IBIS SUMMIT MINUTES 1/29/96

From: Bob Ross <bob@icx.com>
Date: Mon Feb 05 1996 - 11:27:00 PST

 DATE: February 5, 1996
 
 SUBJECT: 1/29/96 EIA IBIS Summit Minutes
 
 VOTING MEMBERS AND 1996 PARTICIPANTS LIST:
 Cadence Design C. Kumar*
 Contec CAE, Ltd. Dileep Divekar*, Norio Matsui*
 Hewlett Packard, EEsof Karl Kachigan*, Henry Wu*
 HyperLynx Kellee Crisafulli*
 IBM (Jay Diepenbrock)
 INCASES (Werner Rissiek), Olaf Rethmeier
 Intel Corporation Stephen Peters*, Will Hobbs*, Arpad Muranyi*,
                                John Keifer*, Aaron Tang*, Duane Quiet*,
                                Donald Telian*, Tim Kruchowski*
 Interconnectix, Inc. Bob Ross*
 Meta-Software (Les Spruiell)
 Motorola (Ron Werner)
 National Semiconductor Syed Huq*, Donald Snyder*, Chune-Sin Yeh*
 NCR (formerly ATT-GIS) Dave Moxley*, Richard Mellitz*
 NEC (Hiroshi Matsumoto),
 Quad Design Jon Powell*, Chris Rokusek*
 Quantic Labs (Mike Ventham)
 Tanner Research, Inc. (Scott Wedge)
 Texas Instruments Roger Cline*, Ben Andresen, Sri Jandhyala*,
                                Tareq Shahwan*, Dan Phipps*
 Thomson-CSF/SCTF (Jean LeBrun)
 UniCAD Canada Ltd. Celso Faia*
 VLSI Technology Dick Ulmer*, Sung Oh*, Swami Gangadharan*,
                                Daniel Kim*, Tom Dockery*
 Zuken-Redac (John Berrie)
 
 OTHER PARTICIPANTS IN 1996:
 Altera Vadim von Brzeski*
 AMP Hank Herrmann
 Apteq Design System Dan FitzPatrick*
 Compaq Mark Leonard*
 Cypress Bruce Wenniger*
 Digital Equipment Corp. (Jeff Chu), (Steve Thierauf)
 EIA Patti Rusher
 IC Works Eric Chen*
 Intergraph/Veribest Ian Dodd*
 Micron Technology Brian Johnson*
 MicroSim Ralph Perez
 Mitsubishi Tam Cao*
 North Carolina State U. (Michael Steer)
 S3, Inc. Sarathy Sribhashyam*, Phap Duong*
 Seimans Nixdorf Arnaud Lacariere*
 Tektronix, Inc. Brock Hannibal*
 TRW & Free Model Foundation Robert Harrison*

 In the list above, attendees at the meeting are indicated by *. Principle
 members or other active members who have not attended are in parantheses.
 
 Upcoming Meetings: The bridge numbers for future IBIS teleconferences are as
 follows:

      Date Bridge Number Reservation # Passcode
      2/16/96 (916) 356-9200 2-48992 5519538

 All meetings are 8:00 AM to 9:55 AM Pacific Time. We try to have agendas out
 7 days before each open forum and meeting minutes out within 7 days after.
 When you call into the meeting, ask for the IBIS Open Forum hosted by Will
 Hobbs and give the reservation number and passcode.
 
 NOTE: "AR" = Action Required.
 
 -------------------------------- MINUTES -------------------------------------

 INTRODUCTIONS
 Will Hobbs opened the meeting and welcomed the participants. A total
 of 45 people representing 24 organizations attended. A big thank you to
 National Semiconductor and Syed Huq for providing the excellent local
 arrangements and hosting the meeting.

 MINUTES REPORT, MISC.
 Patti Rusher requested that the EIA legal disclaimer note, "This meeting
 was conducted in accordance with the EIA Legal Guides and EIA Manual of
 Organization and Procedures." be attached to all Minutes. The January 12,
 1996 minutes have been ammended to include this note at the end.

 Hank Herrman submitted a clarification of the last sentence in the BIRD31
 discussion. It has now been rewritten to read:
 "Hank Herrmann indicated that AMP Spice models do not have any assumed
 ground references. The model is designed so that any of the pins
 can be used as ground or power references. You simply connect the
 pins to the references, just as you would do with the actual connector."

