|
发表于 2009-2-21 16:35:49
|
显示全部楼层
1. Rollett showed that the , 0< 1 S L Γ < condition corresponds to a parameter known as the
“k-factor” being greater than unity, where
. So
in practical terms, if k >1 the device will never display an input or output reflecting
coefficient magnitude which is greater than unity, no matter what passive matching may be
placed at its input or output.
2. When the signal gain drops below 10dB, the extra RF drive power required will often cancel
out any efficiency advantage that had been carefully designed. The upshot of this is that one is
often looking for an optimum situation where k-factor is greater than unity, but not too much
greater. Devices with high k-factor also tend to have low gain, and some extra gain can be
retrieved by allowing positive feedback around the device, while keeping the k-factor above
unity.
3. Something about the k-factor which are worth noting:
(1) Any device which has a k-factor greater than, but not much greater than, unity displays a
more aggressive gain/match characteristic than a theoretical unilateral device. In
particular, the final MAG may be considerably higher, in a nearly matched condition,
than a simple voltage standing wave ratio (VSWR) mismatch calculation would indicate.
For example, such a device displaying a 10dB return loss may show more than the
calculated 0.7dB increase in gain when finally matched to -20dB return loss.
(2) Circuit loss can play havoc with the k-factor, and especially the frequency where it
crosses unity. In practice, devices can be safely used some way below the unity k-factor
point if the k-factor is based on fully de-embedded s-parameter measurements.
(3) The circuit environment in which a transistor is placed can modify significantly itseffective s-parameters, and especially the critical reverse transmission parameter, S 12.
This is probably the main cause of unexpected, or unsimulatable, stability problems. |
|