A Direct Current Eliminator

Radio Record, Volume IV, Issue 14, 17 October 1930, Page 27

A Direct Current Eliminator

Third Instalment

By

MEGOHM

HIS scheme for full plate and filament supply allows of any number of valves being supplied with "A" and "B" current from d.c. mains, together with grid-bias for the audio stages. An anode bend detector is recommended, grid-bias being applied,.and the usual leak and condenser dispensed with. A valve of very high impedance may have a high plate voltage and require only 14 volts bias, whereas the 201A type will’ require a higher bias, obtainable from a 9-volt tapped bias battery. Qne-quarter ampere of current will be drawn from the 230-volt mains, which is economical running-16 hours on one unit of electricity,, or where current costs as much as 3d. a heating unit, the set may be used for 3 hours every day at a cost of 17/- per annum for current, or nearly three hours’ running for one penny. Compare this with the continual expense of buying new "B" batteries and the cost of recharging accumulators. Running a Dynamic, OME constructors will be making inquiries as to the practicability of using the current to excite the field winding of a dynamic speaker, so it may be stated that a suitable field winding is not likely to be encountered. A winding to pass .25 amp. would only require about 50 volts, and consume 413.5 watts. A field winding to conme 60 watts would be considered extravagant. A 230-volt winding consumes about 6 watts. A speciallywound field would be required, using 26 wire, of which 1050 yards would have a resistance of 98 ohms, giving 4900 turns on the average "pot," which is more than would be necessary with .25 amp. passing. The d.c. resistance must be not more than 500 ohms to include in the filament circuit. The field winding takes the place of all or part of the extra resistance in the filament circuit, and would render the use of the large smoothing-choke unnecessary. Valves to Use. : LL valves used in the receiver must take .25 amp. on the filament, but the voltages may be all the same or different. All filaments are ired in series, including the two puXb-pull 171’s. As the filament in each i171 is at a different potential by the amount of drop across one filament, 5 volts, the bias applied to the grid of each must differ by that. amount. That is to say, the grid of the valve (B) nearest to the negative end of the filament circuit requires 5 volts more bias than the other pushpull valve (A). ‘This constitutes the little difficulty that was mentioned earlier, If a push-pull input transformer is used with a centre-tapped’ secondary, grid-bias is applied to this tap, and will therefore be equal on each grid. We can increase the potential on the giNd of B by placing in the cirevit a 4-volt bias battery as a means of increasing the bias on one valve. Although the bias is actually half a volt short of the required deficiency, this is a very small error in 40 volts-one-eightieth-and clears the difficulty.

It may at first appear to be a poor expedient, including dry batteries, but this small battery saves running the push-pull filaments in parallel, which would double the current consumption of the receiver without any other benefit, and the extra current would cost far more than the 44-volt battery, of which good makes in the writer’s pos-

session have kept full voltage up to 18 months, When there are fluctuations in the filament supply, either a.c. or d.c, a heavy filament should be used, as it holds the heat better than a light one, making the fluctuations less noticeable, and this quality is particularly needed in the detector. The valve that best fulfils this condition is the 112A, which should be used for the detector and first audio. The 201A type will suit the r.f. stages. If valve voltages are mixed, all that is necessary is to adjust the extra resistance 4 ohms more or less for each volt difference from 5 volts in the filaments. If 201A’s give noisy working, try 112A’s for rf. stages. The Resistances. OTALLING up the various resistances in the filament circuit gives 861 ohms, and this deducted from 920 ohms, the full resistance required, gives 559 ohms, which is the minimum extra resistance required in the circuit. In the table of resistances 580 ohms appears as the extra resistance, which allows 40 ohms over, covering a differ: ance of 10 volts. For fine adjustment a rheostat of 10 to 20 ohms to carry + amp, may be added to this, but failing a rheostat, the end of the nichrome wire might be made variable by securing to‘a terminal. The resistance of a filament is found by dividing the working voltage by the current passed, so the quarter-amp filaments have a resistance of 20 ohms.

The total resistance required’ across the mains is found by dividing the supply voltage, 280, by ’25, the current to be used, which gives 920 ohms. This is to be the total combined resistance of filaments, choke or field windings, bias resistors, etc. included in the filament circuit, to which a small amount of variable resistance should be added. The total resistance will’ be made up as follows :-

Totalling up the various resistances and deducting the sum from 920 gives the amount of extra resistance required in the circuit, but 50 ohms extra at least should’ be included and made variable so that fine adjustment may be effected to suit mains variation ete. Note that in this circuit every 4 ohms gives a drop of one volt. Actually 39 ohms extra resistance is shown, but this may be _ increased slightly to allow for variation. The yariable portion is shown at the plate supply end, because this will probably be the most convenient point, especially if a lamp is used for the 400. ohms. Obtaining suitable resistances is rather a difficulty owing to the comparatively heavy current to be carried. The best way to make up the 580 ohms is by utilising an electric lamp for 400 ohms, making up the balance of 180 ohms with 28 B. and S. nichrome wire (bare). Only two electric lamps work out to a suitable valuea 230-volt 32 candle-power carbon filament with a resistance of 413 ohms and a metallic filament lamp of 30 watts for 110 volts, having a resistance of 400 ohms. ‘Two suitable lamps may be used in

