Mig said:
I encountered with very strange problem with a ceramic crystal
oscillator.
It took a long time to recreate the problem.
When a power supply voltage is turned on first time there is no output
frequency. When the power supply is turned off and on again the
frequency appears.
You can see it on the video here.
Any suggestions?
The vendor doesn't know why it happens.
Startup problems are as common as dirt.
Most oscillators work in a large-signal regime where the active devices
are nonlinear. Thus the average gain, impedance level, and feedback are
all different when the oscillator is running, compared with the same
circuit in a quiescent state.
The feedback loop may well be stable at small-signal conditions and only
oscillatory at large signal conditions. If the power supply transient
is quicker than the bias time constant of the oscillator, it'll find
itself in a large-signal condition at power-up, which will get the
oscillator going and mask the startup problem. It's really important to
test oscillators under a wide variety of turn-on conditions, including
very slow ramps.
It's also quite possible for oscillators to have too much feedback, in
which case you get all sorts of distortion and self-modulation rather
than a nice CW sine wave. ALC oscillators adjust their loop gain based
on oscillation amplitude, and thus tend to avoid both kinds of problem
(as well as being much quieter).
I gather that your power supply is producing different turn-on
transients, depending on how long it has been switched off. From the
symptoms, it appears that when it's been off long enough for the filter
caps to have discharged, the turn-on is slower than when you turn it off
and then on again, and the difference is enough to expose the startup
problem.
Cheers,
Phil Hobbs