Hallo Ricky,

you should ask the manufacturer of the motors about adequate motor filters. If he is delivering the european market he must fullfill CE requirements, so he should have an 'idea' about what snubbers are needed.

Keep in mind, that all snubbers must be located directly at the source of interference (switch noise). This means, that not only 3phase lines must be filtered directly at motor, but also this certain 110Vdc motor switch. Otherwise, connections between motor and your board will act as an antenna, highly amplifying and emitting switch noise.

You told that your grinder motor is switched-on by applying 110Vdc to special switch coil. Don't forget, that this coil is also an inductive load, which, when not being nubbered properly will produce by induction extreme high overvoltages according to Uind = -L x dI / dt, everytime this coil is switched-off. Snubber circuit, mainly consisting of capacitance allows the current furtherly to run through the coil while decaying of magnetic field, when the coil is switched-off and by this heaviliy reducing dI / dt and accordingly Uind.
It's of extreme importance, that the snubber is located directly at the coil!! Otherwise due to high inductivity of wires heavy resonances can be excited, resulting in amplification and radiation!

In another post you stated to have full isolation by using DC-DC converter. Keep in mind, that these converters often suffer from rather high stray capacitance between input and output. So, for high frequency interference there's no such 'galvanic' isolation, but just the opposite! And if you have no properly choosen ground reference high freqency interference goes through your converter like a hot soldering iron through butter...

The influence of stray capacitance and resonances caused by inductivity of long wires is quite often overlooked. And although people are doing 'everything' to achieve 'full' galvanic isolation, they do just the opposite in terms of suppressing high frequency interference. And finally they are surprised, when the application goes wrong.

By the way, this is a also a good example, why PCB should have a solid groundplane: Electromagnetic radiation caused by coil switch noise will only then have no influence on PCB, when PCB contains solid groundplane. Why? Because only then loop area of all involved copper traces is minimum.

Kai