TRANSMITTER TUBES

operations of HIGH POWER BROADCAST TRANSMITTER TUBES

8984 Tube

The tube has several parts.

Filament, or Cathode

Grid

Screen

Plate

Pentode Power Tube
Electrons pass from the cathode (in red)
through the Grid
through the Screen
and strike the Plate.

The tube has several parts.

Filament, or Cathode
Grid
Screen
Plate

8977 Tube
Tube busted in two parts...

Busted portion at top of picture is the plate
(with lots of air fins and heat dissapation).

The bottom broken part shows structures inside the tube. Screen (vertical wire cage) is the outermost structure.
The grid and filaments are hidden under the screen.

RULERYEL.GIF, 4 kB

CATHODE
The cathode is heated by the filament. In most cases the two are one. RF drive is typically applied to the cathode in high power transmitters, and is relatively large; well over a thousand watts. The impedance, and consequently, the voltage of the RF drive varies, depending upon tunning and loading adjustments. In most high power transmitters, a 50 ohm pure resistive load is also used across the cathode. The load minimises tuning effects, of not only the "reactive" input of the tube, but also the output tuning of the previous stage. In any case, the RF drive voltage overcomes the (constant ) negative grid voltage, at a rate of the RF frequency. The result are groupings of electrons accelerating towards the plate.

The cathode is where the initial supply of electrons takes place. The grid permits a group (packet) of electrons to leave the surface of the cathode. The screen and plate accelerate the group; first past the grid, where a small amount of low velocity electrons collide with the grid structure, resulting in about -50mA of unwanted current in the grid. Next, the group accelerates past the screen. The screen, about 1500 volts, initially attracted the electrons, and would burn and be instantly destroyed, except for the higher plate voltage, of about 10kV, which saves the screen from certain failure. Together, with the electron momentum, the large plate voltage continues the acceleration - successfully - past the screen. The screen also divorces, and buffers, plate effects. The beam collides with the plate; resulting in mostly resistive heat.

RULERYEL.GIF, 4 kB

GRID
Power amplification is realized by the use of High Voltage on electrons. Electrons are accelerated from the cathode to the plate.

The electron beam is controlled by the grid-to-cathode voltage.
The grid is normally about -100 volts DC to the cathode. The result is a operating point that determines quiescent (static) current in the tube. Static currents are typically about a forth that of Dynamic operation currents. (An example for a 50PKW transmitter would be 10kV plate voltage and 1.5A Static plate current yielding about 5A driven plate current.) The grid to Cathode voltage determines the instantaneous values of plate current: in the example all values between 1.5A and 5A.

RULERYEL.GIF, 4 kB

PLATE

Electrons are accelerated towards the plate.
Each electron has this Kinetic Energy:

Most energy in the tube is heat energy, and must be dissipated by either Forced Air or Liquid Coolant.

Source electrons

Heat

Other

V
V=Voltage

I
I=AMPS

And total power will depend upon voltage and current in the tube.
Total power is called plate dissipation.

And RF power is only a percentage of this.

TOWRADIA.gif, 2 kB

Fortunately, in proper operation, the electrons are slowed (in fields) and thereby impart some of their energy to tuned circuits. As electrons approach the plate, they are slowed BEFORE impact.
And this results in the desired RF energy.
The RF energy will later be delivered to transmission lines, and ultimately, to the antenna.

Tube and Plate Physics

RULERYEL.GIF, 4 kB

FILAMENT
Additional current, as support current, is sent to the filament for heating. All of this support current energy is used to heat the filament, which is necessary for cathode emission. In some older transmitters the filament is called the heater. (Typically 150A current (at 12.6V) inters into the filament, and exits the filament in a cable about 3/4 inch in diameter.)

Filament management is crucial for tube life.
An engineer must keep the filament about 10% above filament starvation for any given power level.
An engineer must keep the filament out of the cathode poisoning level at any cost.
In addition to poisoning, cathodes ran too low, and made of an oxide coating, will have the plate voltage "strip" filament oxide material. This will produce an internal - and devastating - arc. Modern tubes use carbonised thoriated tungsten filaments, and this surface will not flake off or strip. The carbonization gives the filament it's emission, but degrades over time. Thoriated tungsten filaments begin life with low current at rated voltage. As the tube looses carbonization, the filament current begins to increase at any constant voltage. In other words, the resistance goes down. As the tube ages, resistance can decrease 20%. I have been forbidden to publish my (company) data: which is 35 years of work on this.

