Radio Broadcast (May 1928-Apr 1929)

30 RADIO BROADCAST MAY, 1928 Experimenters who have The Screen-Grid installed a single screen-grid Tube: Selectivity tube ahead of a detector, regenerative or not, have been disappointed at the apparent loss in selectivity of the system, although the gain increases. Let us suppose that the resonance curve of a singlestage of r.f. amplification using a 201-A tube is as shown in Fig. 1 . Now let us use a screen-grid tube in which the gain may be twice as great. In other words every point of the resonance curve of Fig. 1 is boosted twice as high with the result as shown in B. In any local area, stations are separated by 50 kc, so that on Curve A the incoming signals from a station 50 kc. off resonance are below the line which represents the arbitrarily chosen signal magnitude beyond which interference occurs. Now look at Curve B. Here the 50-kc. signal is up out of the area in which no interference occurs, and is heard in the background of the station to which the receiver is tuned. Let us call the absolute selectivity factor, the ratio between the height of the curve at resonance to the height at 50 kc. off resonance. This factor for curve A is 50 -s 2.4, or 2.1, and for the curve B is 100 4.8, or 2.1 exactly the same. Then, if the apparent selectivity is defined as the number of kilocycles off resonance a signal must be before it is reduced to the non-interfering region, we see that for the 201-A example it is 50 kc. while for the Curve B where the gain is twice as great the frequency is 70 kc. These figures depend, of course, upon the shape of the resonance curve. Each additional stage of 201-A r.f. amplification increases the gain of the receiver and increases the selectivity, but as Professor Hazeltine has shown in his recent patent, No. 1, 648, 808 — (Nov. 8, 1927) — by properly designing the interstage transformers in such amplifiers, an increase in selectivity of 50 per cent, can be obtained at a loss of only 20 per cent, in amplification. In other words, in a system using successive tuned stages, relying upon resonance curves for selectivity, there must always be a compromise between gain and selectivity. Increasing the gain without increasing the absolute selectivity at the same time, reduces the apparent selectivity. Although the signal to which the system is tuned will be louder, so will the background produced by other stations. Adding a single stage of screen-grid tube r.f. amplification under conditions which produce maximum gain from that tube, decreases the apparent selectivity too much. Adding an extra stage, from which we secure maximum amplification, will boost the sensitivity faster than the selectivity is increased, and we are no better off than before. In other words, if three stages of 201-A amplification are needed to secure sufficient selectivity, more than three stages will be necessary when using screen-grid tubes. This seems to indicate that these tubes will not decrease the number of effective tubes in a receiver, but will actually increase that number — so long as we use the t.r.f. system of tuning. The answer, if this reasoning is correct, is to use a different system of tuning. The superheterodyne is one solution; perhaps the Vreeland system described recently in the I.R.E. Proceedings is another. This is a system designed with the avowed intention of making a response curve of a receiver flat on top and very steep on the sides. It is done by using two closely coupled circuits so that the resonance hump is not a single sharp peak, but is composed of two peaks with a dip between. The Laboratory hopes to present some quantitive data on this system 110 100 80 70 60 50 40 30 20 10 rem Grid \ B 4 — 1 \ ,w cvei auuve — nch Signals Interfere , Non-Interering Regior '///A , / / A /W § w 200 150 100 50 0 50 100 150 200 KC. OFF RESONANCE FIG. I soon, and in the meantime welcomes comments on the problem outlined above. Many times per month we Loud hear about some high official Speakers in this or that radio company who indulges in a bit of forecasting, usually about the future of radio. Among other trends, to believe these officials, is that toward greater power which will be handled in the future radio's amplifier, one executive going so far as to state that our sets of the future would have amplifiers turning into our loud speakers at least ten times as much energy as they do at the present time. Judging from what has happened during the past few years, one cannot doubt it, for in the good old days we were satisfied with the output of a 199, then we needed a 201-A to deliver sufficient power, then the 112, then the 171 and 210 tubes, and now the 250 type. Each of these tubes delivers