The Frequency Test from 2YA

Radio Record, Volume II, Issue 4, 10 August 1928, Page 28

The Frequency Test from 2YA

Comprehensive Discussion on Note Losses and Amplifications —

By

Megohm

WiE frequency test con ducted by Mr. Bellingham from 2YX on July 24 was a matter of considerable importance to radio listeners. It gave the owners of good receivers "an opportunity of getting a fairly exact indication of the desirable qualities incorperated within them, and it wfforded all and sundry a useful gauge of the small: or -great shortcomings of their receivers, whether factory ready-made or home-built. Mr. Bellingham set wv high standard, which is good, becanse an ideal’to be aimed at should always be bigh, and from this moment, experipienters..and..constructors should ain at reproducing as evenly as possible, all frequencies between 10,000 and cr. Generally speaking, this is no easy matter, and in many cases will inyolve » considerable amount ‘of experiment. But the knowledge gained will be great. and the improved tone and clarity of reproduction will well repay the trotible, and although it may be ‘found impossible to actually reach the range mentioned, any improvement mude to increase range in either an upward or downward direction will be rewarded by an improvement in quality, provided that necolInpanying — es‘sential factors have also been attended to. a Fhe Ideal Range.

HE ideal frequency range would be from 10,000 to 16 cycles, but it must be said that this near approach to -perfection is not entirely necessary, and for practical purposes a fairly even reproduction from the loudspeaker of 30 to 7000 cycles is very high quality, and is still very good if all frequencies between 80 and 4000 are delivered. This latter is the minimum range in-

dicated on the recording chart, But, us was pointed out in the lecture, with to maximum frequency of only 4000, the reproduction of certain musical instruments will be found to lack the characteristic quality in a greater or less Cegree, on account of the fact that some of these sounds coutain high harmonics with frequencies varying from nan octave ubove the fundamental to as high as 30,000, the complete sound heing composed of the fundamental intermixed with several harmonics or higher notes at varying frequencies. The human eur reaches its limit of audibility at 30,000 or somewhat below that figure, and many of the higher audible frequencies will be so weak, comparatively, that they may well be omitted from musical sounds without their absence being readily detected. It is when the cutting out of harmonics or overtones is carried downwards, ap-. proaching Loo near the fundamental. that the stronger harmonics .are lost, quality suffers noticeably, and reproduction ceases to approximate faithfulness, Kven Amplification.

&) LD E-by-side with a suitable range of reproduction, the ideal is un even amplification of every note from the highest to the lowest, so that all have the sume volume as sounded at the studio. This is perhaps the most difficult condition to fulfil, and only an approximation to it will be expected. All audio amplification systems have in a greater or less degree the fault of reproducing easily the middle frequencies, and falling off badly in their handling of the higher and lower frequencies, as to both amplification and

purity. To extend our range of reproduction in either direction it is necessary not only to secure greater cumplificacion, but to obtain it free from distortion, In some types of audio amplifiers it will be found easy to extend the range in one direction by losing range at the other extreme, but extension in both directions is desired. . Generally speaking, reproduction of higher, notes is a simpler matter than of the lower sounds, and the average receiver casily handles the middle and higher, but not the highest frequencies, whilst tuaking «a more or less hopeless attempt to deal with the deeper sounds, If ‘the loss of high notes begins too low in the seale, the clarity of speech is affected, and to reproduce the 8 and "Z perfectly a receiver must reach 6400 cycles. At 5000 speech begins to suffer very noticeably, though it may not. actually lose much in intelligibility. Yet although ‘the average receiver will handle high notes of normal volume fairly well, any increase in the yolume of, say, a singer's voice, will produce more or less obvious distortion or "blasting." Likely enough, this {rouble is caused by the overloading of the detector, for it is in the radio-fre-queney und detector stages that high notes are inclined to suffer most, and low notes mect trouble chiefly in the audio system. : Megehm’s Receiver. PEE writer's receiver acquitted itself in the test just about as Was ex pected. This receiver gives good qualitv reproduction of music, including all bass notes, 2nd also brings out with good volume the high-pitched instruments, such as fiute, piccolo, cornet, and violin. In its present form the set ix n three-valve Browning-Drake tuner, followed by a two-stage amplifier, the first stage of dual impedance as described on April 6, with a speciallydesigned plate impedance, with an inductance of 250 henries, obtained by employing an inch-square core of stalloy. his, coupled to a PM 3A as detector, gives high amplification of low notes, without unduly losing the higher ones, as evenness of amplification is a characteristic of impedance coupling. The first audio valve is a PM4,.followed by a Ferranti AI.3 transformer aud a PM 256 super-power valve and chokecondenser output filter. An addition has been contemplated for some time {o increase amplification of high notes, but more pressing matters continually assist its posthonement. Two speakers are employed, a Brandes Ellipticon to ensure full justice being done to the lowest notes as well as the remainder of the scale, and to add brilliance to the whole, a smal: horn speaker is connected in parallel. This horn is not capable of properly reproducing the lower notes, and is not required to do so, therefore a fixed condenser of about 01 capacity has been placed within the base in series with one of the leads. By experiment the yalue of this condenser is decided so that all notes below about middle G are cut out, though

