ColorCNC Colorlight 5A-75E/5A-75B as FPGA controller board

About the speed, consider that my approach to stepgen Is velocity only mode, so It accept a velocity command input (and provide the corresponding step frequency on it's pins) and provide internal position feedback based on number of steps emitted.
But you need to provide externally the position control, trough a pid component. Much like a servo.
This way i simplified my work on stepgen validation and pid loops are already available and validated in linuxcnc Hal.
So in summary that depends on you configurazion, I Need some time before arriving where you are with my hardware.
About inputs I'm not sure, are you referring to gpio or encoders?
Encoders are reset when the board Is reset, since in that state the board feedback could not be trustable. Maybe this behavior can be made more smart.

About inputs - i mean gpio.

I see that you are using very high settings for stepgen times, step width and space at 5000.
Combined with a scale of 640 pulses per mm this will slow down you maximum speed.
Consider that the times are expressed in microseconds and not nanoseconds, so 5000=5ms.
This parameter Is internally clipped anyway, but maybe you intentions are for lower times.

About gpio I'm not getting the problem. 
Is some gpio PIN not being read correctly?

muvideo wrote:

I see that you are using very high settings for stepgen times, step width and space at 5000.
Combined with a scale of 640 pulses per mm this will slow down you maximum speed.
Consider that the times are expressed in microseconds and not nanoseconds, so 5000=5ms.
This parameter Is internally clipped anyway, but maybe you intentions are for lower times.

Yes, i mean 5000ns. Change from 5000 to 5 and WOW!!! It's worked!!!!
Will try on real hardware!

About gpio I'm not getting the problem.
Is some gpio PIN not being read correctly?

I'm use binary coding switch DCRS-01J connected to GPIO IN.

# --- FO-INCR ---
net fo-incr-a <= [FABIO](CARD0).input.14.in
net fo-incr-b <= [FABIO](CARD0).input.15.in
net fo-incr-c <= [FABIO](CARD0).input.16.in
net fo-incr-d <= [FABIO](CARD0).input.17.in
net fo-incr-e <= [FABIO](CARD0).input.18.in

# connect feed overide increments - switches

setp halui.feed-override.count-enable true
setp halui.feed-override.direct-value true
setp halui.feed-override.scale .01
net feedoverride-incr => halui.feed-override.counts
net fo-incr-a => foincr.sel0
net fo-incr-b => foincr.sel1
net fo-incr-c => foincr.sel2
net fo-incr-d => foincr.sel3
net fo-incr-e => foincr.sel4
net feedoverride-incr <= foincr.out-s
setp foincr.debounce-time 0.200000
setp foincr.use-graycode False
setp foincr.suppress-no-input False
setp foincr.in00 0.000000
setp foincr.in01 2.000000
setp foincr.in02 4.000000
setp foincr.in03 6.000000
setp foincr.in04 10.000000
setp foincr.in05 20.000000
setp foincr.in06 30.000000
setp foincr.in07 40.000000
setp foincr.in08 50.000000
setp foincr.in09 60.000000
setp foincr.in10 70.000000
setp foincr.in11 80.000000
setp foincr.in12 90.000000
setp foincr.in13 95.000000
setp foincr.in14 100.000000
setp foincr.in15 105.000000
setp foincr.in16 110.000000
setp foincr.in17 115.000000
setp foincr.in18 120.000000

On start LinuxCNC, REAL state of this switch is not readed, and always set to 100%.
If I switch to any non-current position, then only the REAL switch position is read.

Link to video. Feedrate switched by my device.

Oh my god!
Not updated inputs on start LinuxCNC is a very old problems and not solved for a now!!!
github.com/LinuxCNC/linuxcnc/issues/139

New bug