parallel for the 580-ohm resistance, but as they will not be brightly lit, care must be taken to make sure’ that they have sufficient resistance when testing out. A 50-watt 230-volt mf. lamp works out to 1059 ohms, and passes .217 amp. Two in parallel would have a resistance of 529 ohms when fully lighted. They would jointly be capable of passing .434 amp, but would only be getting .25 amp, between them, hence the reduced brilliance. Necessary additional resistance would be made up of 80 B. and S, nichrome wide. When using lamps as resistances it is necessary to arrange that the voltage drop across them is no greater than that for which they are rated. otherwise the filament will burn out. The 180-ohm resistor may be constructed upon a strip of fibre so that heat may be easily dissipated. Of 28 B. and S, nichrome, 43 feet must be used, with a foot or two extra for preliminary use. If the fibre is six inches wide, each turn will be one foot, so 43 turns will be required. Good spacing will be 10 or 12 to the inch, with notches made at the edges of the strip with a hack-saw. It is best to measure off the wire, which may then be wound upon any convenient: size of strip, say 4 inches wide, passing the ends of the wire through fine holes, At the variable end leave 2 feet over, the object of this being to introduce plenty of extra resistance while making preliminary adjustments of the filament circuit. The wire is gradually reduced until the correct reading, 5 volts, is obtained across any of the filaments. nsing a voltmeter. The 1384-ohm bias resistor for the negative end of the circuit must be constructed of the same wire upon the same plan, using exactly 383 effective feet of wire. If preferred, the fibre strips may be conveniently stowed under the eliminator or receiver basehoards, but a clear air-space of at least jin. must be provided on each side to prevent heating. Bearing in. mind that the nichrome wire has a resistance of 4.155 ohms per foot, small bias resistors may be made up as follow: 6 ohms, 18in.; 7, 20; 8, 23; 9, 26; 10, 29; 12, 35; 16, 46;

20, 58in. Wound upon a narrow strip of fibre they may be tapped by twisting a piece of copper wire round at the correct position, not omitting to serape the nichrome bright. Plate Supply. URRENT drawn by the five valves "as specified will total as follows: Rf. 2.5 ma. detector 4 m.a., audio 7 ma, 171’s 40 m.a., or a total of 53.5 ma, The 171 plate current does not traverse the small choke, which leaves it 18.6°m.a. to handle. Specifications for this choke are given later as "Choke A." Passing 13.5 m.a. the drop in this choke will be 7 volts, leaving 196 volts. The audio filament is 61 volts above zero potential, so this must be deducted from the 196 volts, leaving 135 volts for the plate of the first audio 112A, for which the correct bias is 9 volts, 10 being provided for. For the 112A as anode bend detector, a plate voltage of 60 to 70 will be found suitable, with 30,000 or .35,000 ohms resistance. About 6 to 9. volts bias will probably. be required. For the r.f. valves, one or two, 196 less 71 leaves 125 volts to be reduced to 90 volts by 14,000 ohms, or to 70 volts by 22,000 ohms. Half these resistances for two valves. Plate resistances need only be 5-watt rating. Ferrantis specialise a wide range of wire‘wound resistances. By using two in series almost any value may be approximated. By-pass condensers are 2 mfds., 500 test. An extra r.f. valve is worth while, ‘because . the filament current costs nothing extra, and the extra plate current is negligible. Deduct 20 ohms from the "extra resistance" for each extra valve. Some constructors may

adopt @ single output valve and an extra r.f. stage, but this is not recommended if good quality and volume are desired.

Pilate current for the 171’s will bave an actual effective potential of 203 volts less the potential of each filament,

which gives 182 and 177 volts respec-tively-the maximum for the 171 being 180 volts. Where it is noted that on 110-volt circuits only 90 volts can be put on the plates of 171’s, constructors working on 230 volts may feel well satisfied, as the amplifier will also give good reproduction on gramophone records, A probable drop of about 5 volts in,the primary of the output transformer has not been taken into account. In making up a series filament circuit, the valves may be placed in any order, but each filament has also to carry the plate current of preceding valves, and for this reason the power-

valves are placed last in the circuit, so that their plate current does not traverse other filaments. The 171 bias resistor should be placed in the receiver, obviating an extra lead. If the 1 mfd. plate by-pass condensers are placed in the receiver, connect as shown to associated filament at L, M, N, but if they are in the eliminator connect as shown there. For clearness in the diagram the ed

purely radio portions of the circuit are mostly omitted, and the location of a primary winding in the plate circuits is indicated by a few turns ang." the letter "P." 240-volt Mains. F 240-volt mains are to be the supply, add a 40-ohm resistance of nichrome wire (9ft. 8in.), as used for other resistances, placing it in the positive lead before the large choke, Other resistance values in the circuit will then remain as shown. At this point the description will be held over till next week, when the

series of articles will be concluded with a description of how to, alter or build a d.c, mains operated receiver.

Resistance Volt. ohms drop Choke wha. cL OCS 2 Extra Res. 5 8 0 1 4 5 Valves 10 0 2 Bias 2 res. 174 4 9 5 9 2 4 aos to