[HBR] IFT for BFO Coil

[waltah at earthlink.net]

Darrell, WA5VGO wrote:

> Has anyone had any experience converting an I.F. transformer to a
> BFO coil? I have a 100 kC transformer. It has a 110 pF capacitor
> across the coil. I assume if I replace this with a 470 pF and 150
> pF in series, that this will keep the coil resonate and provide me
> with my feedback connection. Is it this simple? 

470 * 150 / (470 + 150) = 70,500 / 620 = 113.71   That's close 
enough!

Cap A *times* cap B divided by (cap A *plus* cap B) -- same 
formula as resistors in parallel.

That'll give you the tank circuit for a Colpitts oscillator.   In the 
traditional Colpitts the cathode of the tube would be connected to 
the junction of the two caps; you'd need a way to ground that for 
DC so the plate current has a place to go.  At higher frequencies a 
choke would work but the choke for 100 kcs is awkwardly large.    I 
suggest instead using a grounded-cathode variant of the Colpitts 
circuit: 

Ground the cathode of the tube directly, also the junction of the two 
capacitors as above.  

Feed the grid through a cap from the 150 mmf end of the coil -- I 
suppose 470 mmf or so, and this cap must handle the high voltage 
from which the stage will be run.   Grid leak as usual.   

Hook the other end of the coil to the plate of the tube directly.   
(Coil is at full plate voltage.)   Feed the plate through a resistor of 
100k, more or less according to power output requirements.   Any 
of the usual medium-mu triodes will work.

Output may be taken from the plate through a few mmf.

If the BFO must be adjustable over a few kcs from the front panel 
connect a suitable variable cap from either end of the coil to ground.
Tuning range fraction will be roughly *half* the ratio to which ever of 
the two tank caps it shunts times the ratio of 1/reactance of the 
shunted cap to the total.   This tuning cap must also handle full 
high voltage.

For example a 15 mmf variable across the 150 mmf fixed cap 
would give:

1/2 * 15/150 * 114/150 = 3.8% tuning range or about 3.8 kcs.   

This is also an excellent VFO circuit because the cathode of a 
tube is the squirreliest electrode of them all -- think of a poor 
quality capacitor between the cathode and the filament shunted by 
a poor quality diode with a low breakdown voltage rating, both with 
very high temperature coefficients -- and the filament usually runs 
on AC.   All the usual hot-cathode VFO circuits *will* have 60 cps 
FM at VHF.

Using modern tubes you can isolate the tube from the tank circuit 
well enough that for HF service you can live with the crummy 
cathode.  With lower gain tubes it's well to consider putting the 
filament at cathode voltage for RF -- the command transmitters ran 
the oscillator filament through parallel windings on the VFO coil, 
did they not?   

When using the grounded cathode Colpitts in a high-stability VFO, 
consider tapping the coil to feed the plate voltage.   (Floating tap -- 
not bypassed.)  That way you don't have the plate feed resistor 
across part of the tank circuit, lowering the Q.   In a 100 kcs BFO 
this is not an issue because the Q is already low enough that 
realistic plate feed resistors will make no difference.

In theory a choke is the answer but in high-stability VFO practice 
chokes are to be avoided because they have the worst temperature 
coefficient of all parts.

Walt 
KJ4KV