Radio Broadcast (May 1928-Apr 1929)

Accuracy of the "Slide-back" Voltmeter MANY engineers take the vacuum-tube voltmeter for granted, i.e., they seldom stop to consider its limitations and inaccuracies. Many times within the past two years we have seen described in more or less technical papers the "slide-back" voltmeter with a direct or implied statement that it is an infallible device for measuring voltage. The circuit diagram for a "slide-back" voltmeter is given in Fig. i. It consists of a vacuum tube biased so that the plate current is quite small and so that a.c. input voltages change this steady C bias, thereby changing the plate current. The plate current is then brought back to its original value by changing the steady C bias by means of the potentiometer. The difference between the two values of bias (with and without a.c. input) is the peak value of the input voltage. The negative halves of these cycles drive the grid more negative, but, since the tube is already overbiased, the plate current changes but little. The positive halves of the cycles, however, reduce the steady negative C bias, and the plate current increases. Then the potentiometer is varied and the steady C bias is "slid-back" until the same platecurrent reading is obtained. To test the accuracy of the instrument we used a Weston model 301 milliammeter with a full-scale reading of 1500 microamperes. With 45 volts on the plate and a C bias of minus 4 volts, the plate current was 200 microamperes steady current was balanced out of the microammeter by using the A battery voltage or by means of a "bucking battery" (B) as shown in dotted lines in Fig. !. Now, when an a.c. voltage of 2.05 peak was placed on the input, the plate current increased, and we "slid-back" the potentiometer until the C bias meter read an increase of 0.35 volts. When an input of 4.0 volts was used a net change of 1.6 volts in the C bias was required before the current plate was the same. Clearly the method fell down. A change of C bias due to a.c. voltages could not be balanced out by an equal change of d.c. voltage. In the next test we put .33 milliamperes through the meter by means of the battery B, in Fig. 1, and the C bias on the tube was adjusted so that no change took place in the deflection of the plate-current meter when the plate voltage was turned on or off — in other words, we placed sufficient bias on the tube so that the plate current was zero. This point cannot be determined exactly, of course, but if a !trays: Laboratory This the meter has an initial reading changes in this deflection are noted easily. Now when an a.c. input voltage was placed on the tube, the C bias changed, the plate current changed and a deflection was noted. The steady bias was then increased so that turning on or off the plate battery to the tube made no change in the reading of the plate-current meter. When an input peak a.c. voltage of 10 was applied to the tube a change of 9.5 volts steady bias was necessary to reduce the plate current to zero; other readings are noted in Table 1. TABLE I Steady current through 500 microammeter=331 microamperes. Steady bias to cause no deflection on meter= 7.7 volts when Ep=45 volts input a.c. d.c. volts peak volts to reduce Ip to zero accuracy 4.05 3.6 89% 6.1 5.8 95 8.2 7.5 92 10.1 9.5 94 14.5 14.3 99 16.3 15.6 96 18.3 17.5 96 1 — WwV^W-i -Tooo.1.1 10'000 ohms FIG. I. SLIDE-BACK VOLTMETER Substituting a more sensitive meter, say a Westinghouse 500 microampere meter, or a Weston zero-center galvanometer with a sensitivity of 60 microamperes per division, increased the accuracy somewhat. But even with the 1 500 microampere meter we could balance out an a.c. voltage with a d.c. voltage with an accuracy of about 90 per cent. Larger input voltages could be read more accurately. When used in this manner the "slide-back" voltmeter is accurate enough for all ordinary measurements. The device is inaccurate when operated, as is often done, so that a fairly large steady current is obtained in the plate circuit, and the bias is so adjusted after an input is applied that the plate current returns to this value. The nearer one can get to the actual zero plate current point, the more accurate the instrument as a whole becomes. FOR A long time we Hum in the "Lab" threatened to throw Circuit Receiver out OUT B-SUpply unit and build a new one. With the four-tube "Lab." circuit receiver (August Radio Broadcast) considerable hum appeared in the loud speaker in spite of rather thorough filtering in the B supply itself. We began to wonder where the noise came from; was it inductive pick-up from the power line running near the audio-transformers, or was it picked up in the first or secondaudio, detector or r.f. tube? The audio amplifier of the receiver was perfectly quiet, as evidenced by shorting its input through a 10,000-ohm resistor. Running the detector from a 45-volt B battery helped a bit, but the hum was still too loud. Running the r.f. tube from a 90-volt B battery killed the hum completely. Larger filter condensers across the 90-volt B-supply lead did no good. What could be the trouble?" Let us look at the r.f.-tube circuit in Fig. 2. Notice that the 90-volt lead from the power-supply unit — which has some a.c. in it no matter how well it is filtered — is connected directly to the detector input coil. This is a very high-impedance circuit, equivalent to Fig. 3, and any a.c. current flowing will build up a large voltage and subsequently will be amplified by this detector and audio tubes. The first experiment was to wind a primary coil about the detector input coil, as in Fig. 4. The noise dropped out. The solution was then simple: isolate the primary winding of the coil from the secondary, as in Fig. 5, and ground the lower end of the detector coil. Now the power-frequency noise is effectively grounded as far as the detector input goes, and, therefore, no hum gets on to the tube's grid. Figure 5, then, gives the circuit diagram of an r.f. amplifier and detector for the "Lab." Circuit which will iron out a.c. hum entering the receiver from the B-supply unit via the r.f.amplifier plate circuit. In the course of the experiment leading to the elimination of the hum from the circuit, the leads to the regeneration condenser were reversed. Considerable difficulty was experienced in "holding-down" the circuit, and it was impossible to neutralize the amplifier completely without placing a shield between the regeneration condenser and the detector tuning condenser. When, however, the regeneration condenser was connected correctly, that is, stator to the detector plate, all difficulty disappeared, and a high-gain stable amplifier resulted. Readers who have trouble with the circuit, evidenced by the detector or amplifier oscillating continuously, might try reversing the leads to this small condenser. FIG. 2. ORIGINAL "LAB CIRCUIT 97