On 5/31/2012 12:39 PM, andy pugh wrote: > On 31 May 2012 17:20, Dave<e...@dc9.tzo.com> wrote: > > >> If you never under run the buffer you are ok, if you do, they fault out. >> That's how they can run with a non-realtime system like Windows. >> They just spool up a second or more of move data, and then go to work. >> > Yes, but this would be very difficult with LinuxCNC. > > If the original device can be persuaded to trust that the next axis > position will be along in 1mS, guaranteed, and not to fret about an > empty buffer then I can see it being relatively easy, as it can be > driven with a HAL component and RTNet. Otherwise it needs to slot into > the LinuxCNC system somewhere where no other hardware does. > If you look at: > http://git.linuxcnc.org/gitweb?p=linuxcnc.git;a=blob;f=src/hal/drivers/mesa-hostmot2/stepgen.c;h=0538d8a833be8d9f16a10d401f2934a5f01d5f6f;hb=0cd2f98d522b87d08f302260ceb571a6b4b46010 > for example, that is the driver for the Mesa Hostmto2 cards, and it > basically sends a new step-rate command every servo-thread. > (the step rate required is calculated from the new position request, > the stepgen scale and the current position) > >
Well apparently Yishin Li must have something figured out with a buffered setup. I am assuming that is required since he is using a USB interface. But I have not had a chance to look at his code. Dave ------------------------------------------------------------------------------ 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-developers mailing list Emc-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-developers
