Review Request for Arc Volt Sensor Schematic

Yes, inverting and decoupling. 
Wouldn't be surprised if there is a similar isolated SPI component.
Best if the ADC has some kind of free run mode. Otherwise the reading of a value can take too much time.
Was looking for a new value every millisecond, with as many bits of resolution as possible.
Have also used ADS1014 with similar good result.

I have these:
www.ti.com/product/LM331
but i never bothered to use them as the time is worth more than the price of THCAD.
Easy to use and can be plugged to parallel port with a single optocoupler.

I moved the voltage divider to the first stage and put a zener after it for overvoltage protection like Hakan.

@tommylight: Do you have a schematic you can share that shows how you use 3 resistors for voltage division. I'm having trouble visualizing your description.

Here's where I'm at right now. I need to add SPI isolation. That's going to be the biggest stage of all because there aren't any good through-hole all in one solutions like there are for SMD.

I also wonder about the 2 50k resistors.

Good find on the regulated TMR-0511.
I looked at the output of the isolated power supply I use. I wasn't terrible, almost ok but not totally stable.

Why the filter step? HF?

I use the filter for a couple of reasons:
1. environmental noise
2. I don't trust cheap plasma cutters to output clean DC

I see a couple of capacitors and resistors as cheap insurance against potential future problems.

I want to be able to confidently support cheaper options like Lotos LTP5500CNC. That's the one I have. It works fine and is a small fraction the price of a Hypertherm.

I actually looked at supporting HF start machines in the past. I just couldn't justify the cost and effort.

edit:
The values are placeholders, i usually use 2x470K to 2x1M for the side resistors and adjust the value for the middle one depending on what i can find.