 
 MISCELLANY/ANNOUNCEMENTS
 The reflector went down again, and Bob Ross will deal with this when
 he returns. You can check email96 for missed, archived e-mail.

 
 PRESS AND WEB PAGE UPDATES
 No report. (Several papers at Design SuperCon96 the following day
 mention IBIS)

 OPENS FOR NEW ISSUES
 Roster Update - Bob Ross

 
 ROSTER UPDATE
 Bob Ross will request roster update information. The current roster
 has not tracked all of the people/organization changes and will be saved.
 The new roster will contain only the 1996 supplied information. The
 distribution lists at the top of the minutes will reflect only 1996
 participation.

 EIA /IBIS WEB PAGE
 Syed Huq requested suggestions for more material and improvements of
 the EIA Web page. It contains the diagrams for the EDN article and
 also links to other home pages.

 EIA/IBIS 1996 MEMBERSHIP
 Patti Rusher sent again membership invoices upon discovering that the
 first ones stated incorrect amounts. Members have 45 days upon receipt
 to pay.

 
 DAC STANDARDS BOOTH
 Will Hobbs led the discussion on IBIS participation in the Standards booth
 at the Design Automation Conference (DAC) in June, 1996. Jon Powell
 recommended that the booth have handouts and order forms and be informally
 staffed since most participants will be busy with company activities.
 Pointers to other booths may be considered.
 
 
 IBISCHK2 PARSER
 Jon Powell will be in charge of tracking ibischk parser bugs. A created
 a bug report form which will be available on vhdl.org/pub/ibis/bugs.
 He new mail address ibischk-bug@vhdl.org can be used to submit bugs.
 Paul Munsey and Ron Neville have completed the previously reported
 problems and are awaiting instructions for the next release. This
 should serve as the new baseline. Jon Powell should sent out the form.

 
 PRESENTATIONS:
 The remaining portion consists of brief notes on the presentations.
 Presenters are requested to submit electronic versions of their talks
 in any format to Bob Ross to be put on vhdl.org.

 
 BRIEF OVERVIEW WITH HISTORY OF IBIS - Will Hobbs
 Will Hobbs gave a brief overview and history of IBIS starting with
 Arpad Muranyi's behavioral modeling, Intel's original Excel Spreadsheet
 approach, and Don Telian's consolidation of formats with initial
 EDA industry involvement leading to Version 1.0 in 1993 in ASCII format.
 Advances continued leading to ANSI/EIA-656 (Version 2.1) plus the
 Overview, Cookbook, EDN Article, s2ibis utilities (from NCSU), ibischk
 utilities, reflector lists and repository of models from Intel and
 National Semiconductor and other information on vhdl.org, and formal
 affilation with EIA. Will listed some proposed Version 3.0 enhancements.
 (More on this at the end.)

 There is still need for updating the Cookbook and providing articles
 on Version 2.1 features. Stephen Peters is doing an internal update,
 but keeping it sanitized so it can be used externally.

 Steven Peters is also working with Don Contrell of CFI on the DCL (Delay
 Calculation Language Standard)

 Will Hobbs reported that IBIS has an unanticipated use of conveying
 buffer specification requirements internally and to customers in the
 early stages of design. IBIS can serve as an electronic databook.

 NATIONAL AND THE IBIS EXPERIENCE - Syed Huq
 National Semiconductor Corporation (NSC) joined just after the V1.1
 ratification, and pushed differential I/O, multiple rails, etc. Some of
 NSC's models are generated from bench measurement. They modified s2ibis to
 ns2ibis converter. It changes their NsSPICE models to IBIS, and they've used
 it to convert a large library to IBIS. NSC has a library management system
 for archival and controlling revisions. They have an internal IBIS home page,
 and library. They've also generated training materials available on the
 internal Web and publicized IBIS throughout the corporation. Several NSC
 groups are generating models. The EDA industry was employed to evaluate
 their models and Syed received excellent support.

 All of the models National has produced are available on their Web site:
 www.national.com. All of data for these components are available in the
 same area, not just the models.