considerably more power than its predecessor, and to jump from a 5-watt tube to a 50 watter represents ten times as much power output, that is, 10 tu difference. We are not convinced. Several years ago we listened to a pretty fair program coming from a speaker that was 10 tu better than any speaker now generally available. That is, it required 10 tu less power to get a good healthy sound from it. This represents the difference between a power output of 600 milliwatts ( 1 7 i-a), and 60 milli Watch for the A. C. "Lab" Receiver TJUGH KNOIVLES' article describing the con±1 structiori of the popular R. B. "Lab" circuit for a. c. operation will appear in Radio Broadcast for June. The completed receiver has been thoroughly tested in Radio Broadcast Laboratory and has proved very satisfactory? indeed. Readers will recall the complete experimental article :n our April number describing tests in the Laboratory on the design of this circuit. The June article describes completely the construction of an a. c. four-tube "Lab" circuit receiver. Some of its interesting features are: Excellent efficiency for four tubes: a. c. operation, an extra socket and control switch enabling quick transfer of the audio system to phonograph pick-up or short-wave tuner unit, and an interesting volume control. — The Editor watts (20 1 -a). In other words, when loud speaker manufacturers are able to build a loud speaker with a good characteristic that is 10 tu better than our present loud speakers, we can all go back to our old 201-A tube, our B batteries, and start all over again. What are the prospects? We have heard that the Western Electric Company builds a loud speaker much more efficient than the 540-AW, for sale in England only. It does not have the same tonal range as the 540-AW and is cited only as an example of an efficient loud speaker. Incidentally, present trends in loud speaker design are toward moving-coil affairs, elecrodynamic units, such as the Magnavox and the Jensen cones, and the Vitaphone, which is a moving-coil unit, coupled acoustically to an exponential horn. All of these have very fine frequency characteristics, with honors at this writing in favor of the Jensen. The Jensen is made in California and is now generally available, we understand. We have seen curves, above reproach, which show the Jensen to have a quite flat response curve from about 60 to 6000 cycles. With a large baffle-board, such as obtained by inserting the loud speaker in the walls of a home, the lower limit of response can be pushed down to about 35 cycles. A six-foot square baffle about one and a half inch thick, is however, very satisfactory. If this Jensen unit, for example, could be made io tu more efficient, that is to say, deliver the same output with one-tenth the input, the result would be distinctly worth achieving. It should not be forgotten, however, that aside from the question of overall "audio" efficiency in this unit, we must supply the coil with 60 mils, at 100 volts — 6 watts of power. The Magnavox, too, is a fine product, but seems to suffer from excess filters which appear to be necessary to prevent too much a.c. hum from getting into it, to cut down the high-pitched heterodyne whistles, and other noises. In other words, a Magnavox loud speaker unit without the removable devices now employed to make up for our present poorly filtered a.c. receivers, or the present unfortunate situation in the ether, is a fine unit. We understand from an unimpeachable authority that the Stromberg-Carlson engineers built into stock receiver models a few years ago, audio amplifiers so good that nearly all of them came back. Criticism arose from the fact that these receivers seemed more noisy than sets of other manufacturers; more static came in, and "whistles" were more evident. It was a simple matter at the factory to put filters in the amplifier system, cutting down on the two ends of the frequency spectrum. Then the receivers stayed sold. It is a fact that many hundreds of these receivers have been sold to Bell Laboratories engineers and their friends. As soon as they are received out comes the soldering iron, off comes the filters, and out of the loud speaker come the low and high frequencies that are so essential to quality. In the article on the fourcAn Error in tube "Lab" receiver in the Coil Dimensions April Radio Broadcast, specifications were given in Fig. 2 for coil dimensions. Coils Li and L2 were shown to consist of 90 turns, No. 24 sec, on a form 2.5" in diameter. The correct designation should have been 90 turns No. 24 sec. on a 2.0" diameter form. If the reader already has a 2.5" coil form, he may use 66 turns of No. 24 s.c.c. wire to cover the same range. — Keith Henney.