On Fri, 2011-02-25 at 08:48 -0800, Kirk Wallace wrote: > On Fri, 2011-02-25 at 15:46 +0200, fi wrote: > ... snip > > Is hal_parport.c the proper file to modify for 16 bit I/O over PATA > > ports ? > > Are other files to be modified ? > ... snip > > >From my study of the parallel port drivers, 90% of the code is for > setting up the structs for managing the setup strings, and sharing > multiple ports. When that is out of the way, it comes down to a simple > outb or inb to a few registers. I suspect the PATA interface might be > the same except that 90% of the code will be PATA specific, so the > parport code may not be helpful. > > If one is going to go through the effort of a new design, I would vote > for pursuing a PLX based PCI or PCI-X interface like this one: > http://wiki.linuxcnc.org/uploads/pci1_15_02_sch_wiki.pdf > (from this page: > http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?AVR ) > > (It seems to me, the old ISA slots would be perfect for real-time > interfacing.) > > I am guessing here, but the only "active" part of the parallel port > driver is in/outb which lives well with real-time. This may not be the > case with PATA or PCI. Study of the PCI FPGA examples may shed light on > this. > > On the other hand, if your time is money, just buying an existing PCI or > PCI-X motion card would be cheaper.
Hi Regarding obsolescence, there are PCI/PCI-EXPRESS to IDE PATA adapters which may already work with pci_8255.c driver and 8255 boards but what I would love to get is 16 bits wide I/O , so for now I'm looking at hal_skeleton.c and hal_parport.c OTOH, I used an internal PATA HDD Rack for some testing with _normal_ kernel. I've played with some 573 574 and 245 , with 245(not being a latch) were fast pulses which maybe are useful for fast steps in step/dir configurations. Next weekend I'll try a PCI to IDE adapter . OT 1: is possible to make timing of step pulse to be half of desired moving time like 1ms pulse for a 2 ms move or 3s pulse for 6s move? Instead of many short pulses to give just a single pulse with half the timing ? EMC's stepgen knows before writing the pulse how long a move will take so it will be easy for a modified step/dir servo to do optimized moves. OT 2: I have bad latency PC's so here: http://wiki.linuxcnc.org/emcinfo.pl?Latency-Test it says: "Numbers over 1 millisecond (1,000,000 nanoseconds) mean the PC is not a good candidate for EMC, regardless of whether you use software stepping or not." >From a hobby point of view I don't care how much time it takes to mill a good PCB as long as it is good ; so according to the statement above I can't use EMC2 ? Sorry for my english Bye. Florin ------------------------------------------------------------------------------ Free Software Download: Index, Search & Analyze Logs and other IT data in Real-Time with Splunk. Collect, index and harness all the fast moving IT data generated by your applications, servers and devices whether physical, virtual or in the cloud. Deliver compliance at lower cost and gain new business insights. http://p.sf.net/sfu/splunk-dev2dev _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