 Syed Huq spearheaded putting an IBIS web page on eia.org, with an FAQ section,
 the EDN article on IBIS, hypertext links to vhdl.org for versions of the
 specication, many models, minutes archives, etc., are there, and a "virtual
 poster" of member companies with potential links to members' home pages. EIA
 IBIS members can have their logo on the virtual poster (a good reason to join
 the committee).

 Syed would like a good How-to guide to support measurement, such as measuring
 package-less dV/dt, use BTL IEEE 1194.1 for C_comp measurement. He'd also
 would like a way of tracking model usage on the Internet. Syed proposes
 taking steps to sync up IBIS with other related standards, including JEDEC
 JC-15 (package parameters) and CFI/Pinnacles/OMF/DIE/DIET efforts.

 Kellee Crisafulli suggested adding a Web counter to the IBIS home pages of
 NSC and eia.org. This may not give full information on who is accessing
 what, but it would give a rough idea of the interest in the web sites.

 Jon Powell suggested we form some subcommittees to focus on some of the
 inter-specification relationships.

 Syed's wish list: Create a dummy model or two that can be simulated and has
 all of the IBS defined keywords and parameters with data tables. Also, a
 glossary/clarification of IBIS keywords and parameters as defined in the
 IBIS standard.

 Bob Ross reported that there does not exist a set of models for testing
 the IBIS Golden Parser.

 IBIS AT INTEL: BUFFER ARCHITECTURE TO CUSTOMER MODELS - Stephen Peters
 Stephen Peters gave an overview of the "Intel IBIS Experience" from the
 technical side. Groups within Intel define the essential characteristics
 of the interconnect between chipsets and CPUs, to bound the problem, balance
 the tradeoffs between what the buffer designer can create and what the system
 designer can tolerate. This results in "specification wars" until common
 ground is found. To handle the communication between the system architect
 and the buffer designer, IBIS is being used. This is tongue-in-cheek referred
 to as "GUMBIE," or Grand Unified Model Buffer Information Exchange. The
 ultimate result of the dialog between architects and designers is an IBIS
 model that can be validated against silicon and shipped to customers.

 Stephen gave a real example of a recent buffer definition project, where
 the designers began with a simple on resistance "Ron" (linear V/I) approach.
 The Ron model didn't allow a system to be designed. By switching to V/I,
 they were able to meet the design challenges. This is because Ron wasn't
 able to handle saturated outputs.

 Stephen enumerated the issues he perceives that affect customer models: How
 do you generate a "worst case" model? Are the models truly a specification
 of the device? How do we guarantee a model (and do we have to?): What do
 you guarantee? How do you test and characterize the model's parameters?

 Obtaining skewed silicon doesn't necessarily demonstrate worst case buffers,
 since the process slow skew may not be the skew that produces the slowest
 I/O buffer characteristics - it may be slow for some other characteristic of
 the IC.

 Jon Powell: Do you really want a worst case model? The system designer may
 be willing to trade off a lower yield against higher performance. Jon also
 raised the issue of how to tell the customers what parameters of a model are
 independent. Kellee Crisafulli suggested a corner table setup map.

 MODEL PROVIDER EXPERIENCES, THE TI EXPERIENCE - Sri Jandhyala
 Sri Jandhyala of Texas Instruments (TI) discussed the work going on at TI
 with IBIS. Many groups within TI have developed I/O buffer models because
 of strong customer demand. They are concerned by the large effort required
 to generate standard SPICE models, and are concerned about the IP protection.
 IBIS is attractive, too, because of the vendor independence, reduced cycle
 time. Customers are asking for IBIS models by name.

 Sri listed many companies that have explicitly asked for IBIS models, such
 as Apple, Bay Networks, HP, IBM, Intel, LG Semiconductor, Unisys and others.
 TI has allocated resources to steer their IBIS efforts and support IBIS.
 They are developing internal tools to automate model creation process. TI
 is committed to the support and development of IBIS. IBIS model generation
 is to be integrated into the product development flows.

 Current issues: standardizing the process, characterizing the models,
 validating the models, establishing common procedures and improving the
 coordination between work groups.

 Question: Whom does a customer contact to get model information? Answer:
 The TI Web page will have links into TI IBIS information. Sri's e-mail is
 srij@ti.com.

