Just a comment or two to add to the good discussion on transient
suppression.

Transient suppression is simply the controlled dissipation of the energy
stored in the magnetic system.  (Note that the mechanical part of the
system also can store energy.  When the armature returns through the coil
it generates a voltage in the coil.)  Subtracting the heat losses in the
coil and mechanical parts of the system, the remaining energy must go
somewhere when the driving current is removed.

Simplifying things a bit, the energy dissipated can be described as the
product of voltage, current and time.  So if we open the circuit quickly,
given the usual low resistance of the component parts, the voltage must
rise towards infinity as the time approaches zero.  Voltage is the usual
culprit in damaging components.

Now the tradeoff begins.  If the system is to recover quickly, the
voltage or current must increase.  The trick is to find the optimum
balance of recovery time versus voltage, given the inductance and
resistance of the system.  (For the more technically minded, the L/R time
constant is usually the limiting factor.)  Note that the system energy
does not have to reach zero for the mechanical part to return to its
initial position.

Robbie's comment about analyzing the waveform is right on.  There are
too many variables in an electromechanical system to predict well which
suppression system will work best.  Based on my experience, I would look
at a snubber network first, possibly including a diode.  Usually a
snubber consisting of a series resistor and capacitor will do the job.
The resulting waveform is a damped sine wave, which will cause much less
damaging arcing across contacts.

Bob Billings
toadhall@sprynet.com