## Info

Delay voltage will be determined by the bias on the tube, and until the incoming signal produces more control voltage than this the control line will remain al ground potential If bias is set at 2 volts, AVC voltage will be zero until the control voltage exceeds 2 volts, and laen will be equal to control voltage minus the 2-volt delay.

Returning I-megohm resistor R2 to a source ol negative bias voltage will establish the AYC voltage at the level of this bias, but will also reduce the delay voltage by the same amount. You can use this to set a minimum 1-volt level on the AVC line, with 2-volt delay, or whatever you want, and get rid of cathode-resistor problems in the if chain; such an approach was used in the BC-779.

A more complex delayed-AYC circuit known as the "sinking diode' hookup is shown in Fig. 7. This one eliminates all differential distortion and allows the AVC action to be tailored to almost any circuit requirements, but is more complicated than inostr Heres how it works:

To start w ith, let s assume that D3 isn t in the circuit Lets also assign some values for Es, Ed, and consequently Rl. Es might well be 150 volts from a regulated supply, which Ed (the delay voltage) could easily be 5 volts. Then if R2 is set to be 470K, Rl will be 14 megohms«

Now, with no signal input, point A will be at ground potential, while the voltage at point B will be that developed by the voltage divider Rl and R2 in series from the 150-volt source, or about 4,89 volts positive.

When the voltage at point A drops to —5 because of incoming signal, the voltage across the divider becomes 155 (from H-150 to —5) and the voltage at point 13 becomes 5.02 positive, in comparison with that at point A. Since point A is —5, the resulting voltage at point B with respect to ground is 5.02 — 5, or 0.02 volt

As signal increases and the voltage at point A becomes say —25 volts, the voltage across the divider becomes 175 and the voltage at point B becomes 5.79 volts positive to that at point A, or —19,21 volts to ground.

When the voltage at A goes to —50 volts, that at B will be —43.5. As the voltage at A becomes more negative, the percentage difference between B and A becomes less.

In this circuit, the delay voitage may be set to any desired value as shown in the illustration. The larger Es is made, the less percentage reduction of control voltage will be introduced by the R1-R2 voltage divider, However, if the ratio is made too high you may have difficulty locating suitable resistors for Rl as few distributors stock anything larger than about 22 megohms,

So far, we have been ignoring the action of D3 and of CI. D3 prevents the AVC line from ever going positive, by conducting and clamping the line to ground whenever its plate is more positive than its cathode. If you would [irefer to have the line clamped at some value of standing bias such as —1 or —S volts, simpiy returning the cathode of D3 to such a voltage instead of directly to ground will accomplish this.

CI is the filtering capacitor, and its value will determine at least the charging time constant of the system. The other component of this time-constant network is R2. The value of CI should be chosen after R2 is selected, so that the product of CI in microfarads and R2 in megohms is equal to the desired charing time constant (0.004 if the 4-miIlisecond recommendation is to be followed). To get a 4-ms charging time with our previous example, CI should be 0.0087 |jf. This is not a standard value; you could use an (10082 in parallel with a 500 pfn or probably just a 0.0082 or an 0,01 alone with little real error.

As shown, the discharge time will be longer than the charge time, since the discharge path includes 1 megohm in series with R2 (the high resistance of R1 prevents it from having much effect in rapid discharging of CI). If a 0:01 [if unit were used, the discharge time constant would be 0,0147 second, or 14.7 ms, too slow for CW and too fast for speech.

Discharge time can readily be lengthened by adding a diode in series with the capacitor as shown in rig. so that no discharge path is provided except by the leakage across the capacitor (this can't be done if divider resistors are used to give partial AVC to some stages as discussed earlier since they will allow more rapid discharge)- However, the only practical way to shorten discharge time in this circuit is to modify component values in a muIti-way compromise. The simplest way to make the compromise is to make CI small enough that discharge time falls to the desired value; in our example this would be 0.004 seconds/1.-47 megohms, or 0.0027 [jf Charge time would now be only 0,0027 x 0,47 or L28 milliseconds. This, however, is no particular disadvantage,

At this stage, the circuit can be converted to fast-slow switching as shown in Fig, 9 by adding a diode and paralleling it with a switch. With the small value used for CL a 0.5-second discharge time constant can be obtained by using a ISo-megohm resistor at Rd in Fig. 9—and this amount of resistance is undoubtedly present in switch and circuit leakage. In fact, you may have trouble getting that ong a discharge time with no physical resistor at all for Rd. This, incidentally, is the basic "hang" AVC' approach.

Before getting into how to modify only the though, let's look at one more AVC circuit.