 Company policy is not to release IBIS models publicly on the Internet. This
 appears to be because management isn't yet comfortable with the models. TI
 has around 30 models available so far.

 Dave Moxley indicated he'd like more ready access to the models, as he
 might use the models as a method to decide whether he wants to buy the TI
 components.

 USING IBIS MODELS IN SYSTEM DESIGN - Dave Moxley
 Dave Moxley of NCR (formerly AT&T-GIS) reported that NCR develops their
 models through simulation and validates them via measurement.

 NCR is in the massively parallel and SMP server business. They stress
 reliability and quality. They routinely develop with sub-nanosecond timing
 margins, with bus speeds up to 100 MHz today, and higher in the future.

 Critical analysis goes into board design and ASIC design. Both noise margin
 and timing margins have shrunk dramatically over the years that NCR has been
 doing these designs.

 Signal Integrity questions: How many power and ground pins are required?
 What layout constraints apply, how many bypass capacitors and of what type?
 What layout topologies will be required? How fast can a design operate
 reliably? What buffer strengths are required to insure reliable operation?
 How will normal manufacturing variation affect performance?

 Regarding manufacturing, IBIS is used here to play what-if games, which is
 why he is interested in obtaining IBIS models over the spread of process
 variations. They use both 2D and 3D solvers.

 Dave showed the general flow they go through in doing the design from
 concept through what-if to layout and verification.

 They've seen good correlation between SPICE, IBIS and silicon in test
 silicon, test vehicles and components from ASICs to clock distribution
 devices, buffers, memory and processors.

 Question: How do chip specications relate to IBIS Models: setup, hold,
 clock to data, access time, clock skew, buffer delay?

 During their design analysis, they look at every circuit, examining for
 corner case variations, timing, crosstalk, threshold violations and other
 anomalies.

 They would like industry wide support for IBIS, with vendor contacts at each
 company and public library maintenance. They are considering making IBIS
 model availability a requirement in contracts with the vendors they are
 dealing with.

 Other issues included silicon corner verification, slow-n/slow-p may be
 fastest edges. Also included are drivers with feedback. Sometimes there
 are application specific work-arounds. They would like to be able to obtain
 IBIS models from customer supplied parameters of Voltage, temperature and
 load. And, they'd like to know the relation between the IBIS models and the
 board level related timing specifications.

 MODELING LOSSY TRANSMISSION LINES - C. Kumar
 C. Kumar of Cadence presented modeling issues related to modeling lossy
 transmission lines. The industry trends that concern him are increasing
 speeds, 200 MHz and beyond, which drives higher and higher slew rates;
 shrinking sizes; and High DC and Skin Resistance. At higher frequencies,
 the resistance of a conductor can increase by an order of magnitude.

 The classic transmission line model begins with a segmented RLC
 representation, with the number of segments going to infinity. A lossy
 transmission line adds frequency dependent losses. The incident and
 reflected waves encounter different resistances due to the differences in
 high frequency components of the two waves.

 Fundamental challenges are in how to approximate these effects. They use a
 "rational approximation." Stability is one potential pitfall. He gave an
 example of an ECL transmission line running at 400 MHz. He compared a lossy
 transmission line simulation with segment and lossless models, demonstrating
 that noise margin losses are shown by the lossy transmissions and not by the
 other two. The lost noise margin of 20 mV is significant in ECL designs.
 He also showed an example of a CMOS circuit running at 200 MHz showing that
 at 200 MHz the effects are not nearly as pronounced as at 400 MHz.

 Kumar showed that modeling bonding wires with inductors versus skin resistors
 produces a less-damped waveform than the skin effect resistor, which
 correlates better with actual components.

 He concluded that at higher speeds, frequency dependent effects become
 significant, particularly with respect to noise margins.

 Jon Powell noted that if and when we ever get to modeling large networks,
 such as MCM, lossy transmission lines will be a significant contributor to
 accuracy.

