Hmmm, no responses? Not sure if that's a GOOD thing or a BAD thing...
I guess I'll just keep going until I'm told otherwise...
Here's one of the many 'thoughts' that I've been pondering over:
From what I understand, it's not uncommon to be throwing out stepper motor steps at 40kHz.
Admittedly, these are most often 1/16-step microsteps so the 'effective' signal being transmitted is a stepped sine wave at 40/16kHz = 2.5kHz.
Nevertheless, at motor currents of a couple amps, this is still quite a decent chunk.
All of the present electronics packages I've seen utilise 'built in' stepper drivers and route comparatively high current / frequency cabling to the individual motors. (4 wires to each motor on a bipolar).
I'm wondering if there would be any advantage to placing the current commutation right next to the motor itself and thus eliminating the losses involved in the parastic inductance / capacitance of the cabling.
Instead of 4 * high-ish current AC conductors to each motor we'd have 2 * high current DC conductors, 2 * LOW current 'signal' conductors (step and direction) and a single LOW current signal return conductor. (I'd assume the latter is needed to avoid ohmic losses on the negative of the high current supply line).
Since the machine I am building is a Prusa i3, there are two Z axis motors. I'm uncertain if a 'standard' i3 design attempts to drive both motors from a single A4988 styles controller, or uses 2 independent controllers. (If it's the former, then my concept would necessitate an addition controller, but hey, it's only $20 more!)
The next issue I'd imagine is thermal considerations. From what I understand the steppers get rather 'warm' under normal use, so I'd expect it would require some form of thermal separation. Perhaps as simple as a small block of PTFE?
For the Z motors, the drivers could be mounted to the frame itself with some thermal grease to act as a heatsink (although I suspect these driver require the least heatsinking of them all!)
I supect the most 'temperature critical' would be the Y axis motor given that it would have a heated bed sitting above it on a regular basis.
I guess I'll find out more once I see the _actual_ signals on an oscilloscope probing both the motor end and the driver end at the same time.
I also need to sit back and just THINK about it a lot deeper too. I'm sure there will be SOME cabling losses, but are they of any significance? This concept would only work to counteract AC losses as the resistive DC losses in the cabling would still be present.
