It is also a great help to have a weak electro-magnet, or its equivalent, so arranged that it tends by its influence to keep the arc to the front. On some lamps this is provided for, as even with a continuous current it is quite harmless and, if anything, beneficial; but, if not, any competent mechanic can easily fit an 'Induction Ring,' consisting of a single turn of stout copper wire, which has sufficient magnetic influence to do all that is required (Fig. 32).
This ring must be wired in series with the arc itself, and as the current passing in it automatically reverses in synchronism with the arc, its effect is always to deflect the arc in the same direction, and care must of course be taken that it is so wired that the deflection is forward and not backward. This is the exact arrangement I have myself adopted, and I never experience any difficulty on the score of the arc wandering.
Right-angled arc lamps, as described on pages [52] and [53], are also very efficient on A.C. mains, and frequently these lamps are already equipped with electro-magnets for the purpose required. The 'hum' of an alternating current cannot be altogether eliminated, but can be reduced to a minimum by reducing the voltage as far as possible.
As has been already said, the A.C. lends itself readily to transformation of voltage, and I find in practice 90-100 to
be ideal. More than this is inclined to be noisy, and less is apt to result in an unsteady arc.
The arrangement, therefore, which I recommend from long experience is to employ a transformer to reduce the E.M.F. to 100 volts or thereabouts, and then work with a resistance in the usual way (if the original current is 100 volts, of course no transformer is required) with a properly constructed arc lamp fitted with an induction ring or electro-magnet. No difficulty should then be experienced in obtaining a good, steady, and fairly quiet light.
Any little 'hum' remaining can be silenced to a very considerable extent by placing the entire lantern on a thick block of saddlers' felt, but in practice I have never found this necessary with ordinary currents, though a few abnormal circuits where the 'periodicity' is very high are noisier than others.
