VINTAGE RADIO

By JOHN HILL

The importance of grid bias

Correct grid bias is vital if valve radio This caused some concern because I knew, even back then, that battery receivers are to function properly. Here's a valves didn't normally run hot. The circuit was checked and every- rundown on how it works and what to look thing was in order except for the two for when restoring vintage receivers. "unnecessary" components. When a milliamp meter was placed in series with the 90V "B" battery it Many years ago I built a little 2- ative usually went straight to chassis. indicated a drain of 20mA. That is valve battery receiver called "Tom Well, it had on my previous home perhaps more "B" battery current than Thumb". It was an old "Radio and built 1-valve receivers. a valve portable would draw while Hobbies" project that incorporated a In my "wisdom", I chose to leave driving a loudspeaker. Could it be that 1T4 regenerative detector, followed out this part of the circuit and con- those unnecessary components had by a 3S4 audio which drove nected "B" negative directly to chas- something to do with the problem? a pair of high impedance headphones. sis due, in part, to the fact that neither After the 1500Q and its It was built from the circuit diagram a 150052 resistor or a 100' electrolytic accompanying 100 capacitor were only, without instructions, using a were on hand at the time. Besides, added to the receiver, three changes small output transformer and low im- their inclusion seemed so unneces- were immediately apparent: (1) the B pedance phones instead of the speci- sary. battery current dropped to less than fied 20000 headphones. The set was built and it worked 5mA; (2) the output valve operated at There was one part of the circuit reasonably well. Just as I thought — the a much lower temperature; and (3) a that, at the time, made no sense to me extra components were put into the degree of audio (originally at all. Why have a parallel connected circuit to confuse novice receiver assumed to be normal for such a sim- 15000 resistor and 10µF electrolytic builders. ple set) vanished. The mystery com- from "B" battery negative to chassis? However, while listening to my new ponents were not as unnecessary as (see Fig.1.). What could such an ar- creation it was noticed that the 3S4 originally thought! rangement possibly do when "B" neg- output valve was decidedly warm. At the time, my lack of knowledge

This old Eveready "C" battery contains three size "D" cells wired in series. The close-up view (above) shows the battery connections. The terminals, from left to right, are OV, -1.5V, -3V and -4.5V.

82 SILICON CHIP regarding basic theory prevented me from knowing what the resistor/ca- pacitor combination actually did. Ac- INTERSTAGE cordingly, I made a very bad error of COUPLING judgement by leaving them out — but FROM TRANSFORMER DETECTOR learnt a good lesson by doing so. SPEAKER Negative grid bias BIAS OR • As some readers would he aware, 'C' BATTERY 10 the reason for the resistor was to cre- 1111Prb- ate a negative bias voltage for the con- trol grid of the output valve. This Fig.1: the back-bias circuit of the Fig.2: many battery valve sets would allow the valve to work under "Tom Thumb" used used a "C" battery to apply the conditions for which it was de- a 1500Q resistor in parallel with a negative bias voltage to the signed. Without grid bias, the valve 10µF capacitor. control grids. overheated, consumed large amounts of "B" battery current and, most im- portant, created considerable distor- (C• 6.3V tion. Correct grid bias is important! This matter of grid bias raises two broad questions. Why do valves re- 240VAC quire a negative potential on the con- trol grid or, more accurately, between FIELD grid and cathode. And, secondly, how HT+ are negative volts obtained when the main supply voltage — from the "B" battery — is positive? These are good BIAS questions and I will try to answer RESISTOR NEGATIVE them as best I can. BIAS VOLTAGE First, why is the grid voltage neces- sary? If one spends time looking Fig.3: a typical back-bias circuit as used in many AC- powered receivers. through valve data books, a lot of reference is made to "characteristic curves". In simple terms, a characteris- tic curve is a line plotted on a graph which shows the relationship between is always measured in relation to the Some special high-mu (high gain) type changes in grid voltage and changes cathode. Almost all valves need some valves are designed to function with in plate current. Each type of valve negative bias on their control grids in minimum plate current without nega- has its own set of characteristic curves order to function properly. Ignoring tive bias. They are designed for class for a given range of plate voltages. this fact, as I did with the previously B operation in push-pull output cir- A characteristic plot is not uniform: mentioned "Tom Thumb", can lead to cuits. it has a curved section at the bottom all sorts of problems. (the "toe") and another curved sec- Of course, there are exceptions. Battery bias tion at the top (the "shoulder"). Be- This brings us to the second point tween these two sections is a substan- in this discussion; the exact circuit tially "straight" section. mechanism by which an appropriate If a valve is correctly biased (typi- negative voltage can be applied to the cally at the midpoint), the voltages on grid(s). There are a variety of arrange- the grid will be confined to the straight ments but one point is paramount — section of the curve and the valve will the basic requirement is to apply the operate with minimal distortion. It required negative voltage to the grid will also draw the specified amount with respect to the cathode. of plate current. However, if the valve Valves have always needed to be has insufficient bias, the plate current correctly biased but, back in the dim will increase and there will be distor- past, in the era of battery receivers in tion. the 1920s, the need for correct bias Conversely, if it has too much bias, was not always fully understood. If it there will be insufficient plate current was used at all, the usual procedure and this will also give rise to distor- was to add a separate bias battery (typi- tion. This distortion is particularly cally 4.5V) to the circuit. This battery serious in the case of output valves. This battery-model late 1930s Radiola was referred to as the "C" battery and Note that the grid potential — in fact uses a 4.5V tapped "C" battery for its it was connected as shown in Fig.2. the potential of all valve electrodes — grid bias requirements. Battery bias is referred to as "fixed

