Quality Improved by a Filter Choke

Radio Record, Volume II, Issue 30, 8 February 1929, Page 26

Quality Improved by a Filter Choke

Details of Constructing and Using

CORRESPONDENT writes: hi 6An output filter is a useful device and is a great aid towards stability in a low-fre-quency amplifier. In certain ’ types of receivers it is prac- >}. tically essential, -for reasons . ' that will be given later. Most readers. know that it is. the ysnel practice to choose a power valved of some kind for the last socket-the output stage-of a set designed for loudspeaker work. Now in. many of the designs published in the periodicals a choke-filter output is included as. part of the set. Tea others there is no filter device. Whether or not such an arrangement should be included depends largely upon the purpose for which the set is intended. ; ‘It really reduces itself to @ question of D.C. resistance. ‘The average power valve tnkes @ fairly heavy current. Readers may not be quite clear as to why a power valve is necessary.

: Power Valves. we lmow ‘that if we wish to work a loudspeaker we must amplify. the signals #o that the volume is adequate for our particuler requirements, I¢ the signal will only work a pair ef "phones at fair strength, it is no good expecting the speaker to give you anything more than ‘phone volume. The. actual signals which the spedker receives must be many times greater than this. Wher so much enefgy is being havidied, the last valve must have a chatacteristic which will pérmit sufficiént grid. bias to be applied, in order to-avold distortion. This necessitates a valve having what is termed a low impedance, or in other wétds a powér valve. . These valves take a fairly high anode current because of their ¢omparatively low impedance or resistance. I have explained this, because it is all eonnected with output filters.

Large Volumes. WtH small sets, such as two or three-valvers, it is. scarcely possible for the volume to be latgé enough to overload a small power valve with, say, 120 volts H.T. and 9 volts grid bias.

This applies in particular to the former class of receiver. The exception might he if the set were employed within the "shadow" of the local station; but even so, one would expéct the detector valve to be overloaded first in the case of a two-valve set. A small power valve can be used fairly safely directly in circuit with a loudspeaker winding, because the current is not so great as to cause damage to thé windings (assuming the instrument to be of high quality) and because the drop in volts across the windings is not large. ‘When, however, we come to the question of four and five-valve sets, the use of an output filter becomes very necessary. Such sets aré quite capable of délivering sufficient signal volume to overload 2 stnall powér valve on the local station. A super-power valve, that is, one designed to handle a greater volume, then becomes’ essential. Now a super-power valve may easily take twite the anodé current required by a: small power valve, and in consequétice it is not wise to connect the delicate windings of a loudspeaker in series with it. It is not solely a question of current, since one has also to consider the mechanical stresses on the windings producéd by the gréatéer power which is béing handled. Preventing L.F. Oscillation. HEN, again, the heavier current increases thé volts dropped, or in other words the volts lost across the loudspeaker windings. This drop is

equivalent to so many volts less H.T. at the anode of the valve, and it may be serious with one of the super-power type-in extreme cases producing distortion. ;

Thus it is an-advantage to pass these heavy anode currents through the robust low-resistance windings of a suitable filter choke, feeding the speech or music impulses to the speaker via a condenser, in this way isolating the loudspeaker from everything except the fluctuating currents, In a large set, which is capable of handling a considerable intensity, there is always a danger of low-frequency oscillation commencing. By separating the steady anode current from the music impulses. with a choke-filter circuit it is frequently possible to stabilise © an otherwise troublesome set. — So much then for the value and use of a choke. Following are constructional details as described by ‘Megohm"’ some time ago. These should interest those who, having their holidays, wish to construct something that will improve their recepon.

