Now suppose the lamp left for a few seconds unattended, while the carbons are burning away and the arc is lengthening; in a very few moments the resistance will have increased, owing to the greater distance between the carbons, and we will suppose it to have become 5 ohms instead of 4.5.

The current pa.s.sing will now be 45 / 5 = 9 amperes only.

In other words, a very slight lengthening of the arc has reduced the current, and therefore the light, by 10 per cent.

Not only so, but 45 volts being needed to maintain an arc of {48} normal length, it is insufficient to maintain a longer one, and in practice the effect of leaving an arc under these conditions to itself for even a few seconds is that it _goes out_, to the annoyance of the lecturer and the confusion of the operator.

It is just _possible_ to work an arc lamp with a total E.M.F. of 45 volts by giving one's whole attention to it and never taking the hand off the feeding handle; but in practice no one with any experience would attempt it. The arc would almost certainly go out several times during the exhibition.

Now, take an example of a similar arc lamp consuming 10 amperes but worked from a supply of 200 volts.

Our equation C = E / R must then obviously become

C (10 amperes) = E (200 volts) / Total Resistance (20 ohms).

The resistance of the arc itself being the same as before, viz. 4.5 ohms, it is obviously necessary to put an _extra_ fixed resistance equal to 15.5 ohms in series with it in order to make up the total of 20 ohms.

_Now_ leave the arc unattended until the resistance of 4.5 ohms has again become 5 ohms; the only effect is that our current, instead of remaining at 10 amperes, has become 200 / 20.5 or 9.8 nearly, a difference which is imperceptible.

This is not all, for it is an elementary rule in electrical science that the total E.M.F. of any circuit distributes itself along that circuit in proportion to the distribution of resistance.

In other words, our original E.M.F. of 200 volts will so distribute itself as to reserve, so to speak, an E.M.F. of 45 volts for the arc, while the resistance of this remains at 4.5 ohms, but directly this resistance increases, the E.M.F. at the arc lamp terminals automatically rises, and therefore the actual diminution in current is even less than the figures above quoted. {49}

Should the arc tend to 'break' or go out, the resistance across it automatically becomes infinite and the _whole_ 200 volts is at that moment available to prevent the occurrence.

Under these conditions, therefore, the operator can safely leave the arc for many minutes at a time. In carrying out experimental work I have often left the lantern, walked up to the screen, discussed results with a friend, and walked back, and the arc has shown no signs of misbehaviour whatever.

[Ill.u.s.tration: Fig. 28.--Resistance.]

In practice any current from 100 volts to 250 volts may be considered as satisfactory for lantern work with a suitable resistance. Less than this involves feeding the arc rather frequently, and more may give a nasty shock, should the operator inadvertently touch a live wire, though I have worked an arc lamp on a current of as much as 500 volts.

The _resistance_ usually consists of a suitable length of wire of high resistance (Iron, German Silver, or those alloys known as Platinoid, Eureka, Manganin, Beacon, &c., are most commonly used) wound in spirals on a frame, and is generally supplied adjustable (Fig. 28), so that more or less current may be used as desired. These resistances get pretty hot in use, and care must be taken that they are placed where they cannot scorch woodwork, &c., and in cases where the lantern is a fixture it is a good plan to have the resistance bolted up against a wall once and for all. The resistance may be placed anywhere in the circuit, so long as the current pa.s.ses through it, then through the arc lamp (or _vice versa_), and back to the other {50} pole of the supply main; it does not matter in the least whereabouts it comes.

In cases, however, where one pole of the supply main is _earthed_, it is a good thing to place the resistance in the 'live' side, as this keeps the arc lamp within 45 volts of earth potential while it is working, to the comfort of the operator should he touch a terminal or wire, though with an ordinary lighting main there is no real fear of a dangerous shock in any case.

The _amount of current required_ depends of course on the size of the sheet, length of the hall, and density or otherwise of the slides; but it is usually accepted in practice that the efficient light from a continuous current arc lamp equals 100 candles per ampere, and therefore a 10-ampere arc will give 1000 candles. This is sufficient for all ordinary halls and slides, but where these latter are very dense, as for example with the Lumiere three-colour process, as much as 20 or 25 amperes may be required.

In these cases some special precautions must be taken for keeping the slides cool, or the result may be disastrous, but this is a question that will be referred to in a later chapter. A current of 10 amperes is pretty safe for all ordinary slides, and may be taken as the normal current used in large halls, though in arranging for the wiring it is as well to stipulate for at least 12 or even 15 amperes, especially as there must necessarily be a momentary increase of current at the instant the arc is 'struck.'

VARIETIES OF HAND-FED ARC LAMPS.--The pattern of hand-fed arc lamp ill.u.s.trated in Fig. 26 is only typical of many of the same general design, and there are others in which the design itself is fundamentally different.

Of these the 'Scissors' arc lamp made by several firms deserves mention on account of its simplicity and cheapness. As its name implies, the mechanism resembles a pair of scissors, the carbons being attached to the ends of a pair of levers hinged together {51} (Fig. 29). In this lamp centring movements are usually dispensed with, the arc being clamped on to a tray pin as in the case of a limelight jet. This is not, of course, so convenient, and a further disadvantage of this pattern arc lamp is that the feeding process gradually alters the position and angle of the carbons. In fact, the one great merit of the lamp is cheapness, and where expense is an object, it should certainly be considered.

