Radio Broadcast (May 1928-Apr 1929)

An A, C. Band-Pass Screen-Qrid Receiver^ The Master aHi*Q 29n By WILLIAM E. BARTON Hammarlund-Roberts, Inc. FOR as many moons as the oldest radio editor can remember, writers and engineers have deplored the fact that one cannot have both selectivity and fidelity— and yet nothing much has been done about it. On the one hand, we have receivers which may get all the audio notes in proper proportion, but which, so far as selectivity is concerned, are as broad as the proverbial barn door. On the other hand, we have receivers which, to use an advertising phrase, are as sharp as a knife blade, but which — and advertising writers say nothing about this fact — get few notes above 3000 cycles. And there you are. You may pick and choose, but you can't have your loaf and eat it. A radio receiver first of all must select the program you want to listen to, and then must amplify the audio tones to the level desired, whether it be for head-phone reception or full loud-speaker volume, usually the latter. But if, in the process of selecting, half of the audio tones you want to hear are lost, no amount of audio amplification will bring them back again in their proper proportion. There will be plenty of "lows" to be sure, lots of the bass drum — unharmonious instrument — but few of the humanlike notes of the violin. And so in a congested district where broadcasting stations are placed far enough apart not to bother each other, but close enough together to prevent any "getting out," people built receivers which were not very selective, and so the quality was good. Fans outside the large cities, however, had a different problem. Surrounded by stations, all over 100 miles away, the receiver had to have selectivity enough to cut out a geographically near-by station which might be poor in quality in order to receive a good station only 10 kc. away from the nearer station. No wonder side bands were clipped. No wonder few notes above 3000 cycles were received. Up until about a year ago such a dilemma as this existed in every constructor's mind — should he build a selective receiver, one that would get out, or should he be satisfied with local broadcasting and build a broadly tuned receiver? In most cases a compromise was difficult to effect. Then, at a meeting of the Institute of Radio Engineers, Dr. F. K. Vreeland gave his paper of band-pass tuning which promised not only more selectivity but greater fidelity of response as well. This started many engineers thinking and remembering their text books and wondering why they hadn't thought of Dr. Vreeland's scheme themselves. For the truth must be told, Dr. Vreeland called to mind the old system of tuning two circuits to the same frequency and then coupling them closely enough together that THE HI-SI IN. A COMBINATION PHOHOGRAPH-RADIO CABINET CT'HE theoretical background of the IQ2Q ■*■ model of the Hi-Q receiver, as outlined in October Radio Broadcast, is amplified in this article Jrom the Hi-Q organisation. In the Laboratory, as tested under average conditions, the receiver seemed to have considerable r.f. amplification, the selectivity was good, and the fidelity of response was excellent. The completeness with which the r.f. circuits are filtered probably has much to do with the stability, and simplicity of operation. There are no trick adjustments. There is one interesting point which is not mentioned in this article, and about which we hope to present data soon. This is the fact that the shape of the response curve depends upon the frequency — that is, the curve will be one thing at 500 kc, something else at 1 500 kc. — when the circuits are coupled by inductance, capacity, or mutual inductance. Just what this effect is, in the present receiver, was not apparent in the Laboratory. At the top and bottom of the broadcast-frequency spectrum good response was obtained. Perhaps the receiver had been adjusted somewhere in the middle of the band so that at the two ends it still had a band-pass circuit characterestic. — The Editor. the response curve no longer looked like a steep mountain, but like twin peaks side by side. It no longer had gently sloping sides down at the interference-frequency region, but a sharp cut-off. And what good is such a curve, you may ask? Why have a broad top and steep sides. The answer is the reply to the prayers of engineers and those who want more selectivity and more fidelity of response. At the top of the curve — where the audio tones are — a broad flat plateau exists, at the sides of the curve — where interfering stations are — there is a steep precipitous drop in response. L1 <*S9(£ BAND-PASS FILTER CURVES ET us look at Fig. 1 which represents engineering data on the Hammarlund Roberts "Hi-Q 29," a receiver making use of the timehonored method of obtaining the flat-topped response curve described above. In the case of the single coil and condenser tuned to 1000 kilocycles, the top of the curve is peaked markedly, and if, as is usual, a little regeneration creeps into the amplifier, this peak becomes even more marked. The dotted line represents the response or resonance curve of a circuit with a resistance of about 5 ohms at 1000 kc. a lowloss circuit. At 2000 cycles the curve is beginning to droop and at 5000 cycles the response has been reduced to only 60 per cent, of the response at, say, 100 cycles. Now look at the full-line curve which represents the band-pass tuning arrangement. This is laboratory data on a single r.f. stage of the "Hi-Q 29" receiver. At 2000 cycles the curve has not even begun to drop, and at 5000 cycles the loss is only 10 per cent. At the bottom of the curves in Fig. 1 other interesting things may be noted. The dotted curve shows a response at 20 kc. off resonance of 20 per cent. In the case of the Hi-Q stage, however, the loss is 90 per cent, which, in a twostage affair, where the loss is squared, gives a response of 1 per cent, instead of 4 per cent, for the simpler circuit. So far so good, but how is it possible for a receiver to be selective and still have good fidelity of response? Fig. 2 is the diagram of a single transformer-coupled stage of r.f. amplification which has the proper electrical characteristics to give a curve like that of Fig. 1 (dotted-line curve). The less the resistance in this circuit, the greater the amplification, and the greater the loss to the high audio tones. Now let us contrast this circuit with the more complicated one in Fig. 3, which is the arrangement used in the Master "Hi-Q 29" receiver. Here, again, we have a transformer-coupled stage of r.f. amplification, but both the primary and secondary windings are tuned — and they are tuned to the same frequency. In fact the primary and secondary coils 117