I am designing a VCO and find that the phase noise from the biasing circuit contribute almost the same big phase noise. The biasing circuit include some current mirrors to supply the tail current source bias.
Anyone has advises to decrease the biasing circuit phase noise? Why reference papers only talk about the simplest biasing such as one current mirror to supply the tail current bias?
tail current is the most important point in vco noise...solution:
1.try to decrease its swing by changing it current level.
2.use a parallel capacitor to fix it voltage swing
3.change W/L of both current mirror transistor to decrease produced noise level
the two books below are good reference books for VCO designer, and there are serveral methods to surppress the up-convert Tail current source noise.
1-. Hegazi, J. Rael, A. Abidi “The designer’s guide to high-purity oscillator”
2-A. Hajimiri, T. H. Lee, “A General Theory of Phase Noise in Electrical Oscillators” J. Solid-State Circuits, vol. 33, No. 2, Feb.1998
There's an option to display the fourier transform frequency spectrum in HSPICE, where the output spectrum in dB is displayed, from the definition of phase noise of the ratio of the output power divided by the noise power at a specified value, phase noise could be computed, or as proposed by Hajimiri tha ISF(impulse sensitivity function), NMF(noise modulating function) in order to compute the ISF effective could be done to look into the noise response of the circuit to a perturbations, I'm still working on executing the ISF and NMF function, and do welcome further comments on this method and how to go about doing it.
There are couple of methods you can use to suppress the bias noise.
First, you can use bias transistor of large W/L ratio but maintain the same bias current to suppress the device flicker noise, which will be upconverted to the 1/f^3 phase noise region.
Then, you can put a simple LC filter oscillating at the second harmoic to the drain node of the bias transistor to preserve the loaded Q of LC tank and further suppress the bias noise.
And you may want to tweak the size of cross-coupled transistor to increase its linear range and tune up the AM-PM noise upconversion factor.
...
Don't know what a tail current is, but I do know that a current mirror is an open loop device. That is, the current is (on the average) equal in the two arms because the two transistors are nearly identical and they are forced to have the same Vbe, hence they should have the same DC current. But if there is any sort of noise in either of the transistors, it will not be suppressed at all. (No offense to Mr. Widlar).
I would suggest some sort of feedback op-amp circuit that actually measures the current flowing into your oscillator, and can respond quickly to low frequency variation in that current with negative feedback.
You can check this out:
**broken link removed**
You also could try a series inductor, but at these low frequencies, it would probably bee too big for a modern circuit card.
tail current is the most important point in vco noise...solution:
1.try to decrease its swing by changing it current level.
2.use a parallel capacitor to fix it voltage swing
3.change W/L of both current mirror transistor to decrease produced noise level
Utilize RC filter to remove high freq noise, and the low freq noise can be removed
by PLL loop.
By the way,large RC filter may cause slower power up procedure.