Grid Bias

[K4RT]

This must be a very elementary concept for many of you, but I've been reading about grid bias in my copy of the 1951 ARRL Handbook and I'm not sure I'm getting it.  I'm thinking in terms of the final power amplifier, but I guess bias applies to the earlier drive stages as well depending on transmitter design. 

As I understand it, bias voltage is applied to the tube grid to prevent the cathode to anode or cathode to screen current from running away and destroying the tube. When signal voltage from the output of the previous drive stage that is fed to the grid increases above the bias voltage that then allows the current to flow from the cathode at levels that correspond with the signal on the grid and amplification occurs.  Is that correct?

73,
Brad K4RT 

[W3RSW]

It is not that elementary.   Bias depends on the class of amplification in which you want to operate, typically A, AB, B and C.  Bias depends on the structure of the tube, the outright size and spacing of the elements, the number of turns in the winding of the grids of the tubes and several other variables including using two tubes in a push pull circuit where each tube handles each side of a 360 degree AC waveform.

Oh where to begin....

Short answer to your question, (and there are zero bias tubes by the way) is that without proper biasing, yes, the tube elements will allow way too many electrons to pass a non-controlling grid, impact the tube's anode, cause it to turn red and melt down if it is not beefy enough or the tube's structure is not large enough to dissapate the heat.  With glass tubes though you generally get a visible warning, an increasingly bright glow tending more and more to the yellow/white part of the visible spectrum, heh, heh.

Well, take a typical triode audio frequency amplification triode.  We are going to amplify in class "A" where the full signal input waveform is faithfully reproduced in the output waveform.

So what about bias to control those electrons passing through the control grid?

Even triodes come in three basic flavors for low signal work, high, medium and low mu (gain).   So let's take a triode and see how bias affects the signal.  When working with tubes as with solid state, to describe one thing all other factors are usually said to remain equal.  So let's assume the AC peak to peak audio voltage feeding the tube stays the same.  Assume the plate voltage is the same and, for now, assume the plate load resistor stays the same as we change bias. 

You can apply bias in a couple of ways in this typical circuit.  You can 'rob' some of the plate (anode) supply voltage by putting a much lower resistor (1/10 of the load as an approximation) in the cathode.  As a constant current crosses this cathode resistor a small voltage is set up so that we have a voltage divider.  Supply voltage = all the drops: across the plate load resistor + the internal resistance of the tube + the cathode resistor.  So now a slightly positive voltage on the cathode means that the grid is now the same amount negative with respect to the cathode as electrons pass by it on their way from the cathode to the anode. You might typically have a few volts negative at the grid.  By the way, the grid circuit has to be normalized back to common with a much higher grid resistor, say 500k ohms from the grid to common supply ground or through a transformer winding while isolating the AC input signal from shorting to ground.  So lets say the grid is now -2 volts.

You could also put -2 volts from an external small supply directly on the grid, with isolation of the AC signal waveform of course and the ground the cathode directly to common supply ground. At any rate (pun intended) an AC signal superimposed on the DC bias will be faithfully reproduced as long as the AC signal does not have peaks that ride outside the envelope of linear amplification for a given tube and amplification/power requirements.  I'm going to stop for now since this type of detail can take up reams of Espace.  Suffice to say that in class B, half the fullwave is amplified, each half by one tube in a two tube push-pull circuit, in Class C, only a portion of half a complete waveform is amplified,  etc. Bias for these classes will be pretty much the controlling factor of which class the amplifier is working.

[W3RSW]

Rereading your post, you sound like you may be familiar with transistor biasing.    Tubes generally take a negative bias to keep from running away and operate in the linear range.  Transistors take a slight positive bias to allow the base to partially conduct and operate in a linear range.

[Opcom]

You said: "When signal voltage from the output of the previous drive stage that is fed to the grid increases above the bias voltage that then allows the current to flow from the cathode at levels that correspond with the signal on the grid and amplification occurs.  Is that correct?"

Yes, when the drive takes the negative-biased grid more positive, towards zero, the anode current increases because more of the electrons are let through. The opposite when the drive swings negative, until the tube electron flow is cut off.

Tubes, aside from a relatively small current during the harder drive in A2, AB2, or B2 operation, are controlled by voltages applied to the control grid with no grid current flowing.
Non zero-bias tubes can be thought of as being 'wide open' when no negative grid bias is applied.
Zero-bias tubes are designed to have the proper operating current when the proper plate voltage is applied. They always require a little grid current because the drive voltage swings positive as well as negative. When operating at their highest voltages, a small amount of negative bias is useful to reduce idling plate current (avoid red hot plates). For an 811 or 3-500Z it is only 3 to 6V negative. Most designs just put a cheap zener diode in the filament return to raise the cathode by the same 3-6V.
In all class A and B circuits, negative bias voltage is chosen to set the idling current to some acceptable value. Class AB is less than 360 degrees and more than 180 degrees. Class B, as usually taught, is exactly 180 degrees conduction and zero plate current at idle. In practice, a small amount of current is used, and the conduction angle is really a little more than 180 degrees, because if the tube is really cut off (>180 degrees), it becomes class C operation.
In class C, the bias voltage is very negative, beyond the point at which no anode current flows, and the drive is great, making the amplifier produce high current anode pulses instead of linearly reproduced signals.

The classes conduction angles apply to transistors as well, but the voltage-current relations are different.

Bipolar transistors are controlled by base current (base is positive on NPN), although the threshold of about 0.6V for silicon must be met before enough current will flow to make the transistor useful as a linear signal amplifier. So the base voltage of an amplifier may measure 0.6 to 0.7V, and over just this little voltage range, the base current and indeed the transistor's collector current will vary widely.

FETs are controlled by gate (positive on N-channel) voltage, with no gate current commonly spoken of although there are charge and discharge currents due to gate capacitance which is typically higher than that of tubes. A FET has a threshold voltage which must be met as well, usually 2-3V but can be lower.
Some FETs designed for logic-level drive (today 0.5 to 1.8V) may have another small FET and a couple of resistors built into the chip to amplify the drive enough to operate the main FET. This type is not as good for a linear amplifier but is used in switching power supplies or contactless power switches or applications where the current is to be pulsed.

If there is an error, anyone feel free to correct.

[W3RSW]

Nice description.  It takes tons of writing to explain what some good illustrations and graphs could show fairly clearly.  I wonder at times of my skill level of writing matching just what level of detail is required.  ..or even my skill level of understanding.

Bias, A vast topic interrelated with just about all vacuum tube theory, commonly right down to the partial differentials defining dynanmic characteristics.  Some of my favorite graphs are the composite characteristics of push pull output stages where Ip vs. Ep of both tubes are shown, one mirrored below the other (upside down) with a common load line running through both. A family of running composit quiescent bias lines with grid input signals overlying is really fascinating and instructive.   It is all just so neat.

[K4RT]

Rick and Patrick,

Thank you for responding. I think your explanations will help.

73,
Brad K4RT

[Jim, W5JO]

Purchase an RCA Receiving tube manual and a Transmitting Tube manual and read the theory sections in the front.  They are  written well and elementary on the operation of tubes.