Notes on "B" Eliminator Construction

Radio Record, Volume II, Issue 45, 24 May 1929, Page 37

Notes on "B" Eliminator Construction

Bias, Resistances, Chokes and Filament Heating

By

MEGOHM

ITHOUT a doubt, "B" battery eliminators have came to stay. In striving for quality, the key-note of modern radio, the adoptien of mainsproduced plate supply is a good first step. -Hven with — a crystal and two-valve ‘amplifier, the adoption of a simple half-wave "B" eliminator will often effect a remarkable improvement in quality of reproduction compared with that produced from dry batteries, especially when the temptation is present to run the latter after the inevile drop in voltage that should cause them to be discarded. Radio dealers now stock many of the components required for "B" eliminators and when all these are purchased ready-made, the connecting up is a comparatively simple matter. The new "Listeners’ Guide" deals fairly exhaustively with eliminator construction, but a few hints regarding details will be of interest to constructors, Dissipation of Resistances. [t is usual to take the full voltage of the eliminator to feed the plate of the power-valve. This voltage may be between 100 and 400, 150 to 200 being the usual, and suiting the average receiver with a smaller powervalve in the last stage. This maxi..f mum voltage of the eliminator then has to be reduced to suit each of the other separate stages, early audio, detector, and radio-frequency. In general yse there are two methods of effecting this reduction. The first is to provide each of the three outputs with a sepa:ate resistance, variable or fixed, and the second method is to connect "B" positive and "B" negative by means of a continuous high resistance ranging of betiveen 12,000 and 30,000 ohms, the various Voltages being tapped off at different points calculated to give the required voltage, which decreases gradually until "B" negative is reached, where the voltage is zero. . If it is proposed to obtain grid bias by the latter system, a suitable amount of resistance is continued past the B negative point, and the grid-bias values are tapped from suitable positions. These points are usually variable along ¢: portion of resistance. It is not ays an easy matter for the amaFur constructor to obtain entire gridbias in this way without the slightest trace of hum, and many prefer to obtain the power-valve bias only, providing lower voltages from the dry "O" battery. One of the objects of this paragraph is to point out that when a bias resistance is u: ‘1 in this way, it is to be remembered that the whole of the plate current of the power-valve returns to "B" negative through the bias resistance, so that the latter must be of higher current rating than might be considered necessary. Where the last ‘valve is a small power-valve, taking "only four or five milliamperes on the late, the ordinary composition variable resistance turned for the purpose into potentiometers, may be used. Where the plate of ‘the power-valve draws 15 to 20 mils. or more, resistance that will not soon disintegrate

must be provided, with the comparatively heavy current flowing. For this purpose a pair of 400ohm. potentiometers, with their resistances placed in series, and the arms each providing a separate bias voltage, will be found admirably suited. Wirewound resistances, in series, may be added to make up the required total, put these should be of high rating, 30 to 50 watts dissipation, otherwise they will heat and soon burn out. Breaking Down the Voltage. GIMILAR high rating is required in the resistances used to reduce the maximum voltage to 90 volts, when 100 to 250 volts are to be dropped. The best type of resistance to use in this position is the Ward-Leonard type of fixed resistance, consisting of a wire winding embedded in vitreous clay and capable of heating to a high temperature without damage. These resistances dissipate 44 watts, and are obtainable in the following values in ohms: 1500, 2000, 2500, 3000, 3500, 4000, 5000, 10,000, 15,000 and higher values to 100,000. handy method of mounting these resistances is shown in a diagram, short lengths of 14’s copper wire being inserted in a strip of wood, the resistances standing upright, thus occupying minimum space. ‘The wire connections are cut short, leaving about one inch to connect in series. The actual potentiometer or voltage divider may be purchased ready-made with a system of tappings provided. and where the maximum voltage for the power valve is greater than that which the potentiometer is constructed to deal with, extra resistance may be added between it and "B" positive power in accordance with the table on page 69 of the "Guide." Between 90 volts and "B’ negative a resistance of low dissipation, as low as 5 watts, may be used. as the current carried by this portion is only a few mils at the higher end, and less still at "B" negative. In this portion variable composition resistances may be employed, high values for separate outputs, or lower values if connected in series to form a voltage divider. Variable Resistances as Potentiometers. BUILDERS of eliminators who wish to make up their voltage divider will find it an easy matter to alter a variable resistance so that it will connect up as a potentiometer-that is. with a connection: to each end of the resistance and a third connection’ to the rotating arm. Some variable resistances now on the market are made with a view to being so altered, and each end of the resistance strip is connected to a terminal bolt. The arm is alsa. connected to one end of the resistance by a strip of metal which is to be liber-

