[Milsurplus] Smart People: BC-9 TX/RX Freq Shift- Q?

[Hubert Miller]

The IR loss is a static figure, not dynamic. What lowers the Q is the decreased plate impedance of the tube with the increased tube current; plus the change in tube interelectrode capacitance.
The higher Q of the loop - not defined - may reduce that effect, but it is NOT going to completely eliminate it. One way to counteract the tube capacitance effects is to use as much capacity as
possible in the tuned circuit. ( I don't know how much capacitance you are using with the coil, compared to what is needed to resonate the loop. ). 
I suggest this frequency shift was always present to some degree with this set. 
I have read about some early Forest Service superregenerative transceivers which also had a transmit - receive frequency offset. As a communication took place, the operating frequency
gradually crawled across the band as each operator readjusted his set to pick up the frequency-shifted other station.
Another possible issue with this BC-9 was that, look, you have an oscillator operating at ~4 MHz with the frequency determining component sitting out there in the breeze. Any shift of the
operator's position, or possibly someone walking by too close, is going to shift the frequency. 
But they used it; somehow it did the job.
The later loop sets at least had the receiver tuning components inside the radio, so the loop on transmit could be zero beat to the constant receiver frequency.
It will be very interesting to learn how the set performs with the large loop antenna!
-H 

>You have substituted a link-coupled coil.
It has an inductance, a different distributed capacitance and a series resistance, which is much larger than that of the loop.
When you switch to Transmit, your greatly increasing the loop currents, the IR losses and distributed capacitance losses, lowering the circuit Q, flattening the resonance curve  and thus moving the oscillation point.  The problem is not the B+; the problem is the degraded Q of the circuit."