Hi KlausST
Thanks for your reply.
Any PLL circuit will have loop filter elements? such as capacitors and resistors, and we do use ceramic type for the purpose.
Can you suggest some differnt capacitor type, which won't show sensitivity to vibrations and shocks?
We see phasenoise degradation in the range of few KHz near the locked carrier during random vibrations.
You may do a forum search about piezo effect of capacitors...
And you may do an internet search.
Basically you may find informations like
* what capacitor type suffer from piezo effect ( and a ranking)
* which type of ceramics has least / worst piezo effect
* how to reduce mechanical stress
I'm not talking about loop filter circuit only. There may be capacitors at the VCO and for power supply decoupling ...and so on...
"A few kHz" is not very informative.
An experienced circuit designer may give you more detailed answers if you give more details about your circuit.
Like: schematic, PCBlayout, BOM (including manufacturer and ordering number), frequencies ....
More specifically, all high Er dielectrics (e.g. X5R, X7R) can be expected to show piezoelectric effect. PLL loop filters and possibly power supply filters should be checked. Also oscillator coils can be vibration sensitive.
I think that before any other LC passive, the first that will be affected by vibrations is the LO resonator itself.
Check first the signal of the LO crystal that generate the reference frequency of the PLL.
I think that before any other LC passive, the first that will be affected by vibrations is the LO resonator itself.
Check first the signal of the LO crystal that generate the reference frequency of the PLL.
You are right, FvM, first thought that came to our mind was crystal oscillator, but we removed it from PLL module and fed reference from bench signal source(which was not subjected to vibration),the phase noise degradation signature of locked carrier still repeated, that is why we are looking for other components, which can respond to mechanical vibrations.
I think that before any other LC passive, the first that will be affected by vibrations is the LO resonator itself.
Check first the signal of the LO crystal that generate the reference frequency of the PLL.
AND, of course, the Crystal Reference oscillator, and the PLL filter parameters (need an open loop bandwidth much higher than the highest frequency of the mechanical vibration).
if the crystal reference oscillator moves in frequency with vibration, the VCO output will move 10,000 times as much when phase locked.
You are right, FvM, first thought that came to our mind was crystal oscillator, but we removed it from PLL module and fed reference from bench signal source(which was not subjected to vibration),the phase noise degradation signature of locked carrier still repeated, that is why we are looking for other components, which can respond to mechanical vibrations.
is your PLL control loop too narrow in bandwidth?
the PLL needs to be able to correct for the VCO wandering around due to mechanical vibrtion. that means it needs bandwidth to see the mechanically induced errors, and enough open loop gain to mitigate the effect
Make sure your connecting cables are not introducing phase changes under vibration. Testing for oscillator phase noise under vibration requires a lot of care to be taken in all aspects of the test set up. Also check that the vibration table is not inducing EMI into the system.
It would be useful to know what oscillator and what phase noise level is discussed here.
A friend is designing ultra low noise dielectric resonator oscillators and he can see vibration from fan rotation in measurement instruments that are some meters away in the lab ... in that case it's the resonator itself that catches tiny amount of vibration through the floor...