Radio Broadcast (May 1928-Apr 1929)

134 RADIO BROADCAST JULY, 1928 ^ (1) 4=8, Rp-10,000 ok b=100hcrmes ^ © -U'S, ftp10,000 uhms R=30,HOO o*im> Clough Amplifier L-^Uhenrits > 30 100 1000 10.000 FREQUENCY-CYCLES PER SECOND FIG. 3 can be done with the same design in the way of approaching perfection of the bass response. Using the same tube constants as were used for Curve 2, Fig. 2 and increasing the primary inductance we could plot curves for increasing values of inductance. It is plain that the bass response improves as we increase the primary inductance. Here again we run into the difficulty of secondary capacity for with each increase in the primary inductance we would have to put more turns on both the primary and the secondary. In practice we would find that after the primary had been brought to a value of about 100 henries, we could not increase it further without either decreasing the turns ratio or putting up with the deleterious effects of secondary capacity. Curve 7, Fig. 3, shows a measured characteristic of a commercial 3:1 transformer which represents about the limit along this line of procedure without reducing the turns ratio. It will be noted that in this case the amplification is yet not ideal at 30 cycles, while the selfcapacity of the secondary starts to impair the amplification at 8000 cycles. Under some conditions it is desirable that the amplification from 100 cycles, or thereabouts, down to 30 cycles should be greater than that on the flat portion to compensate losses in other parts of the circuit, the loud speaker, for example. In reproduction of phonograph records we could compensate the fact that the records are not cut up to full volume from 30 to 100 cycles, due to practical difficulties in recordcutting. By means of a new transformer circuit, an amplifier can be produced with either the ideal flat curve or a rising bass characteristic. WHAT THE NEW CIRCUIT IS THE circuit is shown in Fig. 4. Note that the plate current of the tube is carried by a resistor, R, and a condenser, C, connects to the primary. In this way the primary of the transformer carries no direct current, a feature of importance. The equation for this form of amplification involves a resonance which may be controlled or placed, by proper design, in any part of the frequency band. Naturally, the most desirable place for this resonance is in the bass frequencies. When the amplification is computed for this system, using the same transformer and tube which was used in the illustration, Curve 2, fig. 2, together with a resistance R = 30,000 ohms, the resultant curve is as shown in Curve 9, Fig. 3. it will be seen that this curve approaches the ideal very closely and with the same amount of material as in the designs illustrated by Curve 2, Fig. 2. Here the amplification is lower than was previously obtained with the same transformer, but this condition can be corrected by connecting the transformer as shown in Fig. 5. This auto-transformer connection increases the effective ratio to 4:1, using the same amount of copper windings that we have considered for a 3:1 ratio in ordinary transformer connections. The curve by this connection is shown in Fig. 6, Curve 1 1, where the amplification in the highfrequency portion is the same as we obtained with the same core and windings that were used in Curve 2, Fig. 2. The amplification at 30 cycles has been brought up to a par with the highfrequency amplification while using a transformer having a primary inductance of only 44.2 henries which may be compared with Curve 7, Fig. 3, which shows 27 per cent, decrease at 30 cycles in spite of the fact that it has a primary inductance of 100 henries. It is apparent that having a method of producing an ideal curve with a lower value of primary inductance than was necessary with the ordinary transformer connection to produce only approximately an ideal curve, we will need less secondary turns on a core of smaller cross-section. This mitigates largely the limitations caused by the secondary distributed capacity so that it is perfectly feas FIG. 4 FIG. 5 ible, where large amplifications are needed, to extend the ratio of the transformer greatly by the addition of secondary turns. It was previously mentioned that in the event we need amplification of the low frequencies more than the highs that it could be done by means of this transformer circuit. In order to illustrate this system as well as possible, the value of the primary inductance, 44.2 henries, was chosen in order to satisfy the conditions for a flat characteristic with the particular circuit constants used. This value of inductance is prescribed by the mathematics of the circuit and the equation for the new transformer indicates that if a value in excess of this is used in design that a rising bass characteristic may be produced. Thus let us assume that the primary (in Fig. 5,) is wound to an inductance of 80 henries. The circuit would then produce a characteristic such as is shown in Curve 12, Fig. 7. FINE RESPONSE CURVE WHILE several makes of this device will be on the market during the coming season, it may be interesting to the reader to see what can be accomplished in the way of faithful reproduction by means of measured curves on experimental laboratory designs. The solid Curve 13, Fig. 8, shows a transformer winding of a 3:1 ratio when operated out of a 226 tube with the customary voltages and with the conventional transformer connection. The solid line Curve 14, Fig. 8, shows the same transformer operated as shown in Fig. 5 with the same tube. It will be noted that this design provides reinforcement of the low frequencies and a slight reinforcement at the high frequencies in the vicinity of the cut off. Reducing the size of the windings produces the ideal curve rather than over accentuation of the bass frequencies as shown. A second illustration is shown in Curve 15, Fig. 9. This dotted curve was taken using a small 4:1 transformer in the usual connection; the full line, Curve 16, Fig. 9, with a transformer with 1000 10,000 FREQUENCY CYCLES PER SECOND V= 8; Rj,= 10.000 oht K = SO.OOOohma; L'SOhairu-a 1000 FREQUENCY CYCLES PER SECOND ABOVE — FIG. 6. BELOW — FIG. 7 the new connection. This curve does not go to 30 cycles, but many feel this low limit is not entirely necessary, in view of loud speakers and radio transmission shortcomings in the lowest octave of the musical register. The real point in the above curve is in its extreme departure from the ideal curve when operated in the usual connection, while with the new circuit, the curve resulting from proper design can be carried substantially flat down to any preassigned frequency. It has, no doubt, occurred to the reader that due to the drop in d.c. potential through the resistance, R, of Fig. 4, that it will be necessary to use a battery supply of greater voltage on the detector and first-stage amplifier tubes than is usual to supply the rated plate potential on the tube. This is no serious limitation of the design, because the detector is usually operated on 45 volts and the first audio stage at 90 volts while 180 volts are usually available from the B-power supply. In all the illustrations both resistors will permit operation on 180 volts and yet operate the plates of the tubes at their rated voltages. By supplying both stages from the 180-volt tap of the power unit, we save the voltage-divider necessary for B voltages of 45 and 90, as well as saving the by-pass condensers usually necessary across these taps. Such models have been prepared in the laboratory for operation with the a.c. tubes, producing an effective transformation ratio of 5:1 in the first stage and 4:1 in the second as against values of 3:1 and 4:1 respectively, which have been found to be the practical limit for a high grade audio-amplifier using the conventional circuit. This amplifier then has the following advantages: high quality reproduction, due to its flat characteristics from the very low frequencies to above 5000 cycles, secured at a lower cost than is now possible with standard transformer connections. This is because of the new scheme of connecting the apparatus in the circuit. There is one additional advantage which has not been mentioned here, but which is important. In any transformer in whose primary direct current flows, there is liable to be distortion introduced due to core saturation. Since there is no d.c. in this method of connecting the transformer into the circuit, such distortion cannot result. The writer hopes to present more of this side of his development later. 1000 FREQUENCY-CYCLES PER SECOND V=S. Rp=lOQOO ohms. 1 Oough connection FREQUENCY-CYCLES PEP SECOND ABOVE — FIG. 8. BELOW — FIG. 9