Why does the 75A4 noise limiter tube run at reduced filament voltage?

[k4kyv]

I was  looking over the schematic of my 75A-4, and noticed that V-12, the 6AL5 dual diode noise limiter has a 10-ohm resistor in series with pin #4 which goes to the filament.  The voltage/resistance chart lists the normal voltage on that pin as 4.3 volts instead of 6.3.  The other side of the filament is grounded.

Does anyone know why Collins might have reduced the filament voltage to that tube?  In the circuit description there is no mention of what purpose this is supposed to serve.

[w3jn]

That's typical with detector circuits as well; many Hallicrafters and National 6H6s, 6AL5s, etc use this arrangement..  Reduces the contact potential of the diode, and decreases hum in small signal circuits.

[AB2EZ]

I was thinking about this on the train ride into work today.

I would assume that JN has a lot more background on this... so consider my (alternative) answer as just a guess based on the physics of the vacuum tube diode:

Bottom line: my guess is that by lowering the filament voltage (and therefore the cathode temperature) the Collins folks were attempting to introduce an offset in the turn-on voltage of the diode. I.e., to make it turn on at some positive plate-to-cathode voltage that is higher than zero. Of course, there are other ways to introduce an offset in the plate-to-cathode turn-on voltage (e.g., apply a reverse bias through a sufficiently high impedance)... but those approaches had their own engineering challenges back then (circa 1955).

More details:

In order for an electron to leave the surface of the cathode (i.e., to overcome the surface "work function") it must have a sufficiently high velocity in the direction perpendicular to the surface.

At room temperature (no filament voltage) the average thermal energy of the electrons is such that it is extremely unlikely* that an electron would have enough kinetic energy to overcome the "work function"... and thus leave the surface of the cathode.

[*The electrons have a statistical distribution of thermal energies... so some have more than the average thermal energy and some have less. The statistical distribution is known as the Boltzmann distribution.]

At full filament voltage, the temperature of the cathode is sufficient to produce a  high enough average thermal energy of the electrons... such that a significant fraction of the electrons have enough kinetic energy to overcome the "work function" and leave the surface of the cathode. With no plate-to-cathode voltage applied, they form a cloud of electrons near the cathode. When positive plate-to-cathode voltage is applied, these liberated (from the surface of the cathode) electrons will flow toward the plate... producing the forward current.

With reduced filament voltage, the temperature of the cathode is not sufficient to produce a high enough average thermal energy among the electrons... such as to cause a significant number of electrons to be liberated from the surface of the cathode... and available to produce a current when the plate is slightly positive with respect to the cathode.

However, at this reduced filament voltage, the temperature of the cathode is high enough so that a combination of their thermal energy and the potential energy associated with a high enough electric field (near the cathode) will cause electrons to leave the cathode and flow to the plate.

This field, near the cathode, can be produced by the application of a sufficiently high plate to cathode voltage. [I.e., the field strength is roughly equal to the plate-to-cathode voltage divided by the distance from the plate to the cathode]

Thus, by reducing the filament voltage, the diode will not conduct until the applied plate to cathode voltage is above a certain threshold.

Stu

[w3jn]

You're pretty much on target, Stu.  The effetcive contact potential between the cathode and plate is also affected by the space charge.  Reducing the cathode temperature reduces the space charge thereby affecting the effective contact potential.  Or so I understand it from my nuclear physics class from 25 years ago 

[k4kyv]

I recall once a poorly shielded rf buffer stage in a transmitter would self-oscillate with nothing but the filaments on.  Turned out to be the 5R4 HV rectifier tube.  With the plate voltage turned off by opening the primary circuit to the plate transformer, the 5R4, with no a.c. applied to the plates, would develop about 20 volts DC at the power supply output, and the unexcited HV winding provided the continuity to complete the circuit.  Enough electrons from the cloud that developed around the heated filament would strike the plate to develop the 20 volts, and the rf stage, which must have been very prone to self-oscillate, would take off with that small plate voltage.

I fixed the problem by replacing the 6V6 tube with a 6AG7, which is much better shielded.

That electron cloud may actually shorten the life of a tube as the a.c. filament current makes the  electrons vibrate back and forth at 60~ and literally sand-blast the coating on the cathode or directly-heated filament.  The R-390A manual warns against running the receiver for long periods in the stand-by position.  For extended periods they recommend turning the receiver off entirely to avoid shortening the lives of "certain tubes".

The same thing occurs with a transmitter when it is run for extended periods with just the filaments on.  The tubes would last longer with a few plate and/or screen volts applied, just enough to cause a small plate current to flow, enough to keep the electron cloud drained away.

RCA recommends cutting the filament voltage down to 80% normal during stand-by periods, and cutting the filament off entirely if the transmitter is expected to remain idle for longer than about 2 hours.

But I never thought about the normal filament voltage of a small diode tube like the 6AL5 causing any kind of problem.