ve actually some are only weakened 80 much as to be negligible. The higher the value of the condenser the lower the frequencies passed. When actually listening to the test arrangements were made So that the ralve receiver could be easily switcher on nnd off, to give an opportunity of testing the range of a crystal also, listening on head-phones. This showed that the erystal-a fixed carborundum -has a range from 26 to 7000 cycles, weakening considerably on the higher notes. The valve set took the whole of the frequencies from 26 to T0090, ° losing them at the 8000 mark. Some of the lowest frequencies could not be produced by’ valve oscillations, the method adopted for most of the tones. so i: bass viol and piano were used, and this caused the production of overtones as well. In the case of the 26eycle note, the overtone appeared to be stronger than the fundamental. The low 34, 40 eyeles, on the drums, repro‘duced well, giving the true quality of the instrument. At 32 the overtone and fundamental were both heard, but on the erystal the overtone was not noticeable. All these low notes came through with good volume, which shows that the low-note amplification consistently aimed at has been-necomplished. From 5000 to 7000 the volume weakened considerably, and work will now be put in to increase amplification on the high notes without, if possible, sacrificing anything on the low. _ If low notes are to be heard, they" must be amplified m ch more than high: ones, and this the average reproduction system is unable to do, so the low notes remain weak and are lost, er are sufficiently heard to have a blurring effect, A broadeast station can determine how low and how high a note shall be put out. The cut-off may be at 40 and 5000 ‘eycles, or at 30 and 6000, or as predetermined. The lowest musical notes are put out by 2YA, but if always appears to the writer that BYA has : higher cut-off, and does not put ‘out the lowest frequencies, so that receivers unable to reproduce these low notes do not have them present to distort. Loss of High Notes. N this discussion of radio reproduction, it must be remembered that judgment is passed upon the sounds as reproduced by the loudspeaker, and that faithfulness of reproduction depends not upon one unit or portion of the receivers alone, but upon the whole combination, including the loudspeaker. The loudspeakers used at the 2YA studio, it was mentioned, reproduced every frequency from 26 to 13,000. The speaker in question is the Western Blectrie balanced-armature cone, a very popular speaker in the United States, though little effort has been inade to popularise it in New Zealand. The majority of radio sets in combination with their loudspeakers avg deficient in the frequencies above 38000) cycles. which means that the amplis- | eation obtained at these frequencies is not so great as it should be. So much stress has latterly been laid on the reproduction. of deep notes that this is

not surprising, yet without the low tones a great deal of the pleasant musical effect is lost. ° As already mentioned, the radiofrequency stages usually pass the high frequencies easily, but should be well ‘gesigned, or reproduction will suffer through their suppression of the lower tones. One particular type of set, placed under test in an American laboratory, did not pass any frequencies below 350 cycles through out R.F. stages. Signals suffer less by having several R.F. stages: of medium selec‘tivity rather than one highy selective stage. This is because the "side-bands are with high selectivity received with less amplitude or volume than the centre of the carrier wave. There are several factors that can cause high-néte loss in the audio amplifier. Two stages-of transformer coupling form the popular amplifier, as giv- « ing the greatest amplification per stage, but unless the transformers. are well designed they will nullify the good reproduction of which the remainder of the circuit may be capable. And the choice of suitable valves is also essential. Amplification must be wnaccompanied by distortion, otherwise it is useless. Some of the well-designed modern transformers are capable of siving fairly even amplification over a range of 40 to 6000 cycles, but the small-sized badly-designed type will not appreach anywhere near this performance, and on account of its low primary impedance, will give very poor amplification on the low notes. If the coupling device between audio valves is such that its impedance varies with frequencies, then amplification suffers af some parts of the musical scale, and the same thing happens in the average coupling system, transformer, or resistance, owing to the fact that the amvunt of amplitication falls off proportionately as the strength of signals increases. This latter may not always bé a disadvanlage, though it reduces the contrasts of light and shade in the musi. received. Double-impedance coupling gives mvure uniform response in this respect. If a transformer capable of amplifying low notes satisfactorily is to transinit high audio frequencies equally well, other conditions must exist, and the most important, perhaps, is the effective "apacity across the secondary windings. Small values of capacity across the secondary will seriously reduce the voltage of the higher frequencies. which is reducing volume. Resistance or condensers should never be placed across the secondary of a transformer, as volume will thereby be reduced. Fidelity of reproduction is also reduced by this practice. . One thing should be meade clear, and that is that the troubles now being discussed cannot -all be entirely eliminated, but only reduced to a negligible minimum. As amplification of low notes is increased, so there is liable to be a corresponding falling-off in the higher frequencies, which is a very common cause of high-note loss. Resistance Coupling. N resistance-coupled amplifiers, the umplification depends entirely upon the amplification factor of the walves employed, therefore, only those «with a high amplification factor, about -30, should be employed before the power valve. The value of the resist~ance in the plate circuit must never be less than the actual plate resistance of the valve, and.a high voltage must be applied to the plate circuit to overcome

