[R-390] On 3TF7's...

[Drew Papanek]

On the topic of the 3TF7 ballastube, Dallas Lankford wrote:

(snipped)

>...I found that the 3TF7 does not do a very good job of stabilizing the
>BFO and PTO frequencies when the AC line is varied.

There are other failure modes for the 3TF7 other than just going "open".
I have one which "regulates" at 350 mA; another list member reported having
one at around 240 mA (the spec is 300mA).  The degree of regulation probably
is degraded with that type of failure.  It could be that Dallas' 3TF7 has
that defect.

Dallas continued:

>Whoever designed the original BFO and PTO filament stabilization circuit
>was on the right track.  They just used the wrong method to stabilize it. 
>Current regulation is the wrong approach; voltage regulation is the correct
>approach.

A good 3TF7 is specified to hold current constant within +-1% over the
voltage range.  Perhaps that performance is not adequate for the ultimate
in PTO/BFO frequency stability.

Either current regulation or voltage regulation would accomplish the same
desired result in terms of frequency stabilization.  The designers used
current regulation because that was far simpler to implement with the
technology of the era.  Their current regulators weren't as good as those
we can inexpensively build today.

Current regulation has an added benefit in that it eliminates inrush current
surge on startup.  That high current (a bit over 2 amps for a 6BA6) causes
a brief heater hot spot which eventually burns open.  Tube heaters are
generally designed to reduce that effect, but it still remains to some
extent.  (Extreme example: I have a few Bugle Boy 12AX7's the lowest part
of whose heaters flashes brilliant white on startup).

A few current regulators using the LM317 have been described in this forum.
One is Dr. Gerald Johnson's simple AC regulator.  While his circuit does not
get the best performance from the LM317, it is still quite good and much 
much
better than a recalcitrant 3TF7.  It dissipates no more power than the 3TF7 
and
places neither asymmetric load nor current spikes on the power transformer.

Another is Dave Wise's LM317 DC circuit (not his phase control based "3DW7"
designs).  Its regulation is excellent and is adjustable from 270 mA to
330 mA IIRC, but it generates more heat than Jerry's circuit (I calculated
something around 7 watts compared to the 3.8 watts of Jerry's design).
With its half wave rectification it places an unbalanced load with high
current peaks on the transformer.  Dave incorporated series resistance to
reduce those peaks somewhat.

Dallas' voltage regulator circuit generates higher heat and without the
aforementioned refinement used by Dave draws even higher current spikes
in its unbalanced load from the transformer.  Dallas' circuit is not
adjustable and is simpler than Dave's.

May I suggest yet another DC current regulator circuit?  Connect 25.2 VAC
from ballasocket to anode of diode (1N4002 and up suitable).  Connect
cathode to + side of 1000 uF 50V electrolytic capacitor. Ground - side of
cap.  Run DC thus formed from + of cap to input of LM317 (pin 3).  Connect
load at 3TF7 socket (the connection to the seriesed BFO and PTO tube
heaters) to LM317 adjust terminal (pin 1).  Connect a current sense
resistor (4.166 ohms,  1 watt) from LM317 output terminal (pin 2) to the
adjust terminal.  A 4.0 ohm, 1% resistor will give nominally 313 mA
- close enough to the desired 300 mA. For those who want to set the
current exactly (given the LM317's Vref tolerance) use 5.0 ohms in parallel
with 20 to 33 ohms.  Add the obligatory short leaded, grounded .1 uF caps
on LM317 input and output to quell LM317 oscillation/noise tendencies.
Heatsink well.

That circuit would have close DC current regulation, not be as readily
adjustable as Dave's, be simpler than Dave's, more complex than Dallas'
and have the same high input current spikes, unbalanced load, and excess
heat as Dallas'.

Both the circuit I propose and Dallas' could be made to dissipate a little
less heat and be a little less "spikey" by reducing the filter cap to 470 uF
thereby allowing more ripple (not Chuck) at the regulator chip's input. That
would have no noticeable effect on regulation.

While the current spikes and unbalanced load of half wave rectification with
capacitive filtering will cause additional heating in the R-390x power
transformer, I'm of the opinion that there would probably be no detriment
given the transformer's conservative design and massive size.  The
transformer's existing load is symmetric and spike free, comprising
resistive and full wave rectified choke input.  The added half wave
rectified load is small by comparison and hence would be "diluted".

On 3 terminal regulator noise Cecil Acuff wrote:

(snipped)

>There are any number of complex ways to solve the 3TF7 issue but we should
>not be short sighted and forget about any noise that might be generated by
>the solution.  Linear regulator circuits work great but are quite noisy.

>The new fangled sand box radios suffer from many problems associated with
>noise generated by devices internal to the radio...

I don't know how much of a problem that would be for an R-390x using a 3
terminal device for PTO/BFO tube heater regulation.  The radios having that
malady have operating bias voltages so regulated/adulterated.  OTOH,
heater power is not directly associated with the signal path, though there 
can
still be some leakage.  In the R-390x, encountered first after the tube 
heater
regulator's output is the BFO tube heater.  There and at the detector the
signal level is high enough that 3 terminal regulator noise would be
miniscule by comparison.  Next in line is the PTO tube heater.  That is fed
through a brute-force LC noise filtering circuit included by the designers
to keep PTO signal in and noise out; no trouble there.

Any noise problem caused by the aforementioned heater usage of 3 terminal
regulator would most likely be due to radiation from heater wiring inside
the IF module. That would be dependent on existing lead dress and shielding.
An inductor and another .1 uF capacitor could be connected to the regulator
output to form a pi-section brute force filter thereby addressing any noise
concerns.

Not all of us demand the ultimate in stability from our R-390x. On
ballasubstitution, Jerry wrote a while back:

(snipped)

>The purist restorationist will want to use ballast tubes until there are
>no more. The picky will want to go solid state regulation, and the AM
>listener probably will be super happy with a pair of 12BA6 and a jumper.
>Since the 12BA6 was the standard IF tube in 4 and 5 tube AC/DC radios using 
>miniature tubes, there should be a million of them about or more.

A schematic of Jerry's AC current regulator and a component connection
description of Dave's DC current regulator can be found under the "Ballast
Tube" heading in Wei-i Li's brilliantly conceived "Pearls of Wisdom".
Go to r-390a.net   .  Select "References", "Pearls of Wisdom".  There reside
postings from this forum painstakingly distilled over the years.  There is 
much
elightenment to be gleaned and amusement to be had by perusal of the lively
and animated discourse over this most controversial of R-390x topics.

Vive le Caballo Muerto!

Drew

_________________________________________________________________
Check out Election 2004 for up-to-date election news, plus voter tools and 
more! http://special.msn.com/msn/election2004.armx