[ARC5] Test Oscillator - Lambda conficuration.

[Richard Knoppow]

----- Original Message ----- 
From: "Leslie Smith" <vk2bcu at operamail.com>
To: "ARC-5 List" <arc5 at mailman.qth.net>
Sent: Friday, November 04, 2011 11:19 AM
Subject: [ARC5] Test Oscillator - Lambda conficuration.


> Hello List.
> About one month earlier some-one posted an inquiry about 
> oscillators
> suitable for use as a VFO.
>
> I have been working on a circuit based on the Lambda 
> (negative
> resistance oscillator) configuration, and in particular 
> the frequency
> stability of this circuit.
>
> I have a circuit of the prototype available for any-one 
> interested.
> The interest in this circuit, setting it apart from 
> conventional VFO
> circuits (Clapp, Vackar & co) comes from the wide range of 
> oscillation
> without change to component values.  Specifically I have 
> had my test
> circuits oscillating from audio (eg 850Hz) to mid HF (eg 
> 10MHz) with the
> only change being to swap the tank circuit coil and 
> capacitor.  The
> circuit does not use a tapped coil and has no feedback 
> capacitor
> network.  One side of the tuning components is tied to the 
> common rail.
>
> Stability.  The circuit drifts at the rate of 315 Hz per 
> degree C, using
> the components in the test configuration.  I changed the 
> coefficient of
> drift from negative to positive by adding a 22pF capacitor 
> across the
> main tank capacitor.   A 7pF N750 compensating capacitor 
> would give a
> lower drift, although I've not confirmed this by 
> measurement yet.
>
> At constant temperature (19 C) the circuit on my bench, 
> beside me as I
> write this, has drifted 29 Hz in the last 31 minutes - at 
> 5.5 MHz
> (approx).
>
> In the longer term I have several tasks for this 
> oscillator, one being a
> test and alignment oscillator for the ARC-5 sets.  While I 
> get lower
> drift figures and do a little more work, if anyone is 
> interested in this
> circuit (and drift figures) send an e-mail and I'll reply 
> with a
> circuit. (Format: png)
>
>
>
>  Leslie Smith
>  vk2bcu at operamail.com
>
      It seems to me that there are several factors relating 
to stability of oscillators:
1. Electrical: the circuit, whatever it is, should confer 
reasonable immunity to supply voltages (including filament 
voltage in vacuum tube oscillators) and should give the 
lowest loading of the tuned circuit. In general, the less 
power dissipated by the oscillator the better.
2, and 3, together because they are related: Mechanical and 
thermal. The parts affecting frequency should be as stable 
as possible both mechanically and as free from change from 
temperature variation as possible. I think a lot of designs 
rely on some sort of compensation for the latter but I think 
probably variation should be minimised without the 
compensation before its applied. The problem with 
compensation is that its usually possible to make it linear 
only over some limited range outside of which it can 
actually make the drift worse. Of course, one can resort to 
the extreme of enclosing the frequency determining 
components in a regulated oven and this is done in in many 
commercial designs, but even with the oven the choice of 
materials and configuration of coils and condensers can be 
made to minimise inherent thermal drift. I think this is not 
a trivial excercize and I offer no magic solutions. The 
literature, both amateur and commercial, is full of proposed 
methods of stabilizing oscillators and descriptions of 
practical applications are plentiful.
     This is not to say that a Dynatron, or negative 
resistance circuit is not good, it has some definite 
advantages, but by itself is only one step in achieving 
stability.
     Good luck and success with your project:-)



--
Richard Knoppow
Los Angeles
WB6KBL
dickburk at ix.netcom.com