Tom:

I've tried everything you've said, all reasonable combinations and permutations. There has got to be some missing piece of information I don't have. Last time it was the ch->CommutationOffset that kept Write3PH() from working. One little detail like that is all it takes for a days' effort to become totally futile.

I have uploaded the .c file (to the Easy E folder) that I've been using to horse around with this motor. The goal is to get this to work with the Halls, eventually. Right now, though, I just want the BLDC to work with the servo and the 256 LINE quadrature encoder.

To clarify, what I do towards the end of this program is to execute a Write3PH() to the motor, and hold it there for 3 seconds. I then Zero the axis, and set the commutation offset to 1/4 cycle away. I then enable the axis and try to jog - zero net movement, but the motor has got current and you can tell the servo is struggling mightily as it fights itself. I have "cleared" the filters by assigning all zero coefficients as you can see (I'm not sure what else "clearing" the filters means).

-Erik-

--- In DynoMotion@yahoogroups.com, Tom Kerekes <tk@...> wrote:
>
> Yes do what I requested in the last two emails :}
>  
> TK
>
> From: quelisto <quelisto@...>
> To: DynoMotion@yahoogroups.com
> Sent: Tuesday, February 21, 2012 12:37 PM
> Subject: [DynoMotion] Hmmm, a 256 count per rev diff encoder is giving 1024 counts per rev
>
>
>  
> Hi Tom, I think maybe my troubles with the BLDC have something to do with the afforementioned curiousity. I checked the documentation on my encoder, and sure enough, it is a more standard 256 count per rev encoder (not a 512 CPR). I am using the Kanalog Diff 0 input for the interface, using ch#->InputChan0= Encoder0_Channel.
>
> I wrote a little program to watch the position as I turn the shaft by hand, and sure enough, I repeatably get +1024 counts going one way, -1024 going the other.
>
> I have tried a bunch of permutation/combinations for both the inv_Dist adn the commutator offset (1/256, 1/512, 1/128; 32, 64, 128, 256). None work.
>
> So...any advice?
>