MDL-BDC24 Comutator Mode

[drakesword]

Just digging through the source code for version 7243 and stumbled apon this curiosity.

           //
            // Set the number of brushes in the motor, which determines the
            // number of commutations per revolution for sensor-less speed
            // detection.
            //
            case COMMAND_NUM_BRUSHES:
            {
                //
                // This command has been handled.
                //
                break;
            }

Now the comments state that the command was handled although it is not. Anyone have any more insight into this?

[Al Skierkiewicz]

Bryant,
Doesn't the response return a status that the brush number has been set and nothing else? Since I am not a programmer, this seems flawed in that all of our motors have two brushes but the number of commutations per revolution varies from motor to motor. Some only have three commutator segments, but others have mush more. I wonder what the other part of the code uses to make the calculation.

[Mark McLeod]

That may be just a stub for future capability or something they were thinking of playing with.

[drakesword]

The comments read "for sensorless speed detection" I'm getting antsy!

Haven't actually mapped out the hardware of the jag but how would it be possible to do this? Maybe detect spikes and brakes in current?

[Ether]

Originally Posted by drakesword

The comments read "for sensorless speed detection" I'm getting antsy!

Haven't actually mapped out the hardware of the jag but how would it be possible to do this?

http://ieeexplore.ieee.org/xpl/freea...number=4510598

http://ieeexplore.ieee.org/xpl/freea...number=5951768

http://www.ideadvance.com/sites/idea...hitePaper2.pdf

[drakesword]

Originally Posted by Ether

http://ieeexplore.ieee.org/xpl/freea...number=4510598

http://ieeexplore.ieee.org/xpl/freea...number=5951768

http://www.ideadvance.com/sites/idea...hitePaper2.pdf

Very Nice ...
How do we implement this ^_^

[Ether]

Originally Posted by drakesword

Very Nice ...
How do we implement this ^_^

Use a scope to look at the motor's voltage and current under various operating conditions and see if there's a commutator transition signature of sufficient S/N that you could detect it reliably with some custom hardware.

Or better yet, wait till TI adds that functionality to the Jag, or some third party markets a chip to do it.

[drakesword]

Originally Posted by Ether

Use a scope to look at the motor's voltage and current under various operating conditions and see if there's a commutator transition signature of sufficient S/N that you could detect it reliably with some custom hardware.

Or better yet, wait till TI adds that functionality to the Jag, or some third party markets a chip to do it.

Yea I meant as in how do we implement this with the jags current hardware ...

I have yet to dig into the full hardware availability on the jag ... guess Ill be doing that next

[Ether]

Originally Posted by drakesword

Yea I meant as in how do we implement this with the jags current hardware

AFAIK, the current Jag does not have built-in support for this.

[drakesword]

Originally Posted by Ether

AFAIK, the current Jag does not have built-in support for this.

Well there's hardware built in for current/voltage monitoring could it be bitbanged into our purposes?

[Ether]

Originally Posted by drakesword

Well there's hardware built in for current/voltage monitoring could it be bitbanged into our purposes?

Could be wrong, but my guess is no. Sample rate too slow to reliably capture and detect all commutation transitions.

As I said previously, use a scope to look at the the commutation transition signature to see what would be required to reliably detect it.