[R-390] 3TF7 Ballast Tubes

[2002tii]

Paul wrote:

>What this proves, though, is that your 3TF7 may be working, 
>stabilising heater current, but it might not be the correct value of
>current!  FWIW, I've seen the same stabilising effect in filament 
>globes, but never a convenient value.

I'd have thought that the last zillion times this has come up on the 
list would have been sufficient.

The fundamental goal is to keep the cathodes of V505 (BFO) and V710 
(VFO) at a constant temperature to stabilize their operating 
parameters.  To a first approximation, this will be accomplished when 
the power dissipated in the heater of each tube is constant (this is 
only an approximation because even at constant heater power, the 
cathode temperature will vary with tube operating current and the 
heat loss path from the tube to ambient temperature).

The 3TF7 attempts to regulate the current through the tube 
heaters.  However, (i) it is not a particularly good current 
regulator, and (ii) current is not the right thing to regulate in the 
first place.

Imagine a tube heater fed from a perfect constant-current 
source.  Being incandescent, the heater has a positive temperature 
coefficient of resistance -- as its temperature increases, so does 
its resistance.  But since E = I x R (Ohm's law), as R increases, so 
does E.  And since P = E x I (assuming a non-reactive load), and I is 
constant, P rises.  Thus, when the current is regulated, the heater 
temperature has positive feedback -- greater heater temperature tends 
to increase the power dissipated by the heater, which increases the 
heater temperature even further.  And vice-versa.  Luckily, this is 
not the only feedback loop in the system -- because of the relatively 
low positive tempco of resistance of the heater at the operating 
temperature and the characteristic of the radiative cooling versus 
temperature of the cathode, there is not enough overall positive 
feedback in the system for it to run away -- but there is enough to 
destabilize the cathode temperature to some degree.

By contrast, imagine a voltage-regulated heater.  As the temperature 
increases so does the resistance, thus reducing the current (Ohm's 
law, again), lowering power, and stabilizing the heater temperature.

At the end of the day, nobody will notice if they change to 12BA6s or 
to a resistor across the 3TF7 socket.  And while regulating the 
heater *voltage* increases the stability of the oscillators, once 
again, the improvement is not likely to be noticed by anyone using a 
390A unless they are trying to copy extremely narrow FSK, PSK, or 
something similar (i.e., something way beyond the intended uses of 
the radio).  Note that some replacement solutions use switching 
converters, which have the potential to create interference -- think 
very carefully before you install a switching regulator into a quiet radio.

Finally, with regard to the "messy" looking internals of 3TF7 -- note 
that the wire they are wound with is not of uniform cross-section 
from end to end. The production techniques used to accomplish this 
account for much of what appears "messy."

Best regards,

Don


Copyright (c) 2011.  Not for redistribution