[Central-Electronics] 20A drift: new one just like the old one

[k6jek]

I upgraded my very early 20a to the newer 6U8 oscillator circuit from the old 12AU7 circuit that was only used for three months.  Alas, the new circuit isn't much better than the old one for drift but at least it doesn't have the 50 Hz pull problem.  

Next I suppose I'll try some temperature compensating caps although with the old circuit, just putting a small fan on the outside left blowing in to the lower vent made a big difference, enough that my enthusiasm for fooling with caps is diminished.  

What was the 50 Hz pull problem?  If the CARRIER knob was turned beyond 8, the oscillator was pulled 50 Hz.  I would do just that to join a QSO -- switch to a dummy load, turn the carrier up, zero beat the QSO.  Pretty soon some honest soul would say "Jon you're 50 Hz high" (or maybe it was low). I would check by using the same procedure.  I'd be dead on.  They'd say, no 50 Hz high. How could that be?  It was making me crazy. I thought the crystal was jumping.  No the oscillator was being pulled by the carrier insertion.  No wonder they junked that circuit after three months.

I didn't figure out the mechanism of the oscillator pull.  Instead I just rebuilt to the 1955 schematic.  No more pull.

I am disappointed this version is no better than the last for drift.  I'm not surprised since others with newer transmitters also noticed the 9 MHz drift. Still I had hope.  At the same time that CE changed the oscillator, they put 33 pF N750 caps across the RF phase shift coils which must help the stability of the phase shift. They did not put any temperature compensation in the oscillator itself, however.

Some numbers.  After 5 minute warm-up measured every 10 minutes.  Regulated 115 VAC, 75˚ F controlled room temperature.  Drift in the last ten minutes

Time	Old	New
10		-9	-29
20		-46	-53
30		-46	-45
40		-38	-37
50		-32	-25
60		-26	-16
70		-21	-13
80		-17	-10
90		-12	-8
100		-10	-6
110		-9	-6
120		-5	-2	


Cheers,

Jon