toggle quoted messageShow quoted text
I do just as you suggest on all my dials and masters. 1N400X series diodes. The maintaining of current is minimal (looked some years ago with an oscilloscope and it was insignificant). There is typically some variation of impulse length dependant on the master anyway (Gents tend to be longer and early Synchronome clocks rather short - lighter parts?). Typically, the coil carries about 300 mA and has a resistance of about 5 Ohms, with a parallel resistor of about 50 Ohms.
Diodes work well - but of course you have to have the right polarity.
Surely now we have cheap reliable silicon diodes the best approach is the now-standard practice of connecting a diode across the coil so that when the contact opens the diode shorts out the inductive kick? The diode should be reverse biased when the coil is energised. The only snag with this is that the diode maintains the current when the contact opens for a period which may cause a problem. That can be made shorter by putting a resistor in series with the diode to dissipate the stored energy more quickly at the expense of a higher voltage. For example, if the coil carried 1 amp when the contact was closed, and the resistor was 100 ohms, the peak backswing voltage would be 100 volts, quite a lot but probably not enough to cause breakdown in the gap.