BIRD19 Comments

Stephen, John and Ibis Committee:

Thank you for your comments. There is a strong consistency rational for
at least providing V_fixture_min and V_fixture_max. The [Ramp] dV/dt_f
data is defined with a measurement setup with a 50 Ohm load tied to
"POWER". For decomposition purposes, this should be the same set of
voltages (typ-min-max) to which the device is connected. I believe
this is how models are currently produced. The [Falling Waveform]
typ-min-max tables should be capable of providing a corresponding, more
detailed description of the SAME waveforms. V_fixture_min and
V_fixture_max values allow defining the same voltage variations that
the [Ramp] setup uses.

Stephen, regarding Vref_min and Vref_max, I have given it more thought and
this portion of the BIRD may not be justified because most timing specs
are at one voltage. Even for Open_drain cases, the Vcc is assume fixed
for min/max timing values. The most compelling consideration is that,
for consistency, other parameters would have to be allowed to vary (e.g.,
Vmeas, Vinh, and Vinl which are technically a percentage of Vcc in CMOS
technologies), and this would create an unnecessary complication in IBIS.
So, unless there are objections in the next few days, I will issue a
revised BIRD19.1 next week which deletes the Vref_min and Vref_max
changes and deals only with V_fixture_min and V_fixture_max.

Bob Ross,
Interconnectix, Inc.
Received on Wed Aug 10 19:50:17 1994