[R-390] 6080 in place of 6802 - RMS ???

[Barry L. Ornitz]

Joe Foley <redmenaced@yahoo.com> brought up some
interesting questions on the subject:

> Ok, so let's say that from the theoretical
> considerations of this that Dr. Ornitz et al. are
> right.

> I have no postition to argue that anyway, he knows
> more about it than I do.

> BUT, is it possible that that theory doesn't take into
> account the varing resistance of the partially cooled
> filament, it did cool some between pulses where it
> would have been reheated in the opposite polarity
> voltage if used with AC.

The thermal time constant of the 6080's filament is on the 
order of 10 seconds.  Any variations in filament 
temperature due to pulses occurring 60 times a second will 
be pretty negligible (close to 30 dB down).

> Also, because of the 1/2 wave DC there is a
> possibility that the structure is able to dissipate
> heat just as effectively at the higher voltage for the
> shorter period of time as it would at AC.

> Or does RMS power explain it exactly in practice?

Again the thermal time constant determines this.  The 
temperature the filament reaches will be determined by the 
RMS value of the waveform applied as long as the period of 
the waveform is much shorter than the thermal time 
constant.  The thermal time constant of many small 
receiving tubes is on the order of 20 seconds.

> Just how do they establish the rated voltage/power?

From the needed temperature for proper electron emission...

At low filament temperatures, a tube's emission is solely 
determined by the filament temperature (the principle 
behind temperature limited diodes used as noise generators 
for testing receiver noise figure).  With oxide coated 
cathodes, like those used in the 6080 and most modern 
tubes, cathode emission is far more than adequate for 
normal operation.  So the cathode temperature is chosen as 
a compromise of high electron emission versus filament life 
and excessive "boil-off" of the oxide emitters.

Because tube filaments operate at lower temperatures than 
do the filaments in incandescent lamps, tube life versus 
filament voltage is not quite as sensitive as it is in 
lamps.  However the relationship between lamp life and lamp 
voltage is well documented and is a good thing to 
understand.  Chicago Miniature Lamp, Inc. gives the 
following relationship:
                                                     12
   Life at Operating Voltage     |  Rated Voltage   |
   -------------------------  =  | -----------------|
     Life at Rated Voltage       | Operating Voltage|

[Doing equations in ASCII is difficult, but the second ratio 
is raised to the 12th power.  The voltages are all 
expressed as RMS.  Read this in a fixed width font.]

Basically this says that raising the voltage by 10 percent 
shortens the life to less than a third the normal life.  
Likewise, dropping the voltage by 10 percent more than 
triples the lamp life.  [As an electrician, I am sure Joe
realizes that 130 volt bulbs last far longer than do 120 
volt bulbs.]  This is also why those 24 volt bulbs are 
better replaced by 28 volt bulbs in the R-390 series.
  
This improvement in life is not without its costs though.
The rated light output drops with operating voltage.  In 
the case of a 10 percent reduction in voltage, the light 
output is only about 70 percent of what it was at full 
voltage.  The light is somewhat more yellow due to the 
lower filament temperature.

But as I said earlier, this equation does not directly 
apply to tube filament voltages with their lower operating 
temperatures (at least with oxide coated cathodes; it holds 
fairly well with high-power directly heated filaments).  
However it does show that excess filament voltage is not a 
good thing.  Deviations of the filament voltage in small 
receiving tubes of up to 10 percent are generally 
acceptable, but for transmitting tubes, especially directly 
heated filament types, it is best to keep the deviations 
within 5 percent.

        73,  Barry     WA4VZQ          ornitz@tricon.net