[time-nuts] xtal oscillator phase noise

Bruce Griffiths bruce.griffiths at xtra.co.nz
Tue Jan 1 18:06:20 EST 2008


Henk ten Pierick wrote:
> Bruce,
>
> 1.  Best wishes.
>
> 2. It has taken some time but here is the schematic.
>
> Henk
>
Henk

The circuit diagram helps a lot.

1) Since the crystal current is 1mA the RF voltage across R2 (and the RF
at the input of the HC04) is 220mV rms (622mV pp).
This is a little low as HC04's are a little noisy a higher input voltage
slew rate should help. Ideally about 1.75vrms or so (just avoids
clipping by the HC04 input protection diodes) should significantly
reduce the noise contribution from the HC04.

2) Emitter follower Q4 has a relatively low collector base voltage which
increases the collector base capacitance and associated phase noise. AC
coupling the emitter follower would allow a much higher collector base
voltage, and an inductor in series with the emitter resistor will allow
the emitter to swing below ground.

3) Your AGC circuit  seems to control the  feedback voltage/current at
the oscillator transistor base. An AGC in a crystal oscillator is
typically used to control the crystal current, you circuit doesnt appear
to do this directly.

4) The emitter follower will not drive a 50 ohm load without severe
distortion unless the emitter current is increased. Its not just the
output impedance (with zero input) that is important.

5) As Rick stated a common base buffer circuit would have better
performance. A common base circuit has no difficulty when driving a 50
ohm load. A step down transformer between the common base collector and
the load increases the RF voltage across the load.

6) A good buried zener has lower noise than the ADR445. The noise specs
for the ADR445 are somewhat sparse in that the spectral distribution of
the noise isnt given. One can always reduce the noise by averaging the
output of several references

7) Unless you need a tuned circuit to supress oscillation at unwanted
crystal modes, then using a circuit without one will provide better
temperature stability. If the tuned circuit uses ferrites then they may
increase the flicker phase noise.

8) If you do need a tuned circuit to suppress unwanted crystal modes
then one (there are many) of the Driscoll crystal oscillators is a good
solution.

9) The minimum ft of Q1 is a little low.

I will create some circuit schematics for crystal oscillators that
control the crystal current more directly and use a common base output
buffer.

Meanwhile look at:

http://www.ko4bb.com/~bruce/LPND.html
<http://www.ko4bb.com/%7Ebruce/LPND.html>

This crystal oscillator circuit is a variant of the low noise oscillator
proposed by Wenzel for use with fundamental crystals. It uses a common
base buffer and demonstrates several low noise biasing techniques that
can be used.

The 2nd and third references above the oscillator schematic explain the
mechanisms for generating AM and PM noise in a BJT RF amplifier. The
derivation is quite mathematical (statistics and calculus) but the
conclusions are relatively simple.

To drive an HC04 the common base buffer can use a load consisting of an
inductor shunted by a resistor to develop the drive. The shunt inductor
reduces the dc gain (from base to collector) of the  buffer stage and
hence the low frequency noise voltage developed across the collector
base capacitance. Such noise voltages modulate the output capacitance
and hence the phase shift of the buffer, increasing the buffer phase
noise.  Nonlinearities such as hfe variation with current tend to
increase the buffer output AM noise not the PM noise.

Bruce




More information about the time-nuts mailing list