PID configuration loss.

this is a picture of my PID, I on the right track.

John

or some other kind soul who can help me.

know from my limited language skills, additional limitation in my knowledge.

But just to be sure what I'm doing.

The idea to set PID is as follows.

After reviewing encoder and connection
maximize the P, with everything else at 0 when having oscillation axis, return the maximum point without oscillation.
start moving the shaft and check the error. with halscope
with that error. FF1 increase. to be controlling it.
up to an ideal. negligible error.

well now my question in that time comes I, D, FF2, deadband, bias?

besides, I want to know if what I did with my encoder is fine.

encoder_scale to configure, mount my servomotors, when he told linuxcnc 1mm, had a dial indicator, I recorded more, then started down the scale encoder until as close as possible, this is right?


thanks for everything

friend something I did not understand it is:

Assuming the Following is the drive itself is tuned. The axis Should not drift When enabled and no position is Commanded. If not set the Zero pot if there is one on the drive.

you mean this drive enable and not controlling linuxcnc not have cable connected 14-56?
with those conditions my servomotors have minimal movement. very minimal. this is ok?

or should I move on, zero pot

Peter mesanet technology does not support fanuc servo?

a friend saw my work, and was interested.

Thank you.

John,

I think I'm close. halscope upload photos tomorrow

I have a question.

John in your tutorial you write this:

The low P is only for setting FF1 not for final tuning.

I managed to move my machine to 5000mm/min,

i ranges over 5000mm/min end,

I have the P in 10. others in Zero.

if you increase the "P" to more than 10, 5000mm/min oscillate.

let me say what restless. P low. is only for setting of FF1, not for final adjustment.

because:

P = 10
I = 0.2
D = 0.1
FF1 = 0.31

I have managed to significantly reduce the error.

strange thing I notice is that when I move from 0 to -150mm, the error is greater than when I move from -150 to 0.

also when the machine goes from 0 to -150mm, when get to -150 feels hard braking. as a blow.

but if the rate increase will 7000mm/min. everything loss achieved. (oscillate)



questions:

That would mean more down the P and others to zero? and start again?

to adjust the error really. Should I do it with maximum velocity (10000mm/min)?

not fail if P is very near 0?

Although I have doubts as to establish the max_acceleration, as I can be sure it is correct?

is normal for now I'm adjusting the PID, which feel a blow to the reach of the required position? ( abrupt braking)

thanks.

If you have differences in errors that are dependent on motion direction, I would carefully check the axis mechanics.

Another note: do not use _any_ I term (leave it 0) until you are completely satisfied the the tuning. The I term will reduce the static or long cruise error but will just confuse things if applied before everything else is completely correct.

FF1 is tuned by setting P low (say 1) and adjusting for minimum following error during a long medium/fast speed slew

(with too little FF1 the axis position will lag the commanded position, with too much, the axis position will lead the commanded position)

Once FF1 is set you need never change it (unless you change some setting on the drive or some ini file setting the changes the input or output or feedback scale) , FF1 is independent of other tuning parameters so once set, should be left alone

If it oscillates you need less P rather than more P, but you should make sure that the drive itself is stable (tachometers in particular should be checked)

You can do this by driving the axis back and forth open loop by simply applying a 0-10V DC voltage to the drive analog in pins.

Obviously you have to do this carefully so you dont run it into the stops

If the drive itself is OK, it should slew smoothly (no oscillation or erratic motion) at any speed 0 though 10V

Peter,

understand.


Regarding max_acceleration? to define correct?

Tomorrow I corresponding tests.

thanks