Winding the Coil. THE spool is made on a wooden former 7-8in. by 3-8in. by 2 3-16in. long. This is covered with one or two thicknesses of manila paper and ends of thin fibré or stout millboard fitted on and glued. The winding should be done a jig and the spool ends supported in some way by cheeks fastened to the winding spindle, in .order to prevent them being forced outwards by the wire during the process of winding, or it can be glued firmly to the manila paper. The beginning and end of the winding, which should be of thicker wire, are passed through holes in the spool ends. The spool is to be filled with 36's enamelled wire, over half a pound being required, which gives about 8000 turns,

"The wire is run in irregularly in patches, piling up a certain thickness, say 1-16in., and then passing along to another pile until the other end of the spool is reached. Then a strip of writing paper is put round, care being taken that it closes up to the ends, as although the wire is being put on irregularly, one layer must not be allowed to come in contact with another by wires sinking down at the ends. This easily happens, especially if the spool ends are not well supported. When the spool has been filled and the leadout wire soldered on and passed through the hole in the end, a covering of stout paper and then a piece of empire cloth are put on to protect the winding. The Laminated Core. (THE core is built up of laminations of ordinary tin-plate, 28 gauge, in full size, eighty. pieces being required. The tin has,the advantage of being thinnér than black iron can be obtained, and thus serves to more efficiently suppress eddy currents in thé coré. Stalloy can be wsed, but it is not always procurable. One sheet of tin 28 by 20 inches will be more than sufficient for the laminations.

The tin should be marked out into 80 squares 8} by, 22 inches, these squares being cut out, after which a cardboard. template is made of a complete -lamination in one piece, to be laid on.eath piece in turn and scratched round the two "windows" to mark. their position. oy Now a piece of card is. taken the size and shape of half lamination, but without the windows. This is used as a guide to scratch the diagonal line

across each. Then cut each square’ right through on the diagonal line,. afterwards ¢utting out the "window", pieces. bending the metal where necessary, afterward straightening out by stroking on the bench with thé back edge of a flatiron, the point being held up in thé air. This method of straightening the cut plates is‘easy and effecve,

The laminations are now to be shelced with one coat on both sides, for which -purpose shellac is dissolved in methylated spirits and applied with a brush. This coating of shellac is important, as its purpose is to insulate each lamination from the next. "When the shellac is thoroughly dry, Jaminations, all the same way round, are packed into one end of the coil, the last few being pushed in under the top one, and having the sharp corner cut off the central piece to prevent the manila being cut. A piece of strong paper is to intervene between the two ends of laminations where they meet outside the coil, so large enough pieces are placed there, while the laminations on the second side are packed into place. AS many must be got in as ssible, so that both iots are free from @: looseness. The ends of laminans must both press. against the paper, so that they are only separated by its thickness. This forms a gap in the iron, the object of which is to prevent magnetic saturation of the core by the direct current passing through the coil. . Final Operations. THE wooden clamps, four in number, are 44 inches long and T-16-inch square, drilled at each end to take a

3-8 brass bolt two inches long, about a half-inch of which can be cut off after -all is assembled. A washer should be put under the bolt head to protect the wood, but at the other end the nut will be sufficient. A slip of ebonite about 2 1-8 by 1 1-8 inches is now screwed to the clamps and fitted with two terminals, to which the leads are connected underneath. When completed and stood on end the choke occupies a floor space of about 8 by 1} inches, and, of course, the ebonite slip may be placed in any convenient position. A small diagram of the circuit is included. The wooden clamps should be finished with the shellac applied with a piece of rag and rubbed. The clamping bolts must not be allowed to touch the ends of the laminations, and so connect some of them together. If care and finish are put into the work the coil has quite a neat appearance. If desired the outside of the laminations may be finished with black cycle enamel, Experiment may show that a higher value than .001 gives better results for the fixed condenser across the speaker. With no condenser there, tone is inclined to be harsh and thin. Too large a capacity causes woolliness.

tralisation, and any tendency toward undesired of the radio-fre-quency stage completely elminated. This condition results in better tone quality, as the regeneration in the director circuit may readily be controlled without affecting the stability of the radio-frequency stage. "THE constructor should not find it difficult to complete this new receiver, especially as units comprising the coils, condensers, and controls are already on the market. These new units are very fine, and should add greatly to the popularity of this receiver. One dial control is used, and a special balancing condenser allows’ of maximum tuning capacity.