Peter O. Brackett wrote...
James Early, RIP.
And besides the well known "Effect"... James Early also has a "Voltage"
named after him, the Early Voltage "Va".
I have always admired the way in which the virtual voltage "Va", in much
the same manner as "intermodulation intercept points" do in another field,
summarizes succinctly in a single number [Va] such a complicated "Effect".
Indeed. Speaking of succinct summaries using the Early voltage, there's
an interesting transistor-evaluation parameter used in IC processes, the
"beta-Va" product figure-of-merit. It's related to the fact that, other
things being equal, higher beta (narrow base width) forces a lower Early
voltage (high base-modulation effects), so that if an IC process can be
devised that allows both parameters to be high at the same time, it's a
better process. An early example of the issue would be Widlar's favorite
super-beta transistor, which had poor Early voltages and low breakdown
voltages, forcing the use of cascode stages wherever they were used.
Frank Goodenough of Electronics Design magazine used to write reports
on the beta-Va product for high-f_T transistors in new IC processes.
For example, in some cases the pnp transistors suffered especially-poor
beta-Va products. And advanced SiGe HBT transistors did very well.
Low Early-voltage transistors force designers to use cascode stages and
other space-consuming tricks to improve current sources and raise the
impedance of amplifier-stage collector outputs. These are important
issues for high gain, good linearity and improved accuracy.
Perhaps Jim can fill us in a bit on this convoluted scene.
BTW, I haven't been able to trace the original paper defining the Early
voltage, and calling it Va. I have Early's early papers
>) and, IIRC,
it's not there directly, so it must have come later, perhaps contributed
by another person or by an evolving industry convention?
Thanks,
- Win
whill_at_picovolt-dot-com
