Power supply design for a 230VDC 1.75HP motor controller (DC-DC conversion Q's)

[Sci]

Breif
I'm trying to create a microcontroller-driven motor controller for machine-tool use, but have little experience with the mains end of things.

Specifics
I have a nice big mill, but in my current workshop there isn't space to use it with the stock motor. It makes it too deep. I have fabricated a mount and pulley adaptor and affixed a treadmill motor instead. It fits and ran until the original controller failed catastrophically some months ago.

The original treadmill controller design used MOSFET PWM of rectified UK mains straight to the motor (the PWM section is sheilded and potted. 3 days in acetone did not soften it to allow dissection. All I know is it ran at sufficient power to blow traces off the control board PCB). It's failure mode was to go from what I think was an upper limit 70% duty-cycle (for equivelant 230VDC) to jamming the MOSFET continously on and sending sparks shooting out of the motor housing (~300VDC). I would rather have a design that does not require me having to skim the commutator bars on the lathe again if it fouls up.

On the motor control end I'm looking at an Arduino in 16-bit PWM mode connected via an opto-isolator with schmitt-trigger output to a MOSFET. There'll be an RPM sensor on the spindle for initial PWM>RPM ratio calibration and later stall feedback.
The final goal is to be able to dial in a material cutting speed and cutter diameter and have it automatically spin up to the ideal RPM for the material and cutter in use. Use of PWM and a DC motor should also give me a much wider speed range (as it did before the original controller blew) and good torque. And if the power supply is reasonably adjustable it could likely be applied to a lot of generic Chinese import machine-tools that use DC motors as a nice functional upgrade.

(indeed I would like to try it on my lathe with a position readout hooked in, so it can dynamically adjust speed to match current material diameter)

The problem
As hinted at above, I'm aware the peak voltage of rectified UK mains is going to be a spot over 300v. The advice I've received on that so far elsewhere has been what I feel is more toward the "bodge" end of the design scale. Monolithic 2Kw iron transformers, rectified triac lighting dimmers, etc.. I'm hoping folks here might be able to give more polished suggestions, or at least tell me if and where my ideas are actually unrealistic.

My current line of thought is a buck-converter working off rectified mains, or possibly a buck-boost to enable a wider range of motor ratings to use the same basic circuit/PCB. I've been told a DC-DC converter at these power levels is unrealistic, but I've found an example of an even more high powered one here: http://www.stevehv.4hv.org/boost_converter.htm (massive overkill warning)
I've not built a buck or boost circuit before, but from what I understand so far they're all very similar in terms of componants used, so it's more a planning matter than a material cost difference one.
I suspect that using a buck-boost would enable the same circuit to be used with a wide range of AC voltages worldwide, with only a couple of components requiring customisation based on specific motor power. Good for hackers and makers everywhere.

The questions

• I suspect a pot-set voltage buck-boost would work, but would probably be inefficient if the inductor and cap aren't sized down to match the specific motor it's all attatched to. Yes/no?
• IGBT vs MOSFET. While I can't quote specific sources, it sounds like MOSFET on motor control and IGBT for buck-boost would be most relaible?
• But then it also sounds like IGBTs failure-mode is short-circuit, so in such a situation would risk over-volting the motor? Can buck-boost be made to fail-off?
• I think having a current monitor on the motor control stage would be good for catching mechanical jams, motor over-voltage arcing and other problems, but a software shutdown may be too slow? Would a large polyfuse be better?
• The motor I have is rated for 1.75HP, 2HP peak. So peak of about 1500Watts. A 2Kw PSU sounds like a safe overkill margin to me, or should I aim closer for efficient power use?

My limitations
I have more time than money, so I'd rather plan it out right than burn through components trying things out.

Also, Hi! I'm new here. Any words of wisdom are welcome.