We ran some tests and found a problem with input of 12 vdc @ 1.5 amps pulsed mechanical. It was producing an output in the same line of 200 vac @ 2.5 amps. We looked with a scope and have found this loop to have a big problem as the mechanical points work fine. All other tests have melted the parts.

This is the diagram:
battery +
diode
points (3000 rpm x 6 pulses)
voltage mesured here on oscilloscope
electromagnet
voltage measured here on oscilloscope
on - off switch
battery -

What we saw today was a 2700% increase in voltage.

 

This is normal behaviour for an inductor (which the electromagnet is). As the magnetic field collapses in the inductor, a high voltage is generated. The contacts ("points") have melted due to arcing at high voltage. The usual solution is to put a reverse biased diode or a snubber circuit in parallel with the inductor.

 

And a solid state relay product will last much longer than a mechanical device at this cycle rate.