MARCH 1997 83 Checking the value of Looking at this another way, "B" bias is part of battery negative is the most negative any radio restoration. point in the system, which means that Shown here is a the chassis is positive by the voltage restored early 1930s 6- across the resistor. This in turn means valve superhet receiver made by Eclipse Radio. that the valve filament (cathode) must he positive with respect to the grid by the same amount. The amount of bias produced by a back-bias circuit is proportional to the total high tension current—not neces- sarily the current of the valve or valves being biased. Negative voltages pro- duced by back- are produced at the expense of "B" battery voltage. In other words, if a receiver has a 90V "B" battery and the hack-bias resistor supplies a 5V negative bias, then the effective "B" battery voltage is reduced to 85V. You don't get something for nothing! Back biasing is also used in some AC-type receivers. This involves add- ing a resistor in the high tension cen- tre-tap lead of the power transformer (see Fig.3). Once again, the negative bias voltage is produced at the ex- pense of the overall high tension volt- age. Any form of grid bias that does not use a battery requires a resistor to produce the bias voltage. It is com- mon practice to place a capacitor across the bias resistor to suppress bias" because the bias voltage remains potential with respect to this point. any unwanted signals. constant regardless of the slowly diminishing "B" voltage, which is not Back bias Self-bias the ideal situation. There is virtually However, not all battery receivers Another form of biasing often used no current drain from a "C" battery. Its used a "C" battery. Many, such as the in AC-powered sets is cathode bias- sole purpose is to supply the control previously mentioned Tom Thumb, ing, sometimes referred to as self-bi- grid with a negative potential. have a hack-bias arrangement where- asing (Fig.4). Cathode bias makes use It is easy to understand battery bias- by the negative voltage is produced by of the cathode current through the ing — one only has to look at the cir- the voltage across a resistor (Fig.1). In valve. The cathode current is the sum cuit diagram of Fig.2 to see where the this hack-bias circuit, the voltage of the plate and screen currents and if negative volts come from. If "B" nega- across the resistor is negative at the a resistor is placed between the cath- tive and "C" positive are at the com- grid end with respect to chassis and ode and chassis, then the cathode cur- mon point, then the negative end of can he used as a source of bias for one rent flows through this resistor. the "C" battery must be at a negative or more grids. This current flow through the cath-

IF AMPLIFIE FIRST 1ST IF FROM AUDIO TRANSFORMER DETECTOR - STAGE GRID 10M RETURN RESISTOR • BIAS Fig.6: this grid leak bias system RESISTOR relies on a small amount of grid current which flows through a Fig.4: a typical self-bias circuit. 101%1 resistor. The bias voltage is applied to the grid via the grid return Fig.5: this variable bias circuit resistor. is used as a volume control.

84 SiLicoN CI HP Wirewound potentiometers were used as variable cathode bias resistors in many old radio receivers. When used in conjunction with variable-mu valves, variable cathode bias was an effective volume control. ode resistor produces a voltage across valves; ie, the radio frequency (RF) it and so a positive potential is devel- amplifier, the mixer and the interme- oped at the cathode with respect to diate frequency (IF) amplifier. chassis. Because the grid is normally These valves have variable mu-char- connected to chassis via a resistor, it acteristics and their amplification fac- follows that the grid must be negative tor is controlled by changes in grid with respect to the cathode. bias. In the case of the AVC circuit, the The term "self-bias" is used here bias produced is proportional to the because the bias voltage is propor- signal strength. As the signal strength tional to the total current flow through becomes greater, so too does the bias the valve being biased. voltage which automatically restricts of various bias methods but, as far as Another form of cathode biasing the amplification provided by the RF vintage receivers are concerned, it involves using a variable resistor in- valves. Matters little. What is important is stead of a fixed resistor. Many receiv- Basically, AGC is just another form that the bias circuits are working as ers from the 1930s used such a system of grid biasing and is a variable bias. the designer intended but that is not as a volume control, with the potent- Still another form of grid biasing always the case. iometer in the cathode circuit of a can be found in some first audio stages One problem with old carbon resis- variable-mu IF amplifier valve (see and is referred to as grid leak bias. tors is that they often go high with Fig.5). This involves connecting a high-value age. When a bias resistor goes high so In its simplest form, only one termi- resistor between the grid and chassis too does the bias voltage and that nal and the moving arm connections (Fig.6). It is normal for a small amount means that the grid can swing outside are used. Connecting the other poten- of grid current to flow, the exact the straight part of the characteristic tiometer terminal to the aerial is a amount depending on several factors. curve. This could cause the valve to trick to improve the range of control These include the type of valve and operate near-cut off in extreme cir- but has nothing to do with the bias its operating conditions. cumstances. When restoring a vintage function. In practical terms, this bias system radio receiver, it is therefore impor- is suitable for use with high-mu tant to check the bias resistors and Automatic volume control valves handling low-level input sig- replace them where necessary. The If one goes probing around with a nals Running the weak grid current bypass capacitors should be checked voltmeter underneath the chassis, it through the 10MS2 resistor produces as well. soon becomes apparent that there are the desired bias. The amount of bias is Finally, note that when checking many points in the circuit that will small but can be sufficient to set the bias voltages, it is advisable to meas- register negative voltages. Some of operating point on a straight portion ure the voltage across the bias resistor these potentials vary in magnitude of the characteristic curve. Provided itself; checking from cathode to grid depending on the strength of the sig- the input signal level is held within can give misleading readings. Note nal being received. These variable bias limits, very little distortion will be that it's also best to use a digital voltages are produced by the auto- generated. multimeter. Using a low-impedance matic volume control (AVC) circuit. analog meter can give a false indica- AVC voltages are negative and are Bias problems tion, due to the current through the directed at the grids of the front end Some would argue the pros and cons meter. SC

MARCH 1997 85