Integrate Zmode settings and acceleration into motor control algorithm

[jacobbenlewis]

To implement Zmode I added acceleration percentage to the packet, and created max velocity and acceleration settings on the Rxr. These still need to be rolled into the motor control algorithm so that they do something.

@squarewave, @thayerjbirch, are either of you guys able to do this or should I plan on doing it?

[thayerjbirch]

I can work on it. Should be simple enough.

[jacobbenlewis]

Is this done then?

[jacobbenlewis]

I didn't see accel changing much if at all in my testing of this, but it could be a settings issue. Did you guys see it working?

Also, Zmode is currently making use of the standard FreeMode velocity/acceleration settings. I created new settings specifically for Zmode that need to be used instead. David wanted to be able to set them independently.

[squarewave]

Hmm I just merged things in and made sure nothing broke. I dug into it a little bit this morning though and I seem to always be getting 100's from receiver.cpp no matter what. I couldn't spend more time on it than that, but I'll look into it a bit more tonight.

[davidsthayer]

So pleased that we got through FCC today!!! This will really be so worth it when we get all the rest of it dialed. Thanks guys. Really, really great work everyone!!

[squarewave]

Awesome! So I did some more testing and I'm definitely able to see the effects of the acceleration changes in z-mode. It may just be that with high enough acceleration values the difference between 25% and 100% is not noticeable. I could definitely see it when I set my accel to about 10.

[jacobbenlewis]

I did some testing on this tonight. I too see a difference when pressing accel buttons in zmode if the accel setting is low enough. I think there must be no practical difference in accel when you get up above maybe 32?

Anyway, I was also seeing that the normal Acceleration setting was affecting the Z_mode Acceleration setting. I had Z_mode down at 16, and then a going from normal Accel setting of 32 down to 4 makes a noticeable difference in zmode behavior. In fact, at 32 zmode becomes quite unstable. I think I traced it to decel. In MotorController::set_accel(), decel_denominator_ is calculated using max_accel no matter what mode you are in. Also in MotorController::Configure(), max_decel_denominator_ is always based on max_accel no matter the mode.

Can @thayerjbirch or @squarewave take a look and confirm or deny my suspicions?

[jacobbenlewis]

@thayerjbirch or @squarewave, are either of you guys looking into this? If not I may venture into that code myself and attempt to fix the problem.

[davidsthayer]

Thanks @jacobbenlewis for looking into this! This is a key issue.

[jacobbenlewis]

I spent some time trying to figure this out but after making the change to fix deceleration the problem persists. I just changed the line in MotorController::set_accel() when setting decel_denominator_ for z_mode it uses z_mode_accel_ instead of max_accel_. That does make it so that now in z_mode the decel matches the accel.

But there is still an issue when you set z_mode accel low enough (4 will definitely do it), then push the button to set accel to 25%, the motor will often overshoot, or maybe move a little one direction before starting the real move in the opposite direction.

Fixing this I think gets into the bowls of the motor control algorithm so it would make more sense for @thayerjbirch or @squarewave to look into this if you have time. I think it would take a while for me to just get up to speed on what the algorithm is doing.

@davidsthayer, another possible option to fix this quick though is to remove the different settings between playback mode and z mode. I'm only guessing here though because playback mode doesn't seem to have a problem going down to an accel of 1.

[davidsthayer]

Always a little messy getting into the bowels! Okay, this is gonna be all hands on deck. It's really important to get a handle on this before we ship. I'd briefer to get this wrung out in z-mode rather than playback if we can.

-David

David Thayer
Motion Dogs
805.788.0280

On Dec 29, 2014, at 8:30 PM, jacobbenlewis notifications@github.com wrote:

I spent some time trying to figure this out but after making the change to fix deceleration the problem persists. I just changed the line in MotorController::set_accel() when setting decel_denominator_ for z_mode it uses z_mode_accel_ instead of max_accel_. That does make it so that now in z_mode the decel matches the accel.

But there is still an issue when you set z_mode accel low enough (4 will definitely do it), then push the button to set accel to 25%, the motor will often overshoot, or maybe move a little one direction before starting the real move in the opposite direction.

Fixing this I think gets into the bowls of the motor control algorithm so it would make more sense for @thayerjbirch or @squarewave to look into this if you have time. I think it would take a while for me to just get up to speed on what the algorithm is doing.

@davidsthayer, another possible option to fix this quick though is to remove the different settings between playback mode and z mode. I'm only guessing here though because playback mode doesn't seem to have a problem going down to an accel of 1.

—
Reply to this email directly or view it on GitHub.

[jacobbenlewis]

Did this ever get fixed?