Added TI angle observer code, and related parameters

[mjc-506]

As per https://openinverter.org/forum/viewtopic.php?p=26645#p26645

The new parameters/values, and the old atan2 code probably won't be needed long-term, but I've left them in for testing at the moment.

[jsphuebner]

Very nice, did you test it with some hardware? Could it also calculate angular speed to replace the clumsy 10ms angle difference calculation?

[mjc-506]

Yup, tested on hardware and it runs. Yes, the observer code can output speed too, I'll add that in (separate PR?)

…

________________________________ From: Johannes Huebner ***@***.***> Sent: Monday, 19 April 2021 20:35 To: jsphuebner/stm32-sine Cc: Matt Collins; Author Subject: Re: [jsphuebner/stm32-sine] Added TI angle observer code, and related parameters (#9) Very nice, did you test it with some hardware? Could it also calculate angular speed to replace the clumsy 10ms angle difference calculation? — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.

[jsphuebner]

Yes, that would be awesome!

[mjc-506]

Now uses speed from the observer code for resolver or encoder ICs, but retains the current speed calculation for other position feedback types.
Seems to give nice output, and handles quick speed changes as well as I can test with this hardware here.

[mjc-506]

Turns out 'lastFrequency' must remain positive - the motor will run fine with it negative but trying to regen results in increased speeds! (The TI observer code outputs speed with a sign - +ve for forward, -ve for backward, and the OpenInverter code 'corrects' that later by multiplying by the sign of the direction. This is needed for other position feedback, so I've left this there, and just 'ABS()'d the lastFrequency in the new code)