IBIS Macromodel Task Group Meeting date: 19 Mar 2013 Members (asterisk for those attending): Agilent: * Fangyi Rao * Radek Biernacki Altera: * David Banas Julia Liu Hazlina Ramly Andrew Joy Consulting: Andy Joy ANSYS: Samuel Mertens * Dan Dvorscak * Curtis Clark Steve Pytel Luis Armenta Arrow Electronics: Ian Dodd Cadence Design Systems: Terry Jernberg * Ambrish Varma Feras Al-Hawari Brad Brim Kumar Keshavan Ken Willis Cavium Networks: Johann Nittmann Celsionix: Kellee Crisafulli Cisco Systems: Ashwin Vasudevan Syed Huq Ericsson: Anders Ekholm IBM: Greg Edlund Intel: * Michael Mirmak Maxim Integrated Products: Mahbubul Bari Hassan Rafat Ron Olisar Mentor Graphics: John Angulo Zhen Mu * Arpad Muranyi Vladimir Dmitriev-Zdorov Micron Technology: Randy Wolff Justin Butterfield NetLogic Microsystems: Ryan Couts Nokia-Siemens Networks: Eckhard Lenski QLogic Corp. James Zhou SiSoft: * Walter Katz * Todd Westerhoff Doug Burns * Mike LaBonte Snowbush IP: Marcus Van Ierssel ST Micro: Syed Sadeghi Teraspeed Consulting Group: Scott McMorrow * Bob Ross TI: Casey Morrison Alfred Chong Vitesse Semiconductor: Eric Sweetman Xilinx: Mustansir Fanaswalla Ray Anderson The meeting was led by Arpad Muranyi ------------------------------------------------------------------------ Opens: - Walter: Would like to discuss BIRD adoption plan for IBIS 6.0 -------------------------- Call for patent disclosure: - None ------------- Review of ARs: - Walter write new BIRD for repeater/retimer pin keywords - In progress - Arpad try to apply BIRD 158 shortcut language to his BIRDs 116-118 - In progress ------------- New Discussion: Analog BIRDs: - Arpad showed a list of BIRDs that have been converted to the new BIRD format - Arpad: The content has not changed. - This may be easier once my AR has been done - Arpad showed presentation "Analog BIRD Overview": - slide 3: - Arpad: The first four BIRDs would have to be voted in to implement the approach - The other two are optional. - Arpad: A BIRD-like document has been uploaded but not yet announced - This combines four BIRDs - We should read it - It could be submitted as an official BIRD if that helps - Ambrish motioned to submit the combined BIRD incorporating BIRDs 116,117,118,129 - Bob seconded - There were no objections - Bob: We should look at them first - I have some technical comments, one about parameters and the other about parsing logic - Arpad: We can have those discussions next week Arpad showed presentation BIRD 159 Discussion: - slide 2: - Arpad: There was a question about the receiver threshold voltage - Mike Steinberger and Walter had replied about this - The IBIS specification has clear language - Walter: We should decide that the waveform is centered around 0V - Michael M: What is the reference location? - Arpad: Let's get to questions at the end - slide 3: - Arpad: The specification says there is a single waveform - It does not assume it is differential - slide 4: - Arpad: On page 123 the interface is said to be differential - slide 5: - Arpad: How do we resolve this? - Michael M: Any common mode effect is outside the scope of AMI - Walter: Agree - Michael M: Most flows are statistical - It still needs to be known if the common mode bias is correct - AMI models are not helpful for that - David: So a DC bias will not be modeled correctly? - Todd: We are only saying the network characterization produces a differential waveform - Nothing is said about common mode - Walter: The impulse response is based on the differential output - This uses the TX analog model - David: The channel has even and odd mode propagation - The RX input will never know about common mode - The question is if the differential behavior of the TX can be affected - Arpad: We pass to the algorithmic model an impulse response, not a waveform - Ambrish: GetWave receives a waveform - Arpad: It says "centered nominally around zero volts" - Walter: A stream of ones will not be centered around zero - We transmit only the differential part of the waveform - Todd: David was asking if the mode conversions affect the differential waveform - David: Michael asked if loading that is not nominal changes the result - Can that happen with a Ramp-only TX? - Todd: Yes - Arpad: It can if the differential buffer uses two different models - Todd: Non-linear curves can create those effects - David: So Ramp-only models do not preclude us from catching those effects - Walter: Edge timing skews can cause this effect too - EDA tools with proper analog models will model it - Arpad: The algorithmic model will not care about non-differential effects - I thought we were supporting both single ended and differential models - Walter: I could create an S2P just to represent the differential part of a channel - The intent was not to do single ended DDR4 - AMI does not really support it - Arpad: Are we restricting ourselves unnecessarily? - Walter: We can fix it in the future if the need arises - At that point an RX receiver threshold fix would be needed - Todd: The specification we have now is about differential - Arpad: What about the page 123 language? - Radek: That should be eliminated - David: The model only knows it is receiving a series of numbers - Todd: It won't be right if the channel has not been handled correctly - Walter: I think the answer to "Is it clear to everyone" is no - This BIRD proposal should be withdrawn - Bob: The assumption of centering on 0V is still a problem - Walter moved to recommend voting down BIRD 159 - David seconded - No one objected - David: Does Michael have any thoughts about his questions? - Michael M: Statistical analysis does not account for common mode - That's OK, but it needs to be clearly documented Walter: We should discuss a BIRD plan in a future meeting - Michael M: Agree, we have 18 or 19 open BIRDs now - Members should submit their preferred lists - Michael M: Speaking as an Intel employee for a moment, a repeater specification is very important - Walter: I will send a spreadsheet ------------- Next meeting: 26 Mar 2013 12:00pm PT Next agenda: 1) Task list item discussions ------------- IBIS Interconnect SPICE Wish List: 1) Simulator directives