Tube Bias And The Capacitor.

Tube Bias and the Capacitor ~  By Bill Jones

 

The article is a review of how a tube obtains a negative bias on its grid. Included is a short history of how it has been   done, and   how it should have been done.  Since the capacitor is a major component in the power supply, grid bias, and the radio in general there is a discussion of the capacitor its invention and, application to the power supply because the power supply is often used in obtaining grid bias. When using the power supply for grid bias it generally complicates the circuitry and, adds little to no value to the operation of the radio.   The capacitor is necessary to obtain a “clean” bias from a power supply. Grid bias is very important in the operation of any tube and should always be a consideration in the repair and operation of the radio.  There are some occasions when it may be proper to use the power supply as a grid bias means, but not in the “modern typical radio”.  This will be discussed later.

The “modern” radio is considered as the era of the “miniature tube” by my definition, but modern could imply an improvement in the concept and operation of the radio.  While there has been great improvement in the components used in the radio, the concept and theory of radio has been essentially unchanged. When I started my electronics design career, the miniature tube was the only tube that I used.  I have been a circuit designer for almost my entire career, and perhaps this explains my interest in radio...  I designed many devices using the miniature tube (including communication receivers) and the tube was widely used in the late 1950’s.  The transistor made the tube almost obsolete in the early 1960s.  After 1962 I used only transistors and integrated circuits.

The modern radio generally obtains its tube bias by means of the cathode resistor. The word “generally” is used because some manufacturers used the power supply as a source that furnished a negative bias for the modern radio. This was not necessary, and it did complicate the design, manufacturing and the service of the radio.  The reason for the negative source from the power supply is not clear, but I think the change from battery supply to AC supply was initially responsible for this, because that was the way battery radios were usually operated – i.e. a separate supply source as compared with a cathode resistor for bias. The first radios used various means for bias.  One means was a separate battery called the C battery and it supplied a negative bias to the grids. Cathode bias could have been used but it would have required a separate filament supply for each tube that required a different bias voltage. Clearly, no one would do such a thing. To my surprise, it was done in the Audiola 889. The 889 used four filament outputs with cathode resistors to obtain the bias as required for the tubes. Cathode bias is not a new thing.  I had assumed that it was not used until the indirectly heated cathode appeared.     Patent infringement problems regarding the use of cathode bias have not been found.

When using cathode bias, it is usually considered necessary to bypass the cathode resistor. If the resistor is small, however, it is optional. When the resistor is small, dispensing with the bypass may lower the gain by only a few decibels. There are cases where it is actually desirable to leave the bypass off the cathode.  It is not unusual to find radios without cathode bypass in both the RF and the audio sections. The bypass capacitor for the RF stage is a small capacitor because the frequency is relatively high, but when considering audio amplifiers, the capacitance must be quite large for effective bypass of the cathode resistor. There are cases, where, for example, a push-pull output stage would not require a   bypass on the cathode resistor. If the stage used class A bias the effective cathode output resistance is very low.  Bypass on the class AB cathode is almost optional but good design would suggest the use of a bypass. In the early TRF radios, it is not unusual to find push-pull output stages with no bypass on the cathode.  Over the years, this push-pull arrangement has been very popular.  The class B stage is seldom used because of the crossover distortion.  The class B requires fixed bias.

The capacitor is said to have been invented in 1745. It was known as the Leyden jar capacitor, and was invented in Leyden, Newfoundland at the Leyden University. Cornel Dubilier is often credited with the invention of the mica capacitor in 1909. Credit may also go to Dubilier for the invention of the solid electrolytic in 1936.  The first electrolytic was invented by Julius Edgar Lilienfeld in 1921.  This was the wet type.  Lilienfeld also has an early patent on a solid electrolytic but it is not clear how different it was compared to the Dubilier solid electrolytic. There is a complicated application of chemistry required in electrolytic design.   This capacitor is quite specialized with many patents regarding the various dielectrics and capacitor plates.  There are now many types of capacitors with most capacitor types having a particular application.

The appearance of the AC power supply with the electrolytic dramatically changed the appearance of the radio.  It reduced the requirement of multiple chokes.  Before the electrolytic, it was necessary to use paper capacitors.  The paper capacitor is good and has great longevity if properly built but it is physically large when used in a power supply. The chokes had to be large and thus heavy because the capacitance values of these paper capacitors were so small. These paper capacitors had capacitance values in the two to four microfarads with eight microfarads being unusual.

