All,
I would like to propose the following terminology:
An I/O buffer can have either 2, 3, 4 or 5 rail connections as follows:
1. "Power Rails" (one or two)
a. pullup_ref
b. power_clamp_ref
c. pullup_ref and power_clamp_ref
2. "Reference Rails" (one or two)
a. pulldown_ref
b. gnd_clamp_ref
c. pulldown_ref and gnd_clamp_ref
3. "External Rails" (zero or one)
a. None
b. ext_ref
For devices under test (DUT), there shall be a fixture reference node, all
IV and VT data shall be measured relative to this reference node. When
making measurements of a DUT in a test fixture, the values of the voltages
relative to the fixture reference node shall be specified by the following
table:
Rail [Model]
subparameter
pullup_ref [Pullup Reference]
power_clamp_ref [POWER Clamp Reference]
pulldown_ref [Pulldown Reference]
gnd_clamp_ref [GND Clamp Reference]
[Pulldown Reference] and [GND Clamp Reference] shall default to 0.0V if
not specified.
[Pullup Reference] and [POWER Clamp Reference] shall default to [Voltage
Range] if not specified.
There is no concept of global ground or local ground. There is a concept
of Test Fixture Ground for DUT.
There are three cases of interest.
1. One of the DUT "Reference Rails" is 0.0V.
a. There is almost assuredly a GND pin that the test fixture can
connect the test fixture GND.
2. There is a pin connected to a "Reference Rail", but is DUT value
is not 0.0V.
a. The test fixture can create and supply all of the DUT rail
voltages with values relative to the test fixture GND.
3. There are no pins connected to the "Reference Rails", but the
component has pins with other rail voltages that can be connected to the
test fixture. In this case, the model maker must make a leap of faith that
the chip will generate internally voltages relative the rail voltages that
are connected to a pin.
A_Gnd is a simulator reference node for the digital stimulus and enable
functions that is a reference for Multi-Lingual pseudo digital control
signals such as the stimulus input and enable lines.
All occurrences of the earth ground symbol in the specification where
errors introduced when GND which was a signal_name in the 5.0
specification was incorrectly changed to the earth GND symbol when the
graphics were created for 5.1.
When doing Spice2IBIS a simulator can choose Node 0 to be the "test
fixture reference node".
I would therefore not use names like Puref or Puref, but use names we
already have in the specification for the rail terminals of a buffer,
either
pullup_ref
pulldown_ref
power_clamp_ref
gnd_clamp_ref
ext_ref
Or
Pullup_ref
Pulldown_ref
Power_clamp_ref
Gnd_clamp_ref
Ext_ref
Walter
-----Original Message-----
From: ibis-editorial-bounce@xxxxxxxxxxxxx
[mailto:ibis-editorial-bounce@xxxxxxxxxxxxx] On Behalf Of Bob Ross
Sent: Monday, March 7, 2016 1:00 PM
To: IBIS-Interconnect <ibis-interconn@xxxxxxxxxxxxx>;
ibis-editorial@xxxxxxxxxxxxx
Subject: [ibis-editorial] IBIS Terminal Names
All,
Attached is a list of analog reference terminal names in IBIS, in the
proposed Interconnect BIRD, and in the B-element.
Our challenge is to choose consistent buffer model terminology.
Bob
--
Bob Ross
Teraspeed Labs
<http://www.teraspeedlabs.com> http://www.teraspeedlabs.com
<mailto:bob@xxxxxxxxxxxxxxxxx> bob@xxxxxxxxxxxxxxxxx
Direct: 503-246-8048
Office: 971-279-5325