Setting up Servos in step/dir position mode

The reasoning is that it is impossible to generate a 100 KHz step rate (as required by the 50 mm/sec motion and 2000 steps/mm scaling in your .ini file (50mm/sec*2000 steps/mm = 100K steps per second) with the given step timings.

That is, the maximum possible step rate (without violating the step timing) is 1/(steptime+stepspace) which = about 99 KHz with your 5050 ns steptime and step space settings. If you drive can use shorter timings set these (5050) values lower
and higher step rates will be possible.

Maybe it is a stupid question but when I am driving servos I should not use INPUT_SCALE and ENCODER_SCALE instead of SCALE and STEP_SCALE forgetting about STEPGEN_MAXVEL / MAXACCEL or this is because being the step & dir signal for .INI purposes it is do they treat like steppers?

Thx.

Understood, they will be treated as steppers as long as the command signal is step/dir and it comes determined by the type of controller card.

RoberCNC wrote:

Understood, they will be treated as steppers as long as the command signal is step/dir and it comes determined by the type of controller card.

Its not really the command signal that's important, it s whether or not LinuxCNC gets
external position feedback from an encoder.

Ok, then to treat as pure servos you must connect encoders feedback, but then, as you said before, problems come unless you have scales to fully close the loop.

I will first get acquainted with them in step/dir mode which is how little I know so far and then I will touch on the subject of backlash (pending subject) before going to the next level.

I really appreciate the help and info mates.

Hello again, let's see if you can help me (sure), I'm doing tests and running the machine and it happens to me several times that suddenly one of the servos goes into an alarm state, with the consequent annoyance, I can't give with the reason and I can think of how to identify the problem.
The fault is in one of the Y's (it is dual) and the work is a rectangular matrix of circles. I identify the fault because suddenly he observed that instead of circles there are ovals. What could it be due to?
How can I solve it?

I would disconnect the servos from the machine and see if you still get a servo fault
just to eliminate mechanical problems

If it seems that the problem comes from the mechanics. I have realized that it would be very useful and interesting to take the Alm signal that the drivers have to connect each one to an input of the 7i76 and in case it happens for any reason to know which driver has failed and all the machining stops, at least the pieces won't be ruined, I hope. Looking at HAL I have found the pin joint.x.error, which I am not sure at all is if I am right and it would indeed be that pin where to link the alarm signal or how to do it exactly.