nr 
243 
derive for copper with 7'=3 resp. 8 sec., Hin — 250 resp. 150. 
Taking however the horizontal pieces of the wire into account, 
the resistance increases */, times, the moment of inertia ’/, times, 
whence: 
Hin = 330 resp. 200 
m—6 ay 10 
Not to make the coil too thin and mechanically too weak, I’ll not 
take in accordance with (7) K/K,=m, but = 3. Then according 
to (9) the efficiency will still be 78 resp. 80°/, whereas it is 83 
and 90°/, in the most favourable case. 
When we choose for the faster galvanometer a mirror of 8 mm., for 
_ the other a mirror of 10 mm. we get: 
K,= 09,0005 resp. 0,0013 
and 
D=0,0070 „ 0,0025 
These are about the utmost values which we can use, so that 
smaller mirrors would hardly produce a greater sensitivity. We 
have namely up tll now neglected the airdamping in our calcula- 
tions. This appears to be already quite perceptible here. By further 
reducing the product AD which is 10.10-® here, we should cause 
the galvanometer to be already aperiodically damped on open circuit. 
By using a smaller coil one could get somewhat further. 
To complete the calculation the resistance must be known. As 
an example |! choose 100 2 for the total resistance. From (6a) we 
derive: 
—= 12 resp. 33 and r,=—8,3 resp. 3,0 2 
and / from 
aot 700 — 2,6. 10-5 
jf =13 in both cases. 
As one winding has an area of 1,2 cm? we must take 11 windings 
with a length of wire of 51 cm. From the resistance and the length 
we find for the diameter of the wire 0.035 resp. 0.06 mm. The 
first of these two values is rather too small. We can remedy this 
1st. by taking a smaller coil of 5 X 12 mm. e.g. thus increasing the 
number of turns and the length of the wire. The diameter then 
becomes 0.043 mm. 2ed. By not taking £=3 but eg. k=2. Then 
we find K, = 0.0011, r, = 5,6 2 f=—15.3 diameter 0.046. A com- 
bination of both methods gives the appropriate thickness of 0.055 mm. 
From (4) we derive the voltage sensitivity. For the deflection P’ 