This one employs dc amplification of the control voltage after detection, for a wider range of control than would be possible otherwise. The circuit appears in Fig. 10,

In this one, the delay voltage is set by choice of resistor RL in the cathode circuit of VI. So long as the voltage developed across the 1-mcgohm resistor in Vl's grid circuit is more positive than cutoff voltage for VI, the cathode voltage of V! will be positive. With this voltage positive, diode D1 will be reverse-biassed and cannot conduct, so the AVC voltage will remain at zero.

When signal is applied and D3 rectifies itT a negative voltage is developed across the 1-megohm resistor. When this negative voltage approaches the eutofl value for VI, the cathode voltage ol the tube will move toward the negative-supply level, and when VI cuts off the cathode will assume the negative level of the negative supply.

This forward-biases D2? which conducts and allows the negative voltage to appear on the AVC line.

If signal level is only enough to supply say a 10-volt negative value across D3r the AVC voltage can be made much greater by proper choice of tube type for VI, of negative supply voltage, and of RL. With a 150-voIt negative supply and a tube which will cut off at —10 volts between grid and cathode, the AVC line can be made to go to virtually I he full — 150 of the supply.

Note that D2t D3, and VI can all be a single duo-diode/triode tube such as the 6AT6 or 6AV6, while detector diode D1 can be any semiconductor type. This means that, though the circuit appears more complex than most, it actually requires only a few more components than does the simplest of AVC arrangements. The negative voltage supply can be obtained from the regular power supply through a capacitor and shunt diode as shown in Fig. 11.

Now let's look at the time-constant or reaction time situation with an eye to possible modification of your own receiver. Fig, 12 shows a typical AVC setup, with the actual time constants of your own AVC system, Fig. 10. DC-amplified delayed AVC. omv out

tontrol-voltage takeoil and rectifiers indicated merely by a block. You can compare this to the schematic of your own receiver to determine which components correspond on your own schematic.

You can see that our earlier discussions assumed that reaction time was determined entirely by R4 and C4; tins isn't quite true, i he formulas shown in the figure include the effects of all resistances and capacitances in the circuit. If more stages are under control than the three shown here, their decoupling networks can be included by adding R7+C7 in the same way that the underlined R6, C6 are included here (for two-stage control, omit all underlined components in the formulas).

While R4 and C4 still predominate in the control of reaction time, with the typical values indicated on the schematic the R4-C4 time constant was lengthened by some 2.4 milliseconds on charge and by 42,65 milliseconds on discharge.

To reduce these other effects, \ou can replace a 1 decoupling resistors (R57 R6, and filter resistor Rl) with RI choices, and reduce the associated bypass capacitors to somewhere near 100 pfd, A shorter discharge time constant can be achieved by adding a parallel resistor (Rp) as shown in dotted lines, but this resistor together with R4 will then form a voltage divider which will keep full AVC from being applied to the line. Reducing the value of C4 will shorten both the charge and discharge times; you may find it possible to reduce the value of CAy add Rp, and modify the value to R4 to equalize charge and discharge times. Then adding a diode as shown in dotted lines gives you an option of short or long discharge time; with the diode in the circuit, discharge time will be long.

In making these modifications, refer to the chart of "normal" AVC voltage levels you prepared before starting, arid be certain that you end up with the same voltage on the AVC line for the same signal input that you had in the beginning* (Of course, if you had overload problems too you might want to use a little ess or a little more, but if the control was satisfactory and all you want to change is the reaction time then you must be certain that you get the same control voltages out a iter making circuit changes.) These measurements should be made only with steady signal applied, because even the 11-megohm input resistance of a VTVM will drag off any charge on C4 if it is made small.

The same considerations, of course, also apply to the substitution of any of these AVC circuits for that existing in your receiver. The final voltages applied to the AVC line of the receiver must bo a fairly close match to the w original voltages if undistorted, completely controlled output is to be achieved»

### Not vet mentioned has been the use ol the m

AVC voltage to control other I unctions such as anl bias-setting or squelch circuits. If the AVC is to be so used, it's a good idea to include some form of dc amplication such as the circuit of Fig, 10, with an additional diode connected in the same manner as D2 for each of the additional functions. This will prevent undesirable interaction between the unctions, and will prevent them from having any effect upon the AVC- circuit itself. For economy reasons, such isolation is not usually provided in factory-built gear.

From this point on, AVC design becomes more of a detailed engineering study than a general survey subject. If you care to pursue it furdier, a good starting point is the "Radio-tron Designer's Handbook" available from Radio Bookshop. Look in Chapter 27, An additional reference is k. R, Shirley's book "Radio Receiver Design/' Part 2, Chapter 12, if you can Iocatc a copy—it was published in London in 1943.

But for a general understanding of the subject, additional reading shouldn't be required. And now, at least, you should know what that little knob on the panel does when you turn it! 