for "bumpless" transfer between encoders On Tue, Jul 10, 2012 at 03:49:19PM +0000, Ralph Stirling wrote: > As I am moving through my linear motor project, I've come down to > a place where I need some ideas. The carriage for my motor is 4in > long, and my coil sets are 6in long. The magnets are on the moving > carriage and the coils are stationary. Part of the reason for this is > that I don't want electrical wires going to the carriage; it is entirely > passive. For feedback I use AS5311 linear magnetic encoders, with > a 4in magnetic strip attached to the carriage. I am placing the > encoder sense heads every 3in along the track. This guarantees that > the carriage position is always being measured by one or two sensors. I am not sure this is possible. It seems like it would be very sensitive to the distance between read heads. The basic problem is that you need to switch between sensors during the part of travel where you have information coming from two of them. If they are not counting exactly together, and I am sure that your setup can not guarantee that they are, I think you will tend to lose one (or maybe four if these are quadrature-based) count worth of position certainty every time you switch. Aside from that do you know for sure that the sensors handle going off the end of the strip and back on in a deterministic way? You're asking about implementation details, which are tricky enough, but are you sure the physical design can work at all? I am pretty sure you could rig up a way to use these as velocity feedback, but I am not seeing how you can get correct reliable position feedback. ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
