[Amps] Regulated filament current

[John Lyles]

I agree with W8ZR's recommendation for regulation if it is used for 
filament power. I spoke with a high power tube application engineer at 
one of the manufacturers a couple years ago. He considered three 
methods: constant voltage, constant current and constant power for 
thoriated tungsten direct heated tubes. As a filament ages, and the 
carburization layer thins, the resistance of the filament changes. If 
constant current regulated, then the power into the filament and the 
heating temperature changes as the voltage will track the resistance 
change, trying to keep I = E/R constant. With constant voltage 
regulation, as R goes down, I goes up and vice versa. Over time, this 
has the effect that tube filament heating will increase slowly, which 
will hasten the process of decarb'ing the tube to eventual end of 
emission. Constant Power regulation is probably the best approach but 
not common or simple to implement.

But constant voltage, for the reasons below, is simple to implement, 
more immune to noise and fairly safe for the tube. It does require a 
current limit during turn on, typically set at some value above normal 
hot current, so that during turn on, it inrush of current is limited.

I work with one tube whose manufacturer specifies maintaining filament 
current, not voltage. The RCA/Burle/Photonis 7835 triode is the beast. 
Over time, we have to increase the current (just like the voltage in a 
VR mode) to keep emission adequate for the RF power we need. It has the 
strange result that if a tube breaks (going to air) the filament current 
stays fixed and only the voltage shifts downward, as the tube is destroyed.

Most of the other tubes I work with have voltage as the primary spec to 
watch. Not power supply voltage but the voltage at the socket or as 
close to it as possible. This isn't so easy with grounded grid where the 
cathode is hot with RF, so its measured at the cold end of whatever 
choke or cavity resonator feeds the cathode end of the tube.
73
John
K5PRO


On 9/4/15 10:00 AM, amps-request at contesting.com wrote:
> Personally, for a regulated filament supply, I prefer a constant voltage
> supply, with the current limit set just above the maximun current that the
> fully warm filament will draw. If the voltage is set at, e.g., 12.6V, then
> there it will stay as the tube ages, or is replaced with another tube. The
> voltage can never exceed 12.6V up to the current limit.When the filament is
> cold, the voltage will drop below 12.6V, but the current will never exceed
> the current limit. On the other hand, with a constant current supply, the
> voltage will drift around as the filament changes, but will never exceed the
> compliance voltage setting. The high (nearly infinite) output impedance of a
> constant current supply will be more prone to RF pickup, hum fields, etc.
> The near-zero output impedance of a constant voltage supply will likely be
> much more immune to such things. The power supply itself acts like a huge
> bypass capacitor for AC fields (up to a frequency that depends on the
> particulars of the supply. On lab grade supplies, such as Agilent, Kepco,
> Lambda, and Keithley, the frequency response is carefully specified; you
> probably won't find any mention of it on an MFJ or Astron supply.
> 73,
> Jim W8ZR