[time-nuts] Power Supply Noise Affects Thunderbolt 1 PPS
ed breya
eb at telight.com
Tue Feb 28 04:47:21 UTC 2012
The PNP bipolar and P-channel MOSFET architectures do provide the
best low-dropout performance, but as I understand, do not provide the
best HF line rejection. Looking at the overall circuit - a high gain,
band-limited amplifier driving a "P" pass device puts it in a
common-base (or -gate) mode, with respect to the amplifier's output
port. It provides excellent DC performance, but high frequency noise
(beyond the control loop bandwidth) at the emitter (or source) has a
large voltage gain available. This is of course for positive supplies.
An "N" device operates as a follower, and needs more overhead voltage
to activate the gain as a common-collector (or -drain) amplifier, but
is able to reject the input supply noise much more effectively, given
the same band-limited control loop amplifier.
So, for best HF noise performance where the input noise may be large,
it's best to use a follower or shunt regulator topology, despite the
lower efficiency - unless efficiency is more important.
Ed
More information about the time-nuts
mailing list