If records of filament current, plate current, and sync compression are not analyzed every week, the tube will have a short life. There are no high power television transmitters (tubes, klystrons, etc) that do not incur this small and necessary auxiliary support power.

This current is usually rectified AC (unfiltered DC). The reason is simple: This high current is difficult to regulate directly. Instead, the PRIMARY AC source to the filaments is controled or regulated. Usually regulation is by Saturable Reactors. Older styles may have motor driven transformers. And low power filaments may have autotransformers ("Solas"). Regulation is so important that there are no high power transmitters without it.

AUT8.gif, 115 kB

When the transmitter was going to be idle during the night, my control system would reduce the filament voltages a small amount: about a volt. Actually, a volt is a huge amount on a broadcast transmitter tube. A volt can spell disaster for a tube due to either "cathode stripping" or "emitter poisoning". And only a good control system could attentively prevent (actually delay) a transmitter from going back into a radiation mode with such a dangerously low voltage.
For many years, my system prevented thermal stress by "gradual excitation" and "stepped" filaments.
... And increased tube life expectancy;
... And also saved a little money, along the way, with reduced utility bills.
- All accomplished with my automated filament control.
Filament management was only a small part of a much larger control system. I designed it all.

RULERMAR.GIF, 1 kB

OPERATION

(Note: in a uniform field, separation distance is not a factor for either energy or final velocity) Cathode is a cylinder inside a larger cylinder (the plate). The effect is a nearly uniform isometric electric field.
5A=5q/s
Eqn11.gif, 2 kB

The total 50kW input energy is adjusted by grid bias voltage adjustments, screen grid voltage adjustments, and by filament voltage adjustments. An optimum realistic division of energy in this example of 50kW are ...
29kW Heat
21kW RF
The RF average, must be multiplied by 1.68 for the peak rf power derive peak power A slight misadjustment with tunning controls can dramatically increase heat and decrease RF.

The tube is housed and operated inside a tuned cavity. The Plate is part of the tuned circuit as well as the Cathode. The Plate and Cathode resonate 180 degrees apart in phase. Electrons are slowed

by electrostatic and magnetic forces of oscillating fields in the external tuned circuit. Electrons which are accelerated by the electrostatic forces of the plate "pump" the tuned circuit electrons, and give up energy (velocity) to the desirable RF fields. The objective is to harvest RF energy.

Excess Circulating Current
The cause of the effect is not well understood. But the cause is well known to me. The effect is probably the number one reason for tube breakage. If a tube is tuned for max power with plate tuning controls - and at the same time - tuned for MINIMUM plate current with LOADING controls, (High Q Tuning), then a high power tube will likely - catastrophically - implode by cracking. Ignorance is compounded by the need, as a tube ages, to get more power through efficiency adjustments. A fatal combination!

Oh yes, one more little detail: it can kill you!
You may breath in BeO ceramic dust from the implosion. (Not likely, but possible.)

My good friend Russ Pope promoted the idea of drilling a hole in the side of the cavity and installing a small loop to measure circulating current. RCA liked the idea and employed the technique to their TT-10 transmitter line. Enough energy was obtained to run a passive meter which was hung near the cavity. Of course the meter would rise with increasing power, but also and more importantly, the reading would vary according to tuning. In extreem cases, to the point of instability. If tuning was to sharp the tube was in danger of imploding.

I have received ridicule...
After all - it is the opposite of what is written in text books - on how to tune a transmitter. The problem is that engineers due not write text books. Perhaps because they work in a real world with their hands, and do not have time. On the other hand, schools and conferances DO acknowledge the real world. Such a pleasure to be with "these", and away from "those".

RULERMAR.GIF, 1 kB

Health concerns

Hard Radiation
LABLEXRAY.JPG, 4 kB

Hard Radiation

RULERYEL.GIF, 4 kB

Beryllium oxide (BeO)
Beryllium oxide (BeO)

AUT8.gif, 115 kB

Prevention of thermal stress and gradual ceramic cracking:
My system would also bring RF power down - gradually - before shutdown, and bring RF power up - gradually - after an operator command for Radiation. This automatic procedure also alleviated thermal stress to ceramics, and the rest of the transmitter in general. However, unavoidable cracking has occurred during power outages, resulting in tubes cooling too quickly. My remote control systems have been helpless in this situation. Undoubtedly, in those cases, some beryllium oxide has been liberated.

RULERYEL.GIF, 4 kB

Soft Radiation
Soft Radiation

RULERYEL.GIF, 4 kB