[Ill.u.s.tration: FIG. 29.--'Scissors' Arc Lamp.]

Yet another arc lamp deserving of mention is the 'Parallel,' a name again very aptly chosen, as the two carbons are either exactly parallel to each other or very slightly inclined. In the former case the arc has to be 'struck' by touching the ends of the carbon rods with a piece of metal or carbon. Of the actual manipulation of this lamp I have had very little practical experience, but I have heard it well spoken of, though I believe it has so far only been made for currents of 5 amperes or so.

Yet another type which must not be ignored is the 'Right-angled' pattern (Fig. 30), a name again self-descriptive. The horizontal carbon is the positive, and the vertical the {52} negative, and this lamp again is made by several manufacturers in slightly different forms.

This pattern lamp is in my experience the best of all for _small_ currents, say, of 5 amperes or so, but inferior to Fig. 26 for currents of 10 amperes or more. This last remark perhaps hardly applies to _alternating_ currents, which, however, I have not yet discussed. I cannot conclude this brief category of arc lamps without referring to the _enclosed_ pattern, of which the 'Westminster' is perhaps the best-known and most popular (Fig. 31).

[Ill.u.s.tration: FIG. 30.--'Right-angled' Arc Lamp.]

This is a lamp of the right-angled type, but the arc burns in a cylindrical gla.s.s chamber, not air-tight, but partially so. After burning a few minutes the oxygen in this chamber becomes used up and its place is taken by carbonic-acid gas and other products of combustion, after which the carbons burn away very much more slowly, and therefore require feeding at much greater intervals.

This lamp again is chiefly made for small currents not exceeding 5 amperes (and can therefore be used from any ordinary lamp socket), and for a moderate-sized hall is on the {53} whole as cheap, efficient and simple a lamp as any I am acquainted with. It can be supplied with or without mechanical centring movements as required, and is usually sent out with its own resistance for the particular current on which it is to be used, so that it only requires connecting up to the nearest lamp socket, and is ready for use.

[Ill.u.s.tration: FIG. 31.--'Westminster' Arc Lamp.]

It is _not_ sufficient for anything larger than a 12-foot sheet or for working at a greater distance than, say, 40 feet, but within these limits the lamp, and in fact _any_ good 5-ampere arc lamp, will be found quite satisfactory and saves the expense of putting in a special cable.

AUTOMATIC ARC LAMPS.--Arc lamps for lantern work in which the feeding is done automatically are also made. Like hand-fed lamps, they vary in exact design, but all, or practically all, are so designed that the carbons are brought together by means of springs or weights, and some form of 'brake'

controlled by a system of electro-magnets checks the {54} movement. As the carbons burn away the arc lengthens, the current weakens, the electro-magnets lose their grip, and the carbons move together until the increasing current puts on the brake again. Some of these lamps are 'semi-automatic' only, that is to say, the arc has to be struck by hand, while others perform this operation automatically as well, usually by an additional magnet which draws back the carbons by the correct amount after the arc is struck.

My frank advice to intending lanternists is to leave these lamps alone.

Some of them are satisfactory up to a point, but they are all apt to be 'jumpy,' and on the whole the hand-fed type is in my opinion to be preferred.

ARC LAMPS ON ALTERNATING CURRENTS.--The alternating current is not so good as the continuous for lantern work with arc lamps: the light per ampere is not so great, the light has an irritating habit of travelling round the carbons and there is always a slight 'hum.'

The sum total of these drawbacks is nothing very serious, provided that proper arrangements are adopted, and I have frequently manipulated arc lamps on alternating circuits with such good results that professional lecturers have at first refused to believe that the circuit really _was_ alternating.

As it is frequently stated that to obtain a steady light with an alternating current is impossible, I can understand their surprise, and I can also understand the statement in question, as the problem is usually tackled on entirely wrong lines.

It is almost always stated that arc lamps for alternating currents should be arranged with the carbons _vertical_, and many makers actually so construct their lamps as to allow of this.

To obtain a steady light under these conditions _is_ impossible and I pity anyone who attempts it; but the statement that this is the best method of working has been repeated so often that it seems to have been taken for granted.

The best arrangement (in my hands at any rate) is to {55} slant the carbons as for the continuous current, and also to have the upper carbon cored and the lower one solid, but to use a rather larger lower carbon than would be correct if the main were continuous.

Also the upper carbon should not be _quite_ so far back as with D.C.; to have the front edges of the two carbons practically in line is about correct, but the _exact_ position should be carefully adjusted to obtain the steadiest light, and it will be found that a slight alteration makes a considerable difference.

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 pa.s.sing 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 {56} 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.

[Ill.u.s.tration: FIG. 32.--Arc Lamp with Induction Ring.]

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.

[Ill.u.s.tration: FIG. 33.--THE OPTICAL SYSTEM OF A LANTERN.]

{57}