I didn't see anything like this posted around, but I 'm kind of surprised it wasn't. There's quite a lot of commercial activity nowadays on the topic of using motors for axis control in applications like CNC and 3D printing (additive manufacturing) applications. These usually start with TB6600-style stepper motor drivers for the smaller, lower current steppers, and at the higher currents these appear to have frequently evolved into applications for Piccolo-style DSPs usually driving discrete power MOSFET bridges; there are even models which accept simple quadrature inputs in a "quasi-closed-loop" fashion (just to make certain each step actually was executed and didn't get overwhelmed by a requirement for excessive torque which typically results in errors that are a multiple of 4 steps which can then be re-executed). Now at the very next level there are controls which are supposed to be for brushed DC motors, now some of them CLAIM to be "closed-loop" systems but frankly most of them don't actually have an operational "position mode" because we knew forty years ago that to do a decent position mode you really had to have analog feedback like you'd get from a sin-cos incremental encoder (like for example the Heidenhain ERN 1387 which has been designed into some elevator control systems) but these crude devices use garden-variety digital-only encoders which for this application are complete rubbish. I have no idea why such an obvious scheme seems to have eluded that market so far but what I believe most designers would really like to see is a controller that could be used with BLDC motors. Unfortunately with BLDCs there is so much emphasis on low-cost "sensorless" velocity-only operation (and that's fine and has its place) that when an accurate position sensor IS available there's just no way to integrate it into the system, whether or not such an encoder is "absolute" (that would increase cost markedly and in my opinion unnecessarily) and would therefore be used to somehow assist commutation. There are also a wide variety of simple chip-level BLDC controls that are also velocity-only, and essentially ALL of them require the ability to move the armature "randomly" at startup, which in a precision positoin application makes these idiosyncrasies intolerable! Could someone at least provide some clue as to what a block diagram of what a BLDC FOC controller modified for precision position appli9cation would look like, or is there an alternate to FOC for this type of application that doesn't drive up the cost too much, and some description of how it might work or the algebra behind it?
↧