Add encoder calibration documentation

[anirudhupadhyaya]

No description provided.

[npetersen2]

@anirudhupadhyaya requests that I review this PR in terms of direction of content

[npetersen2]

@anirudhupadhyaya this is on the right track and seems to be written to the right audience, i.e., a user of the AMDC who is trying to interface an encoder to their control system.

There are a few typos and awkward sentences which should be resolved.

However, my main feedback would be to add a "master" block diagram which shows all the possible steps in processing the encoder integer feedback value into something useful for the control system. Then, replace your C code blocks with a single block of C code which implements the block diagram for "all cases" i.e., CW or CCW, etc.

I am not sure about your while() loop for performing the mod(x, 2*pi) operation. Probably better to just use mod() here unless you have a good reason not to use it.

Also, add a section about validating the encoder offset via the d-axis voltage during operation.

I am also curious to see your section on converting position to speed.

[npetersen2]

Thanks for adding this image, but this is not what I envisioned.

Can you change this to a block diagram which is like what you would find in simulink? i.e., a signal flows from left to right and is manipulated? This block diagram should match your C code implementation, just in a graphical form.

[anirudhupadhyaya]

I am yet to update the block diagram, but I have addressed rest of the comments.

[anirudhupadhyaya]

Block diagram updated.

[elsevers]

Thanks @anirudhupadhyaya. I think this is on the right track. I have left you a few improvement requests below, and have gone ahead and implemented most of the needed changes as suggestions that you can simply accept.

In addition to this, I would like to see you add one (or more) reference papers to help users implement and think about the observer approach. There must be some good material from our digital control study group that Nathan and I did with you and Aravind that you can cite here?

You'll get big bonus points from me if you can include a MATLAB / Simulink example of a motor with an observer. I'd really love it if you can do this, but maybe too late in the game for this though.

[anirudhupadhyaya]

@elsevers I have addressed your review comments.

[elsevers]

Hmm, something seems to have gone wrong here.

We're missing the discussion on how the low pass filter always producing a lagging estimate. I remember suggesting changes to this --> how 10 Hz is a good starting bandwidth, that it will always produce a lagging estimate, etc.

I can't seem to find the text I had edited for this. Do you think you accidentally rejected my suggestion?

[anirudhupadhyaya]

Oh.. I guess it went away when I accepted the suggestions?

[anirudhupadhyaya]

@elsevers I added a line on the bandwidth recommendation for low pass filter.

[elsevers]

@mhmokhtarabadi here is the most recent build of this article:
Encoder Feedback - AMDC Platform.pdf

[elsevers]

AMDC v1.5 updated our encoder functions to refer to counts instead of pulses. See: Severson-Group/AMDC-Firmware#511

It is time to move forward now with adding this doc. We'll keep #147 and #140 going to update this doc with observer and PLL structures later (soon! hopefully).

[elsevers]

Let's move forward with this.