Author: 71436.1314@compuserve.com at SMTPGATE Date: 10/28/94 1:13 AM Priority: Normal TO: ibis@vhdl.org at SMTPGATE BCC: Derrick Duehren at JFCCM2 Subject: Resending Bird20, from Kellee Crisafulli ------------------------------- Message Contents ------------------------------- ******************************************************************************* ******************************************************************************* Buffer Issue Resolution Document (BIRD) BIRD ID#: 20.0 ISSUE TITLE: Error correction regarding monotonicity statement in V2.1 IBIS specification. REQUESTOR: Kellee Crisafulli, HyperLynx Inc. DATE SUBMITTED: 10-10-94 DATE REVISED: TBD DATE ACCEPTED BY IBIS OPEN FORUM: TBD ******************************************************************************* ******************************************************************************* STATEMENT OF THE ISSUE: There are two errors in the IBIS version 2.1 specification regarding monotonicity. During the meeting on 10-7-94 the IBIS forum agreed to change the specification to correct these errors. ******************************************************************************* STATEMENT OF THE RESOLVED SPECIFICATIONS: The following changes apply to version 2.1 of the specification and IBIS_CHK program. *************************************************************************** Change 1- Delete the note about 'non-monotonic in one axis' *************************************************************************** In the section defining the keywords [Pulldown], [Pullup], [GND Clamp], [POWER Clamp] and following the paragraph titled "Monotonicity Requirements" the following note is present: | Note that the above conditions allow the data to be | non-monotonic in one axis. The above two lines shall be deleted from the specification. *************************************************************************** Change 2- Add two paragraphs from BIRD 8.2 *************************************************************************** Add the following two paragraphs where the sentence from change 1 above is removed. This is not an error or even a warning. It was down-graded to a simple NOTE: in BIRD 8.2 * * The IBIS_CHK program will test for non-monotonic data and * provide a maximum of one note per V/I table if non-montonic * data is found. For example: * "NOTE: Line 300, Pulldown V/I table for model DC040403 is * non-monotonic! Most simulators will filter this data * to remove the non-monotonic data." * * It is also recognized that the some data may be monotonic if * currents from both the output stage and the clamp diode are * added together as most simulators do. To limit the complexity * of the version 2.0 IBIS_CHK program, it will consider only one * V/I table at a time monotonicity testing. ******************************************************************************* ANALYSIS PATH/DATA THAT LED TO SPECIFICATION: The problems were discovered while the new IBIS_CHK V2.x program was being written. The problems were reported to the IBIS forum at the 10-7-94 meeting by Arpad of Intel. ******************************************************************************* ANY OTHER BACKGROUND INFORMATION: Bird 8.2 address's this issue.