this high resistance, and still leaye sufficient voltage for efficient operation. Correct: valves for the coupling condenser and grid resistor are also important. Increasing the capacity of the ‘coupling condenser reduces tendency to distortion, and increasing the grid leak value increases amplification, whereus too great increase of either produces tenllency to ‘"motor-boat." As already pointed out in this column, high notes are lost or weakened through the higher frequencies being by-passed through stray or unwonted couplings in the circuit which offer much higher impedance to the lower frequencies, and thus preserve their volume. Loss of Deep Notes. ‘ecu of the loss of deep notes is caused by want of even amplification, just as in the case of the high notes. To preserye the deep notes in transformer-coupled amplifiers, high primary impedance is necessary in the transformers, especially the last stage. To pass the deepest notes:a low ratio, say 3 to 1, is to be p-eferred to a high as the well-designed low ratio. instrument generally has more room tospare for increased. primary winding, which gives the higher impedance. Sometimes a high impedance winding following the. detector will give a certain amount of trouble in the shape of either whistling or eliminator hum, if i bB eliminator is used. The whistling or howling may be present if a LB battery is used. Sometimes a resistance of, say, 8000 ohms across the primary will cure this trouble. If a high impedance valve is used as a detector. This measure should not be resorted to if it can be dispensed with, as it slightly reduces volume. The usual by-pass coudenser from plate to filament should be provided aud different capacities. tested, from 0005 upwards. In resistance-coupled circuits, it is imperative to keep the capacity of this by-pass condenser as low as possible. Loudspeakers. HE loudspeaker has a very great influence upon quality reproduction, particularly with regard to the range of notes that may be properly produced. At the same time it is quite a .ommon oce rence for the _loudspeaker to be blamed for glaring shortcomings in the veceiver. Horn speakers are as a rule not strong on the lower notes, but if fed by u good receiver, it is surprising what good reproduction can be obtained from the larger type o. horn speaker, and in such a goou percentage of the lower frequencies will be heard, along witk the middle frequencies, whilst the reproduction of the high sounds will depend to a great extent upon what is vassed by the receiver. Cone speakers are being gradually improved, and some splendid models are now to be obtained. A good cone speaker gives good volume and clarity on the lor’ frequencies without unduly weakening the high. In large sizes the balanced-armature types give increased volume with facility for reproducing the higher frequencies, and in the same class may be placed the double-acting or "push-pull" type, in which there are two pairs of magnets acting in unison on bot: ides of the armature, The moving-coil or dynamic speaker is said to give higher quality than either of the above, but this type of |

-- oo speake. has so far not been imported into New Zealand in. any quantity. It costs a little more to run than ordinary types, but where mains are ayailable to supply current for the field magnets, the running cost would be Very small. Some of these speakers have permanent magnets, and thus dispense with current for excitation. Large-sized Hor-~ ¥F exponential and arthaphonic horns a great deal has been heard of late, and undoubtedly these in large sizes approach very near the present limit in reproduction of the whole musical scale. The chief drawback is their bulk, which really presents a serious problem in the average house. Sizes up to 9 and 1° feet in length are built in folded form anc furnish, a satisfactory solution, reproducing sufficiently lovy frequencies. A straight six-foot horn may be conveniently suspended near the ceiling with the unit

This is the most unobtrusive method of using such a horn. in the corner.