All you can do with thermal drift is heat the assembly to a
temperature , where you can control the core temperature by ensuring
the thermal gradient is steep enough not to be affected by the
ambient changes and the parasitic heat input from the energised
components ... oven will take many times the oscillator wattage
Or place in a bath of insulating / heat conducting fluid , so the
opposite applies , enlarging the thermal heat sink and reducing the
gradients , reducing the affect of point heat source , will drift , but
more slowly ..
G..
--------------------------------------------------
From: "Stefan Schäfer" <[email protected]>
Sent: Saturday, July 21, 2012 6:32 PM
To: <[email protected]>
Subject: Re: LF: RE: Analog oscillators
MF,
Another episode about my analog oscillator for MF story.
After the hints from the LF group i have now collected a Amidon T50-6 iron
dust toroid to wind the coil for a free running analog oscillator. The
oscillator runs on 1900 kHz and is intended to drive my 472...479 kHz PA.
This PA has an internal divide by 4 IC inside, so it can alternatively be
driven from a conventional HF TRX on the 160m band range.
The idea to have a very small analog oscillator covering the 630m band
comes from the need for low power consumtion for /p MF test and volume
limitations in a backpack. My PA was designed to run on 12V DC directly
(https://dl.dropbox.com/u/19882028/MF/100W%20475kHz%20PA.png) and
especially to be very COMPACT. The efficiency is arround 85%, including
the driver power.
For normal CW operation, a drift of say 20 Hz in 5 minutes can be
accepted.
First i took a FT37-77 ferrite core to build the resonsnce coil. All the
capacitors are "Styroflex types". The drift of the oscillator was not
acceptable. See https://dl.dropbox.com/u/19882028/MF/Drift%20FT37-77.png
The new coil which uses about 60 turns of 0.22 mm diameter enameled Cu
wire produces this drift (same scroll rate and displayed range in the
spectrogram): https://dl.dropbox.com/u/19882028/MF/Drift%20T50-6.png
That looks much better :-)
After running the VFO for 15 minutes, the drift is stabilized to 2 Hz in
10 minutes! See
https://dl.dropbox.com/u/19882028/MF/Drift%20T50-6_Langzeitdrift.png
That is very satisfying and so the VFO which uses a 33pF variable
capacitor to cover 472...479 kHz is well suitable for CW!
Of course there will be a higher drift if the supply voltage is changing
or the ambient temperature. And i hope the RF will not affect the transmit
frequency significantly..
This evening i want to test the VFO on the band. Maybe there will be some
QSO partners? :-)
Thanks for the explanations about the temp drifting components. If you
have further useful suggestions, don't hesitate to tell them.
73, Stefan/DK7FC
Am 07.07.2012 23:11, schrieb Stefan Schäfer:
Hello Jim,
Am 07.07.2012 20:55, schrieb James Moritz:
When you replace the existing inductor with a high-Q one, you will find
that the loop gain of the circuit is increased due to the higher Q, and
you will be able to use a lower gain/reduced transistor bias current.
Give me some days. Soon i will report about the progress. A T50-6 os at
my parents QTH where i will be in some days.
For CW enthusiasts, I am suprised more use is not made of simple LC
VFOs. A drift of +/- 100Hz during the time taken for a QSO is usually
not noticeable,
Oh i would say 100 Hz is a bit hard but 20 Hz should be OK in my opinion.
and it is easy to achieve this in the LF/MF range. Also, the narrow
tuning range means an elaborate tuning mechanism with reduction gearing
is not needed - just a knob with a simple hand-marked scale is adequate.
EXACTLY!!!!
73, Stefan/DK7FC
It is much simpler than the schemes needed to obtain an LF/MF signal by
dividing/mixing down the output of an HF rig, etc. A VFO is also a much
better practical choice than a crystal oscillator, which leaves you
stuck on a fixed frequency.
Cheers, Jim Moritz
73 de M0BMU
|
