Power Considerations
Last Edit July 22, 2001
CMOS and BiCMOS Arrays
BiCMOS arrays will simplify the computation since they will only compute
DC power for the bipolar interface macros. Note that BiCMOS arrays still
have an overhead current
CMOS arrays emphasize AC power components and do not usually have DC
components listed
When Power is Specified Rather then Current
Another variation in the computational method occurs when the macros
are individually specified with a power dissipation, i.e., use this macro
- dissipate this much power. The worst-case multiplier and the worst-case
voltage may or may not be accounted for in the computation. Only addition
is required to compute the macro power dissipation
An overhead component for the bias circuitry may need to be computed
and added to the macro sum. If typical power is specified, a worst-case
multiplier or adjustment factor may be provided as previously discussed
Design Rules when Power is to be Estimated
- Before trying to estimate power, verify that the specifications made
for the macros are clearly understood
- Review the overhead (bias) current and how it is handled by the array
- Review the procedures specified by the particular array vendor for
that array series since procedures may vary from series to series with
the same vendor
- Review other vendor-identified power dissipaters - TTL IOEF, ECL static
output
AC Macro Power Dissipation
To compute the AC power for any array, perform the computation shown
earlier and repeated here:
PAC = 0.20 * ( a * f * G )
where
- a is the power constant in microwatts/gate-MHz,
- f is the switching frequency
- G is the number of gates.
There will be some variation in the form of the equation depending on
the use of gates or macros or outputs as the sizing measure. There will
be variation depending on the use of a register ratio and there will be
variation in the types of macros for which AC power will be computed.
|