For the secondary wire No. 34 is best adapted. It is best also to wind the sections in "pies" of 1/8 in. thick. This size of wire can be wound after it has been through a bath of hot paraffin wax, on an average of about 912 turns to a pie of cross section 1/8x1 in. and by a cut-and-try method it is found that 70,000 turns can be placed in 60 pies having an average cross section of 1/8x1 3/8 in. The secondary winding should have about 1 in. for insulation between it and the insulating tube, to prevent the spark from jumping to and across the surface of this tube. At the ends of the coil where normally the potential is highest this insulation should be increased. Thus the insulation between primary and secondary would resemble a series of steps from the centre outward.

Applying these additional considerations to the calculations of the number and size of the secondary pies, the design will be as follows:

18 pies 4 1/2-in. int. diam. and 1 1/2 in. cross see'l length. 18 pies 4 3/8 in. int. diam. and 1 3/8 in. cross see'l length. 24 pies 4%-in. int. diam. and 11/4-in. cross sec'l length, giving 60 pies and approximately 74,000 turns.

This will give a uniform outside diameter of 6 in. which is in good proportion for a 12-in. coil with the chosen core dimensions.

It is proper to assume that the laboratory coil will sometimes be used with condensers across the secondary and that it will be desirable to choke back high frequency oscillations. For this purpose there will be added two coils at each end wound with No. 28 double-cotton-covered wire, to the same thickness, and external diameter as the other pies and 4 3/4 in. internal diameter. This will add less than 1000 turns; so that there will be approximately 75,000 turns on the secondary.

Even when wound with less than 60,000 turns on the secondary, the above coil would give a 12-in. spark but would require, when tested under the same conditions in the primary circuit, one-third more current than the present coil; while a coil wound with about 53,000 turns of the same size wire, cotton covered and thus more wasteful of space, would require one-half more current, using in each case a Wehnelt interrupter. Table I.-Rupturing E.M.F. in Kilo-Volts per Inch. Substance. Rupturing e.m.f..

Ebonite........................ 900-1,500

Glass ........................... 500

Window glass ................... 380-l.000

Gutta percha................... 250-1,000

Mica ............................ l,500-5.000

Micanite ........................ 2,500-7.500

Paraffin wax ................... 330- 650

Petroleum ...................... 230

Impregnated paper:- Thickness. .0047 1,000 Thin printer's paper.

.0035 1,300 Tisue paper.

.0071 1,100 Manilla paper.

.0051 1,600 American linen paper.

.0055 1,350 Linen typewriter paper.

Boiled-out linseed oil .................215

Light mineral lubricating oil .........120

Paraffin oil sp. gr. .28 ................165-250

With the same dimensions a coil for wireless telegraphy work would have required a primary with three complete layers of No. 10 wire and No. 32 for the secondary, wound more effectively since the smaller number of turns for an equal bulk of this wire would give a lower potential, which would require less space for insulation. The coil could probably no longer be rated at 12 in., but it would be far better adapted to heavy wireless work, because of the decreased resistance of a shorter length of larger secondary wire and the increased ability to handle a large volume of current. Since in this class of work an extremely rapid interrupter is not advantageous, the coil could have been improved by making the core of larger diameter and thus increasing the amount of iron, thereby having in common for the two coils only the length of the core.

In joining up the separate pies the insides and out-sides of consecutive coils should be connected together with the necessary thickness of insulation between them. When dealing in such large numbers and using factors of safety of 3 or 4 it is necessary to estimate only approximately, and slide rule calculations are wholly adequate for most of the preliminary work.

Assume that the impregnated paper between the pies has a thickness of .006 in. per sheet. This insulation must withstand, for the method of connection which has been chosen, double the potential generated by a single pie. or 6000 volts, roughly. From the accompanying table it is seen that this paper will rupture at 1200 volts per 1/1,000 inch. Using a factor of safety of 4 there will be obtained 1800 volts per sheet, which would require four sheets between each pie or section.

The winding will now take up 60 pies x 1/8 in. each = 7 1/2 in. and another 1/2 in. for the four end coils. making 8 in. Four thicknesses of paper x .006 in. per sheet x 64 coils = approximately 1 1/2 in. for the total thickness of the paper and 9 1/2 in. for the entire secondary. It has been found that a good length for the secondary might be considered that of the sparking distance of the coil if it is desired to err on the side of safety and have the least danger of the spark jumping inside of the secondary or across the surface. To prevent this on the outside it is necessary to increase the thickness of the insulation if a short secondary is used, and this has its disadvantages. In the coil being designed it is desirable to make everything as immune from breakdown as possible, and therefore another two or three inches will be utilized for insulating purposes.

From the manufacturers' standpoint, paper discs are a more expensive insulation than the wax in which coils are imbedded, so that it might be well to increase the number of paper sheets from 4 to 6 and use the rest of the space for insulating compound, or for oil if the coil is to be immersed. It is convenient to cast the secondary in a number of sections, and between these sections the remaining space will be distributed. The paper discs should extend beyond the winding at both the inside and the outside diameters.

It will now be interesting to compute the length and weight of the secondary winding. The average diameter of the larger pies is 5 1/4 in., which multiplied by N/12 gives 1375 ft. for the average length of one turn: 1375 ft. X 18 sections x 1368 turns per section = 33,900 ft. for the number of feet of wire in the larger sections in the center of the coil. Computing in the same way there are obtained 31,400 ft. for the intermediate and 38,600 ft. for the end sections, making a total of 103,900 ft., or about 19 3/4 miles of No. 34 wire weighing about 15 lb., and having a resistance of about 27,000 ohms at 20° C. The four end coils add a weight of about 1/2 lb. of No. 28 wire. The primary contains over 400 ft. and weighs over 10 lb. and the core will weigh about 20 lb., so that the wire and iron alone in the coil weigh nearly 50 lb. without insulation or mountings.

The Condenser

Working with a Wehnelt interrupter the coil will require no condenser. With a slow vibrating hammer break and heavy current it should not require more than one microfarad of condenser capacity. The capacity should be made adjustable for different conditions of working and different currents and interrupters.

The formula for the capacity of a condenser in microfarads is

_ 885 x effective area in sq. cm. x K x 1010 thickness of dielectric in cms. If the same paper is used as before, .015 cm. thick, there should be two sheets, which will make a thickness of .03 cm., and from

Table II

Average Specific Dielectric Capacity

Air ................................. 1

Ebonite ............................. 2.5

G lass

Flint •.............................. 5-10

Crown ............................ 3-6

Gutta percha......................... 2.7

India rubber ................... 2.3

Mica................................. 6.5

Shellac..................... ........... 2.8

Sulphur .............................. 3-4

Paper impregnated with paraffin ...... 3.7

Paraffin wax........................ 2

Petroleum ........................... 2.5

Turpentine .......................... 2

Table II. it is seen that the value of K tor this paper is 3.7. Let us now find the capacity per sq. ft. of active surface. One sq. ft. = 930 sq. cm.

C = (885 x 930 x 3.7 x 1010) / 0.03 = 0.03 m.f

Thus one can easily calculate the number of sheets of foil and paper for any desired capacity by taking the value of .01 m. f. per sq. ft.

Summary Of Results

Core. 2 1/8 in. by 24 in.; insulating tube, 26 in. by 2 3/4 in. inside diameter by 3 1/2 in. outside; primary,

2 1/2 layers No. 12 wire, 230 turns per layer; secondary, No. 34 wire wound in 1/8 in. sections, each 6 in. external diameter.

The above coil when worked on with direct current at 110 volts with a Wehnelt interrupter will give a 12-in. spark on about 12 amp. of current, using 165 primary turns, and a 14.5-in. spark using 13 1/2 amp. The secondary terminals should be so placed as not to allow a greater spark than 14 1/2 in. Using two full layers of the primary winding and 10 amp. of current there can be obtained a heavy and powerful discharge across an 8-in. gap.-"Electrical World."