It isn't the Q but rather the mutual coupling between primary and
secondary windings. Hammarlund had a patent on this kind of transformer
which it used in the Super-Pro series of receivers. In the Super-Pro two
of the IF transformers have movable coils which are controlled via cams
operated by the selectivity knob on the front.
Richard Knoppow
Well, to be sure, the scheme (old as you and others remember) sure
does change the mutual coupling between the primary and
secondary windings, but mutual coupling as a factor by itself does not
affect
bandpass. The mutual coupling adjustment, in this instance, by moving one
coil
into or out of another's magnetic field does affect the "Q" or "quality"
factor of
the tuned circuit's inductor. This, of course affects the resonant
bandpass shape,
Changing the coil's mutual coupling mechanically (pulling the rod)
also
affects the tuning of each coil slightly, and it was sometimes
recommended to
"repeak" the tuning of the '453's cans after changing the coupling. I
found that
I could not tell any difference by "ear", and in those days (late 1940's)
I had no sweep generator
or oscilloscope to "see" what was happening.
Old Chief Lynn, W7LTQ


I tried to find an illustration of the effect I am talking about on
the web but could not. It would make things simpler.
Q is a measure of the ratio of inductive reactance of an inductor to
resistance. the higher the value of Q the better the inductor but there
are circumstances where the Q may be delibrately limited. The bandwidth of
a resonant circuit at resonance is affected by Q, in fact, the definition
of Q is the ratio of the half-power bandwidth to the resonant frequency.
Varying the Q of a resonant circuit also varies the amplitude, the lower
the Q the greater the losses and the lowe the amplitude.
Varying bandwidth by varying the mutual inductance of a transformer
behaves in a different way. Up to a value of coupling known and critical
coupling the bandwidth of the transmission curve does not change
significantly but does increase in amplitude. If coupling is increased
beyond critical the transmission curve becomes double peaked. Where there
is no other coupling than magnetic the two peaks are symmetrical around
the center frequency. Their deviation from the center frequency increases
as coupling is increased but the amplitude does not decrease until very
large values of mutual inductance are reached. The Q of neither side of
the transformer is affected.
There are many variations on the idea of providing for variation of
mutual inductance. The Hammarlund method, using a physically moving
coupling coil, allows the coupling to be varied without introducing
variations in capacitance. Other methods, such as the one used in the well
known Hallicrafters SX-28, vary both mutual inductance and capacitive
coupling so that the two peaks gotten with more than critical coupling are
not symmetrical about the center frequency. In fact, one tends to stay
about at the center frequency while the other moves.
It is possible to get symmetrical variation without using a moving
element and this is done in some later variable coupling IF tranformers.
Again, there is no effect on the Q of either circuit.
Now, the bandwidth of an IF or RF transformer at critical coupling _is_
affected by the Q of the component coils which also affect the efficiency
of the transformer. However, the variation of this Q is not generally used
to vary the bandwidth of the transformer.
All of this stuff is covered in many books on receiver design and basic
circuit theory. The trick is finding one which is not overly mathematical.
Richard Knoppow
Yes, I must agree with Richard's well written explanation of
the effects of overcoupling (more than critical coupling). However
I do not believe that in the instance of the 453's 85kc IF cans that
with the rods pushed all the way in that they became overcoupled.
With resultant double peaked bandpass.
This is of course is just my observation with a bit of practical measurment
using relatively crude methods. It would seem then,to me, that the double
peaked
curve encountered with overcoupling (more than critical coupling)
is, in this instance, not the condition that results in a narrower bandpass
when the coils are moved further apart.
During the glory days of TV service, IF alignment with sweep
generators and
oscilloscopes, it was quite fascinating to see theoretical bandpass tuning
in the
real world. Overcoupling was often (always?) used to achieve the necessary
wide
bandwith for a TV channel. While transformer coupling was not normally
adjustable, the peaks and skirts of the quite wide bandpass could be moved
around with the adjustments available, and measured with fairly simple
equipment.
But I digress (a lot, sorry)
Old Chief Lynn