crystal Quality factor

if crystal is connected between the drains of cross coupled mosfets,
the crystal Q will decrease due to loading of the mosfets
Q = wl/R , now do i need to make the Z across xtal too high to prevent loading on xtal ?
but they are parallel (r xtal // Z ct) that always leads to r< rxtal
iam confused about the in ct. Q , can any body explain to me how to get max in ct. Q
how can i draw the in ct. Q on spectre too

The circuit that is attached to the crystal should display enough 'negative' resistance at the frequency of interest to wipe out the loss factor in the crystal. Simulations have to be done to ensure this with some safety factor. Loaded and unloaded Q does not make sense here.

yes. thank u
may i ask why the loaded and unloaded Q doesn't make sense here ?
i think it do , as they invented the term loaded and unloaded Q ?
can u explain why u get to this conclusion ?

"i think it do , as they invented the term loaded and unloaded Q ?" The power transfer in a tuned circuit equals (Qu -Ql)/ Qu . This tells us for a low loss, the Qu must be as high as possible and the Ql as low as possible. For the output circuits in transmitters you use a value of Qu = 100 and a loaded QL of 10, so the loses are 10%, but unwanted harmonics are still filtered out by the working Q of 10. If the working Q is the same as the unloaded Q then the tuned circuit transfers no power at all.!!
Frank

thank u chuckey
however these values are not for crystals , crystals Q is in range of 10^4 to 10^5 according to the harmonic u r working on
i understand that there should be some loss , but how to minimize it
the crsytal is modeled as L, C, R in series these are // to C (5p) al this will be added to a certain ct.
should the Z across crystal be very high in order not to decrease the Q ?
and it is not about the harmonics , it is about the stability spec (+-1.5 ppm) that i need a very high in ct. Q

ali kotb said:

yes. thank u
may i ask why the loaded and unloaded Q doesn't make sense here ?
i think it do , as they invented the term loaded and unloaded Q ?
can u explain why u get to this conclusion ?

Click to expand...

Of course loaded and unloaded Q are important and play significant role in RF matching networks, but I only said that these dont mean much in this situation. The circuit that adds to the crystal makes it sustain the oscillations, so the Q becomes inf, isnt it.

saro_k_82 said:

Of course loaded and unloaded Q are important and play significant role in RF matching networks, but I only said that these dont mean much in this situation. The circuit that adds to the crystal makes it sustain the oscillations, so the Q becomes inf, isnt it.

Click to expand...

sorry , can u explain more about Q being infinity ??
the ct that adds to xtal initiate osc. and due to non linearities it sustains.
i think it de-Q the crystal heavily.

If you have a tuned circuit and you apply a narrow pulse to it, it "rings", that is to say it produces a series of sine waves at its resonant frequency, each cycle being 1/Ql less then the preceeding one. So a low Ql circuit produces , say 10 cycles, a higher Ql one, 100 cycles and in an oscillator an infinite number of cycles because there are no losses, i.e. a Q of infinity.
I think you are getting mixed up between the mechanical Q of a crystal, set by how it was manufactured and its electrical equivalent used for modeling.
If you are really worried about the loading on the crystal, build your self a FET multivibrator with very small coupling capacitors, split one into two in series and put your crystal from the split point to earth, this will give you virtually no loading (10M ohm +10pF).
Frank