 EGG8: MODELING PACKAGES WITH PHYSICAL DESCRIPTIONS - Kellee Crisafulli
 At Syed Huq's request, Kellee Crisafulli also discussed the Visual IBIS Editor
 The Visual IBIS Editor is a Windows based standard text editor including
 special features for source code editing, with the syntax checker built in.
 At any point, you can check the syntax, like a spell checker. There is also
 a graphical viewer for VI tables, and rising and falling waveforms. Finally
 there is an IBIS file creation wizard, a skeleton file for either V1.1 or
 V2.1 that can be used as a foundation model to begin developing the model.
 It can be used on Windows 3.1, Windows 95 or Windows NT, but not UNIX. It
 is available at http://www.hyperlynx.com for download for free, but HyperLynx
 asks that models that are created on it are made publicly available.

 Kellee then discussed Egg 8: Physical Package format for IBIS. There is
 presently no standard way to create or deliver signal integrity models for
 integrated modules, such as SIMMs and MCMs. SIMMs from different vendors
 for the same target socket are laid out differently and therefore behave
 differently. Kellee proposes a SIMM module physical package file at the
 top, linked to individual IBIS files below.

 After looking at the state of the art, Kellee concluded there are no standard
 PCB formats that PCB tools read and write today without purchasing special
 translators. Existing standards, such as EDIF, are too complex. Instead,
 Kellee proposes a simple ASCII PCB description. He gave an example of a 4
 layer package. Starting with the stackup (signal and power, etc., layers
 can be defined), pad stack, layer to IBIS model association, and an
 interconnectivity portion that defines specific nets with bias, pad stacks,
 segments, pins. The segments are defined by their endpoints. The ordering
 is "unimportant." One issue discussed was the problem of line segments that
 connect, but whose x,y endpoint coordinates don't always coincide due to
 trace length, trace overlap, via width, etc.

 Kellee contends that the IBIS committee needs to endorse a standard for
 physical descriptions.

 Kellee used curly braces instead of square brackets to make it simpler for
 parsing, but he is open to alternatives.

 Cutouts in the ground plane can be modeled. There is no support for cross-
 hatched planes, however.

 Several wondered if it would be possible to specify min and max for such
 things as dielectric constants and trace width.

 Don Telian expressed concern that unless there were netlisters to generate
 these files for a given format, the SIMM suppliers will not use it. Jon
 Powell and Kellee said that there are lots of people that create writers for
 new formats for as little as $10K.

 Jon would like to make sure there is something in the specication that ensures
 that non-simulated conditions will not meet the specification. E.g., a net
 that is two completely discrete segments would be illegal.

 Ian Dodd added zero length nets should be checked. Kellee hesitates to add
 these checks because all layout packages today put out certain amounts of
 garbage. There is general agreement that this is a problem, but not on
 where to put the burden of guaranteeing that such bad data doesn't torpedo
 the simulation.

 
 BIRDs 28.3, 32: MODELING PACKAGES WITH ELECTRICAL DESCRIPTION -
 Stephen Peters and C. Kumar
 Stephen Peters discussed BIRD 28.3, wherein an electrical approach is taken
 to describing packages. He showed drawbacks with the current IBIS
 methodology that are addressed by the BIRD. He also identified a problem
 in modeling BGA packages because of a plating stub that can't be represented
 in a standard section by section description.

 C. Kumar said that two values to an electrical description are: it can't be
 reverse-engineered as a physical description can, and the data can be very
 compactly represented. Kumar filed BIRD 32 as an extension to BIRD 28.3 to
 add arbitrary coupling and handle an arbitrary number of sections compactly.
 The coupling information is represented in a named matrix and multiple
 models can be referenced by any number of models.

 Question: Is a pure physical description sufficient so we don't need the
 electrical version, or do they each have a place? The consensus seems to
 be that each has value. There may be cases where one would be preferred
 over another. For example, there are cases in packages where the signal
 path may not be planar, causing the physical description to be inadequate.
 If they both are supported, Ian Dodd would like to see the same syntax be
 used for each type of representation.

 The discussion highlighted a number of issues the forum will have to resolve
 before incorporation into the specication.

 BIRD 31: CONNECTOR MODELING - Bob Ross
 Bob Ross of Interconnectix discussed BIRD 31, extending IBIS into connector
 utility. Bob has already decided BIRD 31 has to be replaced by a new one
 that Bob will write soon. Connectors are usually described by lumped
 sections, transmission lines, sections and edge rates, their symmetry, pad
 capacitance and the ground reference. It usually is supplied in SPICE
 subcircuit format and also for single line connectors, in a data sheet.

 Karl Kachigan of HP asked how to deal with connectors that have mated pairs
 when they're not mated, such as backplanes. This could be handled in a manner
 similar that decribed by BIRD30 which has been approved. Normally unmated
 models are not provided.

 Bob showed an actual right angle connector with multiple rows and columns
 and different path delays.

 He then showed what would be in a SPICE model format with coupled pins. The
 pins are grouped, the descriptions are in sections and edge rates, with
 symmetry (e.g., 3 inductor, 2 capacitor sections), making it not directly
 convertable to transmission line format needing an equal number of L's and
 C's. Other Details such as explicit or implicit ground references,
 pad capacitance, fins (connected metal planes dividng groups of pins),
 etc. need to be considered. Exact pin mapping to actual connector circuit
 pins are details critical to the EDA industry for correct path simulation.
 So the set of subcircuits must accurately present pin descriptions that
 can be used for any pin.

 Spice subcircuit connector models are derived from field solver extractions
 and converted into equivalent L and C transmission line sections. Exact
 pin mapping of the circuit board contact of all pins to the connection
 path is critical to the EDA industry for correct path simulation. Details
 on certain pins (or treatment of missing pins at ends and other places
 due to construction by modules) may be missing. Thus connectors are
 suitable for IBIS syntax because of the exact pin mapping and because
 IBIS would use the field solver matrices directly. The alternative is
 to reverse engineer models from Spice Models.

 Bob will revise BIRD31 to have one or two pin columns for numbering,
 cross-connection, and symmetry variations, will embed cable within
 connector model groupings, and offer pad capacitance options. Larger
 models would have to be auto-generated by expanding from a known set
 of comprehensive field solver extractions.

 STORED CHARGE MODELING, DIODE DELAY DISCUSSION - Jon Powell, Bob Ross
 Jon Powell of Quad Design opened the discussion on diode transit time.
 Charge storage results from excess minority carriers injected across the
 junction in forward bias. Transit time (TT) represents the minimum time
 required to store or remove charge from the junction. For moderate or more
 forward bias, the stored charge dominates.

 Jon recommends the reference "Semiconductor Device Modeling with SPICE," by
 Antongnetti and Massobrio. This book has a good model of this effect by
 relating a capacitance in parallel to a current sinking diode, with the
 parallel net in series with a resistor Rs. The TT effect can create heavy
 but delayed ringing, an effect missed without the diode TT being modeled.

 Jon proposes we add TT support to IBIS. Stephen asked if there was a
 measurement technique that could derive TT. Jon says that this is being
 done today, so there is a way. He also noted that the diode that has the
 intrinsic capacitor is the static protection diode, which swamps the effects
 of input and output capacitances of the buffer structure.

 Bob Ross continued the presentation. He showed that Ct can be derived from
 the standard diode equation to be a proportional to TT * Id * (q / KT).
 Id is the diode current, q is charge, K is Boltzmann's constant, and T
 is absolute temperature. The theoretical dependence on temperature is
 motivation to have typ, min and max specifications.

 Kellee Crisafulli asked whether we should specify TT or Ct? He feels that
 Ct can be measured by forward biasing the diode at a specified current and
 measuring the capacitance.

 Bob and Jon provide a BIRD to bring this issue to closure.

 EGG 6: DETECTING CMOS VERSUS TTL - Jon Powell
 Jon Powell has concluded that just knowing the technology is not necessary
 or sufficient. Rather, the standard 50 ohm measurements do not adequately
 demonstrate TTL characteristics. He recommends instead 500 ohms to ground
 and 7 ohms to power. He bases this on where the load lines intersect the
 pullup and pulldown tables to best isolate the pullup from the pullup
 characteristics for the rising and falling transitions. This doesn't violate
 the IBIS specication, but implies a needed change to the cookbook. Even
 more loads allow greater accuracy, but these two are sufficient to produce a
 good model. For CMOS, the 50 ohms to power and 50 ohms to ground is still
 sufficient. Jon will help contribute to the cookbook.

 DOUBLE COUNTING AND OTHER MODEL GENERATION ISSUES - Arpad Muranyi
 Arpad Muranyi of Intel discussed double counting that can happen easily
 with internal pull-up resistors. The pull-up shows up in both the power
 and the ground clamp portions of the model in extractions. This results in
 double counting the effect. Arpad proposed an algorithm which eliminates
 this effect.

 The double counting effect is minor but might affect the DC float voltage
 of a net. Arpad will write this up and submit it to Kellee Crisafulli for
 inclusion in the cookbook.

 Another issue Arpad addressed was the way pull-up and pull-down IV curves
 are defined, which is the difference between floating and on values. These
 curves add correctly only if the same reference values are used at the time
 of measurement. He used a 5-volt safe 3.3 V input device to illustrate the
 problem, as there may be two different reference voltages at different times.

 A third issue is the derivation of the value of the R_load sub-parameter. For
 backward compatibility, the ramp keyword is still required in V2.1. The ramp
 could be derived from IBIS 2.1 without this keyword. R_load is required if
 not equal to 50 ohms. Deriving the ramp number is problematic if all
 R_fixtures are not the same.

 Both of these cases are worth noting in the cookbook. Kellee says they've
 already run into this problem with some V2.1 models, and since it is
 superseded by the ramp tables, they ignore it. One possible solution for
 V3.0 is to add an R_fixture to the ramp.

 RAIL, RULES AUGMENTED INTERCONNECT LAYOUT SPECIFICATION - Donald Telian
 Don Telian of Intel introduced the RAIL specification aimed at automating
 the PCB layout and simulation process. It is an industry-wide initiative.

 Circuit simulation has become critical as speeds increase, with flight
 times being a significant percentage of the clock period, and crosstalk
 increasingly problematic due to ever finer pitches.

 RAIL addresses the gap between good simulation and manufacturable
 boards. The three elements of system design are physical, logical and
 electrical. RAIL complements IBIS to complete the electrical piece of
 the pie. Rail is a "container" for electrical constraints.

 RAIL has been developed to an initial specification level by Intel and five
 simulation vendors. Most tool vendors already have elements that can accept
 RAIL data, which is an electronic (ASCII) version of layout guidelines,
 providing clear communication of electrical constraints, such as maximum
 trace length, acceptable flight time skew, etc.

 Don described the RAIL syntax, which looks like IBIS syntax. Don obtained
 permission from EIA to make the two look this similar. Typical keywords
 include [Stackup], [Group Nets], [Topology], [Budgets] (flight times, max
 overshoot, crosstalk limits), [Clock Skew]. He provided a typical example.

 RAIL can be used as a post-processing tool, where the route is done,
 then checked out and the PCB is modified and the post-processing is done
 again, re-iterating until everything comes out clean. It can also be
 used interactively, where nets are hand routed while the simulator is
 running in the background to tell the user if there is a problem with
 the net in real time. The third approach takes a netlist, IBIS and RAIL
 and feeds an autorouter with integrated simulation, resulting in a nearly
 effort-free and fast layout.

 VHDL has volunteered to make resources available for RAIL on vhdl.org,
 including rail@vhdl.org, rail-info@vhdl.org and rail-request@vhdl.org
 reflectors. There are slides, minutes, the V1.0 version of the specification,
 etc. are available on vhdl.org under /pub/rail. [Note, the rail reflectors
 have been experiencing similar difficulties as the ibis reflectors.]

 The five EDA vendors that put the specication together with Intel are
 Cadence, Mentor, Quad Design, HyperLynx and Interconnectix.

 Don would like the simulators to extract RAIL files from successful
 designs for later use. He pointed out the value of the RAIL data for
 sustaining engineers, who can go back to the original design database
 and see if it met constraints, without having to track down the
 original engineer on the project.
 
 The RAIL committee will meet on a monthly basis, and hopes to have
 demonstrations at DAC. RAIL files will be released from PCD at Intel,
 along with the IBIS models. A RAIL golden parser will probably be needed
 for RAIL.

 
 USEFUL NON-SIMULATON DATA INCLUSION - Jon Powell
 Jon Powell asks: Is there a place in IBIS for information that doesn't
 affect simulation? Threshold is already included, so Jon explored
 additional ones. Timing related information could include additional
 thresholds, other timing information, such as pin-to-pin information.
 Several others are already addressing this, including DIET, DCL. Another
 idea is to correlate between IC timings and the IBIS models. E.g., an
 annotation could be made in the IBIS file that a particular specification,
 such as Tco, was made with a particular corner of the IBIS model.

 Perhaps we should expand the threshold to include a rising and a falling
 threshold, which could handle hysteresis. The 54/74LS132 defines a Schmitt
 trigger with a vlow and vhigh for the rising edge, and a vlow and vhigh
 for the falling edge.

 Component reliability could be addressed by adding information on low and
 high voltage requirements - overshoot limits and time in threshold limit.
 Other factors include: monotonic requirements, pulse size requirements,
 illegality of leaving the input floating.

 John will submit a BIRD.

 BUS HOLD CIRCUITS AND EIA JEDEC JC40 - Bruce Wenniger
 Bruce Wenniger of Cypress Semiconductor is involved with the JEDEC JC40
 committee which is working to make sure components from different
 manufacturers are interchangeable. He discussed bus hold, in response
 to the discussions on the reflector.

 Bus hold is the ability of an input or output pin to retain the last
 valid logic state after the source driving it enters the high impedance
 state. Such circuits reduce noise in a system, and improve hold times.
 The bus hold circuit is particularly important with CMOS devices which
 can float to their linear region and draw excessive current and possibly
 self-destruct. Bus hold circuits also eliminate the need for pull-up
 and pull-down devices, reducing power consumption. This is a real
 benefit for lap top computers. The circuit is simply a weak output
 connected to the pin with positive feedback.

 There is no JEDEC standard for 5V logic. The de facto standard was
 established by the TI ABT Series. The LVTTL standard is currently being
 debated by the JC40 committee, which consists of a number of IC vendors.

 
 OPEN DISCUSSION: IBIS 3.0 POSSIBILITIES, WISH LISTS - Will Hobbs
 Will Hobbs of Intel led the closing discussion on what people want in
 Version 3.0 of IBIS and other areas. The results of the discussion are
 summarized:

 Will listed the top technical enhancements for Version 3.0 that the
 group proposed:

  Packages and connectors:
    Physical description
    Electrical description
    MCM support
  Diode modeling (transit time, charge storage)
  Multiple drive level buffer-dynamically switched or staggered turn-on
  Feedback support (e.g., dynamic pre-driver effects such as gate modulation)
  Multiple drive level-selectable (done)
  Balanced drive outputs (already can be handled)
  Scaling of parameters (possibility)

  Other ideas:
    Refine SSO (simutaneous switching outputs)
    Power modeling
    Support for static timing
    New challanges for higher speeds

 The informational and technical clarification and model availablity
 wishes were:

  Model validation guidelines
  New Modeling Cookbook
  More models
  Clarification of worst case/best case?
  Presentations, overviews, intros to IBIS
  Press Release after summit

  Other ideas:
    Enhance s2ibis, golden parser
    Create awareness of library
    Move IBIS into new uses & arenas

 In addition to Intel and National, the attendees indicated that the
 following companies also have developed IBIS models: S3, Micron, Cypress,
 TI, and VLSI.

 The attendees felt that the most important issue to address was model
 validation. That is an obstacle to many companies releasing models
 widely. Kellee Crisafulli reiterated his offer to provide a HyperLynx
 simulator to companies willing to develop and make public IBIS models.

 CELEBRATION DINNER:
 Thirteen people celebrated the Adoption of ANSI/EIA-656 at dinner.
 

 NEXT MEETING:
 The next telephone meeting is set for Friday, February 16, 1996.

 ==============================================================================
                                       NOTES
 
 IBIS CHAIR: Will Hobbs (503) 264-4369, Fax (503) 264-4210
             will_hobbs@ccm.jf.intel.com
             Server Chipset System Validation Manager, Intel Corp.
             2111 NE 28th M/S JF1-57, Hillsboro, OR 97124 USA
 
 VICE CHAIR: Jon Powell (805) 988-8250, Fax: (805) 988-8259
             jonp@qdt.com
             Transmission-Line Products Manager, Quad Design
             1385 Del Norte Rd., Camarillo, CA 93010
 
 SECRETARY: Bob Ross (503) 603-2523, fax (503) 639-3469
             bob@icx.com
             Modeling Engineer, Interconnectix, Inc.
             10220 SW Nimbus Ave, K4, Portland, OR