7i40hv pid

More
9 years 1 month ago #98448 by billykid
7i40hv pid was created by billykid
hello
i am going to use 2 drivers 7i40hv can you tell me what terms PID to use?
I've read jhon's example but talk about velocity mode, these are nude bridges I think.
as a starting point,% of P of ff1 ff2.
regards
Mauro

Please Log in to join the conversation.

More
9 years 1 month ago #98460 by andypugh
Replied by andypugh on topic 7i40hv pid
I would start with just P at 0.00001 then make it 10x bigger until something happens.

Please Log in to join the conversation.

More
9 years 1 month ago #98499 by billykid
Replied by billykid on topic 7i40hv pid
thanks Andy
the pid works already very well with only P = 160.
more than anything I was interested in knowing which tuning parameters and what not.
I'm testing with an unmanned engine I know is not a good condition
but g1 x10 f1000 stops + - 0.001 mm.
according to you is a good system with this closed loop or better how do I have drives that hold their position commanded step-dir?
regards
Mauro

Please Log in to join the conversation.

More
9 years 1 month ago #98504 by andypugh
Replied by andypugh on topic 7i40hv pid
An PID will tend to stop at the right place. You should plot how well it follows dynamic movements using halscope.

You will probably find that some I-term is needed to remove steady-state error.
D-term might let you run more P, for a stiffer response.

But, until the servo is mounted to the machine and is under load there is no point taking the tuning too far.

Please Log in to join the conversation.

More
9 years 1 month ago #98508 by billykid
Replied by billykid on topic 7i40hv pid
I have four homemade drives and work fine
elm-chan.org/works/smc/report_e.html
as I'm now making a small plasma and I wanted to use 7i40hv + 5i23
I was interested in knowing whether the results will be as good as other drivers
the next few days I will try to calibrate the pid on a real axis, but I'm quite satisfied
sorry but I can not explain well with the translator

Please Log in to join the conversation.

More
9 years 1 month ago #98575 by billykid
Replied by billykid on topic 7i40hv pid
I did the tests work very well.
now i use a 1024 encoder but i think it's a little for a closed loop system.
I would have found this: HEDR-55L2-BY09 3600 cpr line driver. with 5i23 I should not have counting problems.
Usually what do you use as a resolution?
regards
Mauro

Please Log in to join the conversation.

More
9 years 1 month ago #98582 by andypugh
Replied by andypugh on topic 7i40hv pid
1024 lines or 1024 counts?

Assuming a 5mm pitch screw and that f-error is 4 counts (typical) then the accuracy of the system will be +/- 0.02mm with a 1024 count encoder (0.005mm with a 1024 _line_ encoder).

What do you need?

Please Log in to join the conversation.

More
9 years 1 month ago #98590 by billykid
Replied by billykid on topic 7i40hv pid
in fact you are right.
I'm 1024 cpr, I wanted to change them for two reasons: when I get the 7i40 i would like to use differential encoders (my not) and index.
I thought I would use HEDR-55L2-BY09 3600 cpr which I can replace on the fly and have both features.
they will go 100khz maximum i do not think there are any problems.
I found the output-scale parameter at 10 from pnconf is right for this configuration?
thank you
Mauro

Please Log in to join the conversation.

More
8 years 6 months ago #109703 by billykid
Replied by billykid on topic 7i40hv pid
Hello
finally I mounted everything, calibrated the pid only with P, D, FF1, connected encoder -velocity to pid.n.feedbak-deriv as said Pcw.
but it seems to me even more noisy.
I was in doubt ... output scale must be 1? I saw that others put 10 but maybe for analog drives.
the error as you can see from the picture is within 0.005 mm at F200 it seems good to me.
thanks for your help
Mauro

Please Log in to join the conversation.

More
8 years 6 months ago #109706 by andypugh
Replied by andypugh on topic 7i40hv pid

billykid wrote: connected encoder -velocity to pid.n.feedbak-deriv as said Pcw.
but it seems to me even more noisy.


What is your encoder/resolver velocity scale? If the encoder scale is positive then the velocity scale should be positive (probably +1) but if your system requires a negative velocity scale then you need a negative velocity scale too.

(I was caught out by exactly this a couple of weeks ago)

Please Log in to join the conversation.

Time to create page: 0.097 seconds
Powered by Kunena Forum