IBIS Macromodel Task Group Meeting date: 14 May 2019 Members (asterisk for those attending): ANSYS: Dan Dvorscak * Curtis Clark Cadence Design Systems: Ambrish Varma Ken Willis Intel: * Michael Mirmak Keysight Technologies: * Fangyi Rao * Radek Biernacki * Ming Yan Stephen Slater Maziar Farahmand Mentor, A Siemens Business: * Arpad Muranyi Micron Technology: * Randy Wolff * Justin Butterfield SiSoft (Mathworks): Walter Katz * Mike LaBonte SPISim: * Wei-hsing Huang Teraspeed Labs: * Bob Ross The meeting was led by Arpad Muranyi. Curtis Clark took the minutes. -------------------------------------------------------------------------------- Opens: - None. ------------- Review of ARs: - Fangyi to create a BIRD197.3 draft 4 to incorporate his DC_For_Statistical parameter. - Done. -------------------------- Call for patent disclosure: - None. ------------------------- Review of Meeting Minutes: Arpad asked for any comments or corrections to the minutes of the May 07 meeting. Mike L. moved to approve the minutes. Bob seconded the motion. There were no objections. ------------- New Discussion: BIRD197.3_draft_4(DC_Offset): Arpad shared the new draft Fangyi had sent out. Bob and Arpad noted that this draft should be draft 4 of 197.3, not 197.4 as shown in the draft itself. Bob asked that the BIRD NUMBER: field be corrected, and suggested the new entry in the BACKGROUND INFORMATION/HISTORY section be folded into the previous section labeled 197.3x. Fangyi reviewed the changes: - This version reverts DC_Offset back to Usage In only. - The physical waveform at the Rx input is equal to DC_Offset plus the waveform input to GetWave(). The waveform input to GetWave() is zero centered. The model can convert back to the single-ended signal internally if necessary. - The output of GetWave() can have a non-zero DC component. - The EDA tool does NOT add DC_Offset to the GetWave() output waveform. - Introduce a new DC_For_Statistical parameter. It is an Out parameter. It is the mid-point of the steady state high and low voltages at the Rx latch. This is provided so the EDA tool can produce the proper final DC offset when doing a statistical simulation. Randy restated his understanding of this proposal to make sure it was clear. If the model has defined DC_Offset, then the EDA tool is now expected to display the output of GetWave() without adding anything to it. Randy said this might be different than what some tools are currently doing. Arpad noted that the tools currently don't know anything about the DC offset. Randy agreed, and said that since the models currently don't say anything about it, different tools are deciding whether or not to add a DC offset to the output of GetWave(). Fangyi noted that this proposal clears up the ambiguity that currently exists. If DC_Offset is defined, then entry 2. in the Other Notes: section explicitly defines what to expect from the GetWave() output. Randy said that if the model wants statistical and bit-by-bit simulations to match, then the value returned for DC_For_Statistical should match the DC offset of the GetWave() output waveform. Fangyi agreed. Fangyi noted that Walter had replied with several comments. Bob asked Arpad to take the AR to forward Walter's response to the ATM list. The group agreed that no action was necessary on the comments at this time. Arpad asked about one comment and previous discussions about a new threshold being required. Fangyi and the group agreed that a new threshold parameter, similar in concept to the PAM4_XXXXThreshold(s), might be required given that the output of GetWave() could have a non-zero DC offset. Fangyi suggested a separate BIRD should be drafted to deal with a new threshold parameter and possibly define its relation to Rx_Receiver_Sensitivity. Arpad asked if this proposal was ready to be submitted to the Open Forum. Radek suggested it be posted to ATM and given another week for review. Bob took the AR to update Fangyi's latest draft with the revision corrections he had noted earlier. Jitter Amplification: Michael Mirmak noted that he is engaged in internal discussions with the experts on his team, and it will be some time before a BIRD is produced. Arpad asked if Michael wanted the topic to stay on the agenda. Michael said yes. Fangyi asked that Michael pay particular attention to defining how the tool would generate the time domain jitter sequence from the new parameters. He noted, for example, the way Tx_Rj and Tx_Dj specify how to generate the random sequence. C_comp improvements: Randy noted that he had recently made a few updates to the BIRD. He said we could review it next week. He noted that he was interested in working with people from EDA companies who would be willing to try an example of one of these models he had created. He noted they had some questions about using the new C_comp model during simulation but using the old style C_comp for K(t) generation. - Curtis: Motion to adjourn. - Michael M.: Second. - Arpad: Thank you all for joining. AR: Arpad to forward Walter's comments on BIRD197.3 to the entire ATM. AR: Bob to update the latest BIRD197.3 draft with revision info corrections. ------------- Next meeting: 21 May 2019 12:00pm PT ------------- IBIS Interconnect SPICE Wish List: 1) Simulator directives