zero return with stepgen 7i76, 7i77, and 7i92

Hi,
I would like to use the index pulse when referencing my machine.
I have my encoders connected, as I was using closed loop, but because I struggled too much to get a good surface finish, I added a 7i76, and am now using step/dir functionality on the servo drives. My question:
Can I use the zero pulse when homing for acurate homing?
and my next question, I would like to know the secret in pid tuning. I just can't get a smooth finish on this machine, That is my reason for adding a 7i76, With pulse/dir I get very nice finish, but ... I'm having concerns about homing acuracy, and hope someone can point me in the right direction

It is possible to still have and use full encoder feed back while using step/dir commands. Configuring homing to index is exactly the same regardless of the command signal (step/dir or analog servo).

As to the tricks to servo tuning??? I am no master.

It might be as simple as just adding "HOME_USE_INDEX = YES" to the INI file.
(Or, depending on the setup, it might not).
If you are using a 7i76 then are the encoders even connected to LinuxCNC?

Yes. I was using 7i77 at first, closed loop analog , but due to poor surface finish, I added a 7i76, controlling servos with pulse/dir. This gives excellent surface finish, but now having accuracy problems. Homing precision and ovality on fast circular interpolation

I tried that, but then on final approach the axis goes to completely different destination. Not sure what it's doing, I am going to play some more today though

I managed to use encoder feedback via 7i77, and pulse output via 7i76, managed to get it to run closed loop, but same problem with poor surface finish. I played around with pid tuning, but no joy. I'm contacting the Servo manufacturing company to find out more information on tuning on the servo drive side. It's a DS2-42P3-AS and MS-130ST-M15015B-22P3 servo pack from Xinje. I've used many of these servo's with great success, but am now battling with this machine, because of the fast cutting speeds required.
So I was struggling with the homing only setting HOME_USE_INDEX = YES, so decided to write a component to do the homing closed loop, then switch over to stepgen feedback after axis.n.homed = 1, but this gave me following errors, So this problem is still not resolved...
I'll update my progress as I go, but if anyone has suggestions, I'll appreciate it.
My initial problem is I basically can't get a good surface finish at high cutting speeds. If I can somehow get a component or a way to fine tune the PID settings, That would really be the solution.

When using real encoder feedback with a Mesa hardware step generator, I believe that you will need to use much lower P settings than the default P=1000 that you use with a normal open loop stepper config.

It might be helpful if you try not connecting the encoder feedback to the PID loop (leave the stepgen configured as an ordinary open loop system) the way that you were getting good results. Then use the encoder feed back to monitor what is really happening with Halscope. This could give you a baseline of what you are aiming for when tuning your PID for the closed loop system. Then try reconnecting the encoder to the PID and reducing P until the halscope plot looks about the same as the open loop test. If you can't get there. You could just run with the open loop PID setup, and only connect the encoder feed back to the motion pins.

Diederik wrote: I managed to use encoder feedback via 7i77, and pulse output via 7i76, managed to get it to run closed loop, but same problem with poor surface finish. I played around with pid tuning, but no joy. I'm contacting the Servo manufacturing company to find out more information on tuning on the servo drive side. It's a DS2-42P3-AS and MS-130ST-M15015B-22P3 servo pack from Xinje. I've used many of these servo's with great success, but am now battling with this machine, because of the fast cutting speeds required.
So I was struggling with the homing only setting HOME_USE_INDEX = YES, so decided to write a component to do the homing closed loop, then switch over to stepgen feedback after axis.n.homed = 1, but this gave me following errors, So this problem is still not resolved...
I'll update my progress as I go, but if anyone has suggestions, I'll appreciate it.
My initial problem is I basically can't get a good surface finish at high cutting speeds. If I can somehow get a component or a way to fine tune the PID settings, That would really be the solution.

were you able to get this resolved? A friend of mine and I are both doing the same thing as you, using a 7i76 to use pulse output to our servo drives and a 7i77 for encoder feedback, we are at the very beginning of the setup functionality through Linux and are currently getting following errors when using the encoder input. this eems like a very similar situation to what you are experiencing and we were curious if you have any updates on the setup?

Thanks in advance!

Chris

Can you provide a halscope plot of the following error (set the ferror limit large so you dont cause a following error fault)