The very early tube sets often used a grid bias resistor called the “grid leak”.   The grid leak was also used in conjunction with the C-source.  This grid-leak is explained in several ways, all generally confusing.  The vacuum tube grid will provide a very slight grid current. This current passing through a large resistor (in the range of 2 megohms) will give a negative voltage on the grid, thus grid bias.  This works great in RF amplifiers, because it simplifies the circuit.  There are no wires going back to a C source that require bypassing.  The grid-leak detector requires the RF signal to be large enough to draw a very slight grid current on positive peaks.  The capacitor driving the grid then is sized such that it charges to the peaks of the signal (a rectifier) but the resistor is allowed to discharge through the capacitor. The resistor may also be adjusted to allow some bias.  It is important that the tube be operated in a non-linear manner for a grid-leak detector.   The RF signal is actually detected by the grid and then amplified by the tube.  The grid-leak bias, and the grid-leak detector and the negative C source were used in various combinations in the early radios. Bias reliability was a problem, partly because of the very large meg-ohm resistors, and the tube itself did not supply a reliable value of grid current.  There are other means of bias such as the so-called contact potential from a diode plate. This is similar to the current from a tube grid but more reliable.  In superhets this bias may be taken from the AGC line and used for the RF amplifiers. It is doubtful that it would be used without a cathode resistor.  It is also possible with the superheterodyne to use the oscillator negative grid voltage as bias, but unusual.

Obtaining a negative bias from a power supply is not difficult.  Simply inserting a resistor in the center tap return lead of the power supply transformer will make a negative voltage available. The ground lead of the first input filter capacitor is returned to this resistor. The negative lead for the normal power supply is the center tap of the power supply transformer. The resistor will go from this center tap to ground.  The first input filter capacitor ground lead then will be returned to the center tap.  This provides a negative voltage at the center tap, usually with a large amount of AC present.  To use this bias in an audio application it is usual to obtain further filtering. Since there is only an insignificant grid current flow, it is possible to use

large resistors and a small paper capacitor for further filtering before the bias voltage is used in the radio.  

The negative supply adds complexity to the radio. The first or input electrolytic must be isolated from ground. If this is not observed and the circuit is grounded – it will cause the radio to operate without bias and may cause extensive problems.  The resistor inserted into the negative lead is often of a physically large size because of the power dissipated in the resistor.  All the current drawn by the radio will flow through this resistor and is an unnecessary power loss.  The negative voltage is subtracted from the positive side of the power supply. That is, if the negative supply is, for example, 30 volts, the B+ will be reduced by that amount. That is, the B+ could have been 30 volts greater.

The first of the AC power supply radios generally used directly heated filaments. This made the use of the negative supply necessary when the filaments were connected in parallel with one side of the filament connected to ground. With the introduction of the indirectly heated cathode, I assumed that it would have done away with the negative supply but it did not.  Philco and Zenith (and others) often grounded the cathodes of the audio tubes and used the negative supply even into the 1950s in some sets.  The first AC radios needed a large negative supply when audio tubes like the 45 became available, because the grid voltage for these triodes, for example, could be as high as minus 50 volts. The negative supply was usual, because small electrolytics having a large capacitance value were unavailable for a bypass.  The electrolytic can be a fairly leaky device and also a relatively noisy capacitor. Its value may drift depending on how it is used voltage wise and this was probably one reason that it was not used in early audio applications but by the late 1930s I believe that there were very good dry electrolytics available. Many companies used them. I surveyed a few of the 40s and 50s radios and found many using a bypassed cathode resistor. I noted that most, but not all, of the car radios, used cathode bias with a bypass.  I consider the car radio to be of a better quality and performance (when the vibrator works) than the normal household radio.

Without the invention of the capacitor, radio would not have been possible.  The paper capacitor is great but like most capacitors it has limitations.  The paper capacitor does not do well in RF applications because of its lack of stability and high RF resistance. Each capacitor type will have a particular area of application and there are many types of capacitors.  The invention of the mica capacitor was an important event for RF applications because of its low loss and stability. The mica capacitor has been used for many years and has had an unusual longevity.  The ceramic capacitor has replaced the mica in many low voltage applications but the mica is still useful in high voltage applications.     

The terms used for radio history, such as early, very early, are at least vague, but my interest has been in the various means used for tube bias. This note has been written from the viewpoint of a circuit designer and not radio historian. The early designers led the development of the modern radio. They did remarkable work with little test equipment, but they had great imagination and intuition.

It had been assumed that the indirectly heated cathode would have removed all requirements for a negative supply with its expense, and complication, but a short review of the 1940s and 1950s schematics have shown the notion to be without great merit.  While the cathode resistor was widely used for bias in modern sets, some sets still used bias from the power supply center tap. This bias means increased the cost, complexity and was (is) not necessary.

.

Questions or Comments? Please e-mail me at whnj@att.net  Thanks, Bill.