Upgrading to a motion control board

More
03 Aug 2012 20:15 #22736 by andypugh
dangercraft wrote:

On the Minas drivers, and most of the other AC drivers I've come across, you can usually specify one of three types of control schema: position, torque and velocity. I'm not really quite sure what torque is useful for, but velocity control is what you are referring to where it wants +-10v DC. This is akin to using it as a spindle drive.

In position control you use pulse and direction precisely the same way as you control a stepper motor.

This way the amplifier continually updates the amount of power required to accelerate or decelerate according to the changes in input with the least amount of error and lag.


Ah, but this isn't the "LinuxCNC Way". The design principle of EMC (back before EMC2 even) was to move all the processing off of special-purpose hardware onto cheap generic computer hardware.
So, a typical LinuxCNC servo system uses the drives in velocity or torque mode and runs the position control loop in LinuxCNC. (one advantage of this is that you can see exactly what is going on (with Halscope) and change it)
LinuxCNC can comfortably run a 2kHz servo loop, and that is plenty fast enough, especially considering the raw horsepower of a desktop CPU.

Please Log in or Create an account to join the conversation.

More
03 Aug 2012 21:37 #22741 by dangercraft
Erik,
I am very interested in learning about what you just mentioned, but I don't want to hi-jack the thread so to speak.

I've been meaning to ask someone about how to get the original NIST EMC documents and source because I'd like to start from there and then move into how it changed into EMC2. I don't want to toot my own whistle, but I think with some (more like a metric-shit-ton) of study, I may be able to contribute to the developement of linuxcnc.

Frank

Please Log in or Create an account to join the conversation.

More
03 Aug 2012 21:41 #22742 by dangercraft
In regards to the amplifier calibration, you are right, this is mostly useful only when you are sending a pulse train and its so the amplifier can calibrate itself to try and eliminate the response times. On the other hand, I don't see how it can hurt to offload that from linuxcnc's duties if you already have the hardware or the means to aquire it. this can free time and resources, for instance, for a proper look-ahead functionality.

Frank

Please Log in or Create an account to join the conversation.

More
04 Aug 2012 12:04 #22762 by BigJohnT
I've googled for the NIST EMC docs before and found them... never tried to find the original code but I assume it is out there somewhere.

John

Please Log in or Create an account to join the conversation.

More
14 Aug 2012 16:16 #23171 by dangercraft
Andy,
I've been looking through the documentation and I am starting to understand what you are saying about using velocity controls and encoders to linuxcnc. I posted a thread about this here:

www.linuxcnc.org/index.php/english/compo...ew&catid=38&id=23170

I'm hoping maybe you can take a look at it and correct me if I am wrong.

Thanks,
Frank

Please Log in or Create an account to join the conversation.

Moderators: PCWjmelson
Time to create page: 0.212 seconds
Powered by Kunena Forum