Colin McAndrew
Guest
|
Re: itzf contribution to vco phase noise
Reply #1 - Dec 2nd, 2002, 10:35am
Ryan, the transport current in VBIC is broken into two parts, the forward current and the reverse current. This is done because the excess phase shift is only applied to the forward current and not the reverse current. Practically it should not matter much as the reverse current is negligible under normal operation (except when you saturate the device). Itzf is the un-phase shifted value, when the phase shift is applied it becomes Itxf, the It means transport current, the final "f" means forward (the part controlled mainly by Vbe, not Vbc), and the "z" means zero-phase and "x" means with the excess phase shift applied. The phase shift is not done as an exact linearly frequency dependent shift to gm, as is sometimes done, as that is only implementable in the frequency domain. Rather, a 2nd order "filter" is used to implement the shift, and for small phase shifts it approximates a linear phase shift, with a small change in magnitude of gm. For the shot noise, for Ic it is calculated from Itzf. This will pass through the excess-phase filter and appear in the collector current. I am not sure what simulator you are using and what it is printing for noise. What you see at the terminals is not just what a simple analysis gives, the noise sources are placed in parallel with the elements they generate noise for, so what you see at the output is a function of the complete small-signal model of the device and the circuitry it is connected to, and of all the noise generators in the model. The shot noise is often an insignificant contributor, the base resistance is often the dominant noise source (assuming you are above the 1/f noise corner). So internally to the model the noise source associated with Itzf is the normal shot noise, however if you are looking at the output noise it is affected by more than just this noise term. If the simulator prints out the individual noise sources, as some do, then you should see the Itxf component track with Ic (provided you are out of saturation), but if you are looking at the noise voltage across a load the device is driving it will not simply scale with Itzf as (a) other noise sources contribute, and (b) it depends on the whole small-signal network. Hope this helps. Colin
|