Minutes, IBIS Quality Committee 18 September 2012 11:00-12:00 EST (08:00-09:00 PST) ROLL CALL Cisco Systems: Tony Penaloza Ericsson: Anders Ekholm Green Streak Programs: Lynne Green Huawei Technologies: Guan Tao IBM: Bruce Archambeault Greg Edlund IOMethodology: * Lance Wang Mentor Graphics: John Angulo Micron Technology: Moshiul Haque, Randy Wolff Nokia Siemens Networks: Eckhard Lenski QLogic Corp.: James Zhou Signal Consulting Group: Tim Coyle Signal Integrity Software * Mike LaBonte Teraspeed Consulting Group: * Bob Ross Texas Instruments: Pavani Jella Everyone in attendance marked by * NOTE: "AR" = Action Required. -----------------------MINUTES --------------------------- Mike LaBonte conducted the meeting. Call for opens and IBIS related patent disclosures: - None AR Review: - Eckhard test Open_sink AMI model with different [Voltage Range] values - No report New items: Lance showed Waveform Comparison and S2IBIS3 Roadmap: - slide 4: - Lance: - Need to look at: - Horizontal vs. vertical measurements - Difference peak and difference average measurements - Two different regions can have the same area but different shapes - The index is a percentage value - Bob: The horizontal has to be shifted to make vertical measurements? - Lance: Yes - slide 5: - Lance: The same time window must be used for both - slide 6: - Lance: The average is computed over a specific window - slide 7: - Lance: The timing difference area is calculated between the thresholds - Bob: It could be 20% and 80% - Mike: Golden waveforms are really for drivers only - There is no absolute time for IBIS drivers - Would this provide a way to decide how much to shift? - Lance: The average time offset could be used - Bob: With measured waveforms the shifts are different than SPICE - Lance: Some use a clock as the basis to decide how much lead time to chop off - Most people care more about shape than timing difference - Mike: I often compare entire PRBS sequences - The time shift is adjusted to minimize total difference - Lance: Usually lining up the first edge works - slide 8: - Lance: We would not want to use the beginning portion of this - slide 10: - Mike: The red difference waveforms should how large the V difference can be when slope is high - What is 97% based on? - Lance: It is based on the swing on the golden waveform - DAI under 10% is usually good - If both DAI and DPI are small the waveforms probably line up well - Mike: Is there a horizontal difference measurement for this? - Lance: That was used for shifting - Mike: We might produce a set of waveforms for testing purposes - This would have pairs of waveforms with different types of errors - DC offset - duty cycle distortion - ringback - overshoot - slope - Bob: What does this tool do if the DC voltage is off? - Lance: That shows up directly in DA - Bob: One approach is the start with an idealized waveform with sharp edges - Then make specific measured alterations to it - Mike: That could be done in a spreadsheet - Bob: Anders might have done that - Bruce Archambeault might have done something similar - Mike: That IEEE technique is not really feature selective at all - Lance: Voice recognition systems often use that technique - David Banas has made FSV proposals - Bob: Is this proprietary? - Lance: I need to check on that - We did file a patent Lance: We should have an IBIS Quality presentation for the Asia summits - Mike: We might adapt a short presentation I did recently Bob: We may want to discuss next time whether this group still has critical mass AR: Mike produce starter set of test waveforms AR: Lance check on patent status for correlation methods Next meetings: - Next meetings Oct 2 and Oct 16 - Meeting ended at 12:16 ET