All,
The derivation of the IBIS measurement levels and IV and VT curves are
relative to "Node 0" because the measurements (or simulations) are done in
an environment where the local ground is the same as the Test Fixture
ground voltage which is "0.0 V' or Node 0 in many simulators.
However it is critical that when simulators apply these IV and VT curves
during simulation, and use the measurement rules to analyze waveforms,
these IBIS measurement levels and IV and VT curves are relative to the
local reference node of the buffer which is typically the voltage node at
the Pulldown Reference or Ground Clamp reference of the device. This needs
to be carefully worded to deal with cases when the pulldown reference and
ground clamp reference is not 0.0V when the IBIS data is created from test
fixture measurements or SPICE to IBIS simulations.
There is no place in IBIS that states that when a buffer is hooked up to
floating power and ground supplies, that the IBIS measurement levels and
IV and VT tables are to applied to voltage levels relative to simulator
Node 0. Any reference to measurements or connections to Node 0 (or 0.0
Volts) need to be corrected in the spec (other than in data derivation
sections).
I may have not worded this correctly, but based on the comments in today's
meeting Mentor, Keysight and SiSoft agree to this. And I expect all of the
EDA companies will agree to this interpretation just as all of the EDA
companies have agree that C_comp should not connect to Node 0, but the
local reference node of the buffer.
Besides Bob, does anyone disagree with these statements? We cannot do
anything with respect to cleaning up Ground in the IBIS spec without
accepting this as the correct intent of IBIS.
Walter
Walter Katz
<mailto:wkatz@xxxxxxxxxx> wkatz@xxxxxxxxxx
Phone 303.449-2308
Mobile 303.335-6156
