Mesa cards are add-on cards that offload the high-speed tasks to an FPGA.
The FPGA runs a firmware known as "
Hostmot2
"
The interface between LinuxCNC and the card is in HAL, but at one level above that which you configure with the parallel port.
So, with a parallel port you connect motion to a stepgen, and then choose which physical pin to use for the step output, with a Mesa card you connect motion to a stepgen input pin on the FPGA interface, and the firmware on the card determines which physical pin is step and which direction.
A parallel port configuration which contained:
Code:
net x-pos joint.0.motor-pos-cmd => stepgen.0.position-cmd
net x-step stepgen.0.step => parport.0.pin-01-out
Would look like the following with a Mesa card:
Code:
net net x-pos joint.0.motor-pos-cmd => hm2_5i25.0.stepgen.00.position-cmd
Some cards load the firmware at every startup from the /lib/firmware directory. (5i20, 5i21, 7i43) whereas some have it permanently loaded and require it to be changed using the
mesaflash tool
Mesaflash can tell you the physical pin allocation of the installed firmware, but finding a firmware with a specific pin allocation is more difficult. One way is to download the Hostmot2 source file from Mesa and to look through the .vhd files (which are just about human-readable)
Mesa FPGA cards can connect to LinuxCNC through
PCI, PCIe
, EPP parallel port (
7i43
,
7i90
),
Ethernet
and
SPI
(
Pi version
) interfaces.
Many FPGA cards can additionally connect to expansion cards using the
Mesa "smart serial" interface
via CAT5 cables.
