New Plasma build question

You would expect an error if the stepgen MAX_ACCEL and stepgen MAX_VEL
are not at least 25% greater than the joint values

PCW wrote: You would expect an error if the stepgen MAX_ACCEL and stepgen MAX_VEL
are not at least 25% greater than the joint values

Thanks Peter I have run out of time now until next Friday I will get back to it then.

Clive s wrote:

PCW wrote: You would expect an error if the stepgen MAX_ACCEL and stepgen MAX_VEL
are not at least 25% greater than the joint values

Thanks Peter I have run out of time now until next Friday I will get back to it then.

Just could not leave it.
[AXIS_Z]
MIN_LIMIT = -100
MAX_LIMIT = 1
MAX_VELOCITY = 10
MAX_ACCELERATION = 12


[JOINT_1]
AXIS = Z
MIN_LIMIT = -100
MAX_LIMIT = 1
MAX_VELOCITY = 10
MAX_ACCELERATION = 12
TYPE = LINEAR
SCALE = -200
STEPGEN_MAX_VEL = 10.0 )tried 12.5
STEPGEN_MAX_ACC = 12.0 )tried 15 same error
FERROR = 1
MIN_FERROR = 0.25
DIRSETUP = 5000
DIRHOLD = 5000
STEPLEN = 5000
STEPSPACE = 5000
DEADBAND = 0
P = 1000
I = 0
D = 0
FF0 = 0
FF1 = 1
FF2 = 0.000127
BIAS = 0
MAX_OUTPUT = 0
MAX_ERROR = 0
HOME = 0
HOME_OFFSET = 2
HOME_SEARCH_VEL = 8
HOME_LATCH_VEL = 2
HOME_SEQUENCE = 1
HOME_USE_INDEX = False
HOME_IGNORE_LIMITS = True


With the same error

Yeah something basic is broken, not sure what

What LinuxCNC version is this?

Can you plot the commanded Z position and the actual (feedback) Z position when homing?

FF2 should be 0

tommylight wrote: FF2 should be 0

I have just tested it with FF2 0 no change same error

It will home without error when P = 10 instead of 1000

But then I cannot get the speed up.

Ps. Hope to see you Monday ish.

Small amount of FF2 is helpful (and the .000127 looks about right) the problem is elsewhere

The actual amount of FF2 required is the seconds between reading
the stepgen position and writing the stepgen velocity, but this is a very
minor tweak, there are major problems somewhere

Clive s wrote:

PCW wrote: You would expect an error if the stepgen MAX_ACCEL and stepgen MAX_VEL
are not at least 25% greater than the joint values

Thanks Peter I have run out of time now until next Friday I will get back to it then.

Try 50%. I found f had the same problem and the errors went away when I increased these.

rodw wrote:

Clive s wrote:

PCW wrote: You would expect an error if the stepgen MAX_ACCEL and stepgen MAX_VEL
are not at least 25% greater than the joint values

Thanks Peter I have run out of time now until next Friday I will get back to it then.

Try 50%. I found f had the same problem and the errors went away when I increased these.

If I get time, I will try that in the morning before I have to drive down the country.


Peter is there any chance it could be the 7i96, it is new but I do have another one that I could rob from another control box .

rodw: The stepgen headroom should never need to be more than 25%
(1.25 X) if you needed more its a symptom of something else setup wrong,
though you may need double this (2.5X) if you have backlash compensation
enabled.

Clive s:
I tried your config and it works here after doing the fixes i suggested (on all joints)
I also did a faster mod of your configuration (200 mm/s, 2000 mm/s^2)
(all stepgen settings 250/2500) and it works fine (pictures below)
Its very unlikely a 7I96 problem, that would show up in different ways

tommylight: here's what FF2 buys you:

No FF2:



correct amount of FF2:



Note that this corrects a _very_ small error (less than 1/2 micron at 1/5 G accel typical of a CNC machine)
but can be significant on devices with high acceleration and high accuracy (say laser plotters)