Wow, much to read after a few hours. Thanks to all for the useful
informations!
One more question: Currently i use the BF862 since it was on hand.
However i could use a normal BF245 instead. I thought it could have an
advantage since the housing is bigger and so the temperature differences
should be smaller. Also the BF862 has a relative high Id, so the power
dissipation and hence the temperatures may be higher. Vdd=5V and the osc
didn't start to oscillate when using a source resistor of 100 Ohm. When
using 50 Ohm then it runs and the DC source voltage is 0.4V, i.e. 8 mA
DC drain current. Could be a bit lower.
There are several solutions to have a VFO but now i started this analog
version and i want to spend some more time to make it work reasonable
with acceptable drift. Once it runs stable (and i know it CAN be
achieved since my 137 kHz analog VFO is stable too) and the scale is
labeled from 472...479 kHz, all will be fine! I don't need a 1 Hz
stability for CW. Even a 10 Hz drift is acceptable. Fine adjustments can
be done by listening to my own signal in the RX...
All right about the Inductor. I would have choosen a T50-2 instead but i
only have the small FT cores available to time. So i'm looking forward
to see the difference of the drift soon.
Am 06.07.2012 01:45, schrieb James Moritz:
Dear Stefan, LF Group,
In my experience, the oscillator circuit topology is usually less
important than the component values used, the type of components,
setting the bias current in the transistor, etc.
It is a bit disapointing. The oscillator is drifting several 100 Hz
when looking to the divide by 4 side, i.e. the PA output i.e. close
to 475 kHz.
I used 9 turns on a FT37-43.
I suspect this inductor may be the problem. The temperature
coefficient of the permeability of ferrite is high, so the inductance
will drift a lot with temperature. Also, the Q of ferrite cored
inductors is usually rather low, with the effect of requiring higher
transistor gain to make it oscillate, and greater loading of the
resonant circuit by the transistor as a result. I suggest using a
Txx-2 iron dust core insted of the -43 ferrite. It will need many more
turns, but the inductance will be much more stable and the Q much
higher. The higher Q will enable you to reduce the bias current in the
oscillator transistor, and/or increase the shunting capacitor values.
This will reduce the loading caused by the transistor.
Another way to increase stability is to reduce the operating frequency
- an oscillator at 475kHz or 950kHz will tend to be more stable
because the larger fixed capacitance values will mean the junction
capacitance of the transistor will have less effect.
I need the 1900 kHz operating frequency though since the PA is designed
by using the /4 divider, allowing the use of a HF TRX to drive the PA.
Thanks for your help.
73, Stefan/DK7FC
Cheers, Jim Moritz
73 de M0BMU
|