ated from contact with the resistance and secured by a bolt passed through a hole already provided in the bakelite body between the two resistance terminals, Transformer Construction. [THE one-inch core transformer is in every way quite suitable for an eliminator, and has the advantage of being compact, standing 5in. high in a floor space 4}jin. x 24in. This transformer is as large as is usually provided in commercial eliminators to supply as much as 350 volts full-wave. The only consideration is that thin gauges of wire must be used in ordez to economise space. The length of the spool is to be 24in. inside, and the size of "window" will be about 2 18-16 x 18-16in. The primary winding of 1850 turns may be of 30’s or 32’s s.w.g. enamelled, occupy ing 13 or 11 layers respectively, 150 or 180 turns per layer. Next follows the filament winding, which should be put on whether required or not, to act as an electrostatic shield between primary and secondary. If not required for filament heating, one end is connected to earth and the other coiled up and insulated. Double cotton-covered 22’s will carry one ampere, and the necessary 53 turns for 6 volts will oc cupy nearly a layer. A centre-tap is provided at the twenty-seventh turn in case a potentiometer is provided outside across leads, A secondary winding giving 200 volts in each half is well suited to the average receiver, aS when

rectified and smoothed, will give a maximum of 150 to 160 volts. The secondary winding of 3700 turns is made continuous, occupying 15 layers, about 250 turns of 36’s wire on each layer. The centre-tap is taken out half way at the 1850th turn. Half a. pound of 386’s enamelled wire will be required, and the same weight of primary wire. For an eliminator it is verv imnortant that the stalloy laminations are tightly held, so that there may be no mechanical hum. They must be packed as tightly as possible inside the spool without cutting the interior of it with sharp edges or corners. The last two or three strips pushed in should have the entering corners rounded off, and should be placed under the top strip, so that there will be no chance of cutting the spool. As a final tightener, two 4%in. strips slightly narrower than the rest, should be hammered in separately near the centre of the core after the clamps have been lightly screwed up. In sizes above the one-inch core there may be a considerable amount of "spring," depending to some extent upon the flatness or otherwise of the stalloy, and it is a good plan to make a small allowance for this by building the core a little over size, increasing the size of the former accordingly in one direction, say, 1-8in. for the 1jin. core, and up to tin. in the larger sizes. It will be found that when the clamps are tightly screwed up, the size will be very near that required, and there will be full compensation for shellacing that may be thicker than necessary, and for. uneven cutting and consequent gaps at the joins of laminations. Any core is better built slightly over size than under. The small iron belts used by coachbuilders may be used to secure clamps. The total cost of this transformer, including wire, 2 dozen 8ft. lengths of

stalloy, panel, clamps, ete., as described in the "Guide," should not exceed 18s. Sometimes spool ends are dispensed with, and a square cardboard tube made to fit not.too tightly over the former, and on this the layers are wound. When complete, the ends are covered with a mixture of equal parts of resin and beeswax melted together for pouring. The composition off the tops of old dry batteries may be used. Heating Valve Filaments with A.C. HE filament of any valve heated by alternating current must be protected from any voltage in excess of the stipulated maximum. This applies to both ordinary power-valves and al-ternating-current valves. The danger to filaments is caused by variations in mains voltage at different times during the day, and by "surges" of short duration at any time. The usual method of making as. much provision as possible against such variations is to always keep some reserve in the filament voltage by running a reasonable amount below the maxmium. A ten per cent, rise in 230-volt mains, though rare, is possible, and would mean an increase of about half a volt on a 5 or 6-volt filament, so that a half- volt reserve may be considered fairly safe on a power-valve. ‘ Alternating-current valves cannot always be run with this amount of reserve, and some means is usually provided whereby the input voltage to the filament transformer can be regulated to suit the line voltage. One method employs a small auxiliary auto-trans-. former with variable tappings, whilst another uses a variable power rheostat of suitable resistance in one of the mains leads. The actual value of such resistance is governed by the mains voltage and the amount of current drawn by the receiver. Another met’ easily carried out is to put additional turns on the transformer prima: beyond the specified number. On the lin. core 150 extra turns could be ovided, .ipping at every 25 turns. each tap covering a rise in the mains of 5 volts, a switch providing easy regulation. This principle .f additional primary turns can be used on eliminator « experimental transformers. giving an extra means of varying the output by a small amount. or adjustir to suit particular conditions. Very often the specified turns leave room for 50 or 60 turns to complete the last layer, and this space can be wound in, with a tap at the specified number. Tzppings on Separate Secondaries. ) HEN the two secondary coils of a transformer are wound as a continuous coil with a centre tap unalfway, tappings are taken out, if required, near the beginning and end of the coil, and nresent little difficulty. In the case of winding each secondary coil separately, each to occupy barely half the total length of the spool, leaving a small space between to be filled witt insulation, care is required to ensure that the two coils are connected eoriactly. Mach must be wound in the same direction, and the end of one and beginning >of the other are connected together and from this connection the centre tar is taken. When a tapping is required in each coil so that a lower voltage than the maximum is made available, the posi-

tion of the taps. requires caref 1. notice. Suppose that in each secondary of 1000 turns a tap is to be provided 200 turns below. the maximum. The first coil is wound with 800 turns, then the tap taken out, and 200 turns added to finish. The second coil is then wound, turning in the same direction as the first, and when 200 turns have been wound, the tap is taken out, after which 800 more will ec plete the coil.

The beginning of the first o(! Bana end of the second are connected an& the centre tap is taken from this connection. If the lower voltage is to be used, the two taps will be connectd to the rectifier to form "B’ + 1 and "B" + 2. For maximum voltage the end of coil 1 and beginning of coil 2. will be taken to the rectifier. The beginning of the coil is where winding is commenced when putting the turns on. Centre of Filament Winding. ENTRE taps need not be provided on filament windings if the electrical centre is determined outside by means of a 60-ohm potentiometer connectéd across the filament terminals of- ‘the power valve constituting the ideal method, as by adjusting the movable arm, the position giving least or no hum is easily found. A potentiometer for this purpose can easily be made

~~ rom a 80-ohm rheostat, which may be ounted flat on the baseboard along‘Side the power valve. The resistance of the potentiometer should not be less than 60 ohms, but no harm will result if it is considerably higher, say up to "80 ohms, All that is necessary is to cut the ‘connection between the arm and one end of the resistance. The arm is then connected to earth or "BR" negative by means of a wire soldered to its connection. Bach end of the resistance is then to be connected to its respective filament lead by a resistance wire of not less than 15 ohms. Nine inches of No. 36 s.w.g. nichrome resistance wire on each side would be right. This should, be wound on each end of a strip of mica or other insulating material, about jin, long, cut as shown, notched at the edges with a pocketknife, the wire wound on and passed through holes, and the strip bolted to one of the rheostat terminals as shown in the diagram. One end of each nichrome wire connects to one end of the rheostat resistance,and the other end to one filament lead of the power valve by means of a copper wire soldered to the end of the nichrome. Radio Frequency Chokes, THE inclusion of a radio-frequency choke in each secondary lead to the rectifier is well wortl while, as selectivity of the receiver is m.intained owing to the chokes keeping back any unwanted R.F. signals picke’ up by the mains acting as aerials, These chokes are easily ma. by winding 1000 turns of enamelled wire gauge equal to that. used on: the secondary coils of the transformer, upon a flat pod. of thin fibre, ebonite, 01 1-8in. cardboard, and about 2in, in ¢* meter. If 36’s wire is. used, the slot need be only 1-8in. wide, but if 32’s is used the slot must be jin. wide. The centre of the spools are cut f m a narrowcored wire bobvin, secured with seccotine, and may.be bolted through the centre or secured under baseboard with a screw. The sides of the spools may be square, octagonal or round. A Few Don’ts, Dort mount the transformer and chokes di.ect upon a tin-covered L 3eboard, or the tin floor will vibrate and cause hum. 2st to place the tin under the board, with a space for the wiring betwecn tin and board. Don’t cut stalloy strips until coils are wound and complete. Correct sizes are thus eusured. Don’t forget to switch off the nower supply before touching connections of aM fhe a and always switch on J. the filamert battery, and switch off before the filaments are cut out, Don’t build .e container until all the components nd condensers have been assembled. The required space only may- then be provided. Don’t omit to shellac double cottoncovered filament win’ igs, and don’t cove: them until dry. Don’t forget that all wires carrying alte. .cting current should be shielded . from those carr)’ ig the output vyoltages, and from the voltage divide? and _ resistances. Don’t omit to bind the outside laminations of a transformer with adhes._Ive tape-where there are no clamps. 75) aceenepemeasichestgnsttsnetatseemananereepassanassasisindnamtaabenimaasteantnnnE .- YF