(10) ERROR DUE TO REVERSAL OF CURRENT THROUGH THE

Connect up as shown in Fig. 14, but instead of connecting M directly in series as shown, join it up now to the circuit through a reversing switch or key, so that the current through it may be reversed in direction though that in the rest of the circuit and therefore through M, is still unidirectional.

With R large, close S, and obtain say full scale deflection on M, noting that on M, which must be constant; now reverse current in M and again note its value for the same one as before on M. Next re-reverse and note it again.

(11) Repeat this operation for about 5 scale readings on M up to the maximum at roughly equal intervals.

N.B. This test of course only applies to direct current instruments. Tabulate as follows

(12) ERROR IN VOLTMETERS (ONLY) DUE TO HEATING OF COILS BY PASSAGE OF CURRENT.

Close S, and adjust R to obtain about

(F.)

scale reading on M, note the corresponding reading on M, which must be an electrostatic voltmeter. Maintain M, constant for, say, quarter of an hour and again read M.

8

(13) Repeat 12 for a full scale reading on M, and tabulate as before.

N.B. The error in voltmeters due to change in the temperature of the room is readily calculable when the latter is obtained by a thermometer.

(14) Plot the following curves for tests

8

B. Having readings on M as ordinates, and M, as abscissæ for both ascending and descending curves.

C. Having readings on M as ordinates and frequency in per sec. as abscissæ.

D. Having readings on M as ordinates and frequency in per sec. as abscissæ.

Inferences. State very clearly and concisely what you can infer from the results of your observations.

(16) Calibration of a Wattmeter by Comparison with a Kelvin Composite Balance used as a Wattmeter.

Introduction. The following is a convenient method of calibrating a Wattmeter by means of direct currents, using a Kelvin composite balance as the standard Wattmeter with which to compare the one to be tested. The construction of the balance is detailed on p. 361, where the mode of using it as a Wattmeter is also given, and it will merely suffice to say here that it is used very similarly to the Hekto-ampere meter, the only difference being that as a Wattmeter, the fine wire movable coils (only) are placed in series with an extra antiinductive resistance across the mains supplying the power measured by both Wattmeters. It may here be noted that it is not necessary for the current through the thick winding and the pressure across the thin coils to be developed by one and the same source. For since the Wattmeter deflection is to the products of the currents flowing through the two coils, clearly these may come from two totally different sources. In fact it is distinctly preferable to have them separate when possible, for then the variations of the main current will not affect the constancy of the pressure on the fine coils.

This same test serves to determine the "constant" (K say) of the Wattmeter, or in other words the number by which the scale reading must be multiplied so as to obtain the power in Watts.

The following reasoning will no doubt render this clearer.

Assuming the general principle and construction of, suppose, a Siemens Wattmeter to be understood. Let C and c = the currents flowing through the fixed thick- and movable thin-wire coils respectively when a deflection of the torsion head and its pointer on the scale is D° or divisions. Then the force acting between the coils is Cx c, but (c) to the pressure V at the

terminals. Hence the deflecting couple acting between the coils ∞ C x V ∞ Watts. Now when the index is brought back to 0 again by turning the torsion bend, thereby twisting up the spring and introducing the control, we have-torsion of spring Watts CV; but the force of torsion is to angle of torsion of such a spring,

.. D∞ CV

or KD = CV Watts measured and causing a deflection D, where K is the "constant' " of the Wattmeter tested. It may be found that K is not perfectly constant throughout the whole scale. In this case the Watts should be obtained from a calibration curve rather than by the product KD.

Apparatus. Kelvin composite balance (K.B.), with its anti-inductive resistance r (p. 357); switch S; variable power-absorbing resistance R; accurate voltmeter V (preferably electrostatic); main current battery B; Wattmeter (W) to be calibrated ; [pressure battery b and adjustable resistance R1 if available]; adjusting rheostat R

Observations.—(1) Connect up either as in Fig. 15 or 16, and in the present test assume the latter for actual experiment, and make quite certain that the connections are as indicated in Fig. 16.

(2) Carefully level the instruments that require it, adjusting their pointers to zero, and if W has a suspended coil see that this is quite free to move.

Note. Care should be taken to run the "leading in" and "out" wires carrying the main current to W, and in the rest of the circuit close together or twisted in order that the currents flowing in them shall exert no magnetic influence on the instruments.

(3) Turn the switch on the balance to "Watts" so as to place the movable fine wire coils across the small terminals. Observe whether (r) is numbered the same as, and therefore belongs to the balance in use, and make quite certain that the correct resistance is being used in () (p. 362).

(4) Adjust the balance and its sensibility by using the proper weights as given in the table of constants (p. 362), so that the maximum Watts to be measured on W would give, as nearly as possible, a full scale reading on K.B.

(5) With R1 fairly large, close S1 and adjust the voltage as read off on V to the desired amount by altering R1, and then maintain this voltage constant, observing that it is so before taking every reading.

(6) R being fairly large, close S, and alter R so as to obtain about th of the maximum scale reading on W. Note simultaneously the reading on Wand position (d) of the slider of K.B.

(7) Repeat 6 for some ten different deflections on W (by varying R) rising by about equal increments to the maximum, the pressure remaining constant all the time.

(8) Repeat obs. 7 for a similar set of decreasing readings on W, and tabulate your results as follows—

(9) Plot a "calibration" curve for the Wattmeter tested having values of D as ordinates and true Watts as abscissæ.

Inferences. What can be inferred from the results of your Are Wattmeters subject to any sources of error, and if so,

test?

how can they be minimized or got rid of?

(17) Calibration of a Wattmeter by Comparison with a Standard Ammeter and Voltmeter.

Introduction. The following method of calibration by direct currents entails the use of an accurately calibrated standard ammeter and standard voltmeter. These may be either Kelvin balances or ordinary instruments which have recently been carefully compared with accurate standards, and a record of the calibration curves of which are obtainable.

It should be remembered that the constant of a Wattmeter obtained with direct currents will only be true for alternating currents providing the self-induction of the fine wire moving coil or its circuit is practically zero or very nearly so. In other words, the instrument must contain no iron and also be very nearly "non-inductive." This is a matter of great importance, for Wattmeters are in most cases only required to measure power in alternating current circuits.

Apparatus. Standard ammeter (A) and voltmeter (V); Wattmeter (W) to be calibrated with its anti-inductive resistance (r) if there is one; battery of secondary cells B; switch S; suitable resistance R for absorbing power; carbon rheostat (Rh) (p. 393).

Observations.-(1) Connect up as shown. Carefully level all the instruments, adjusting their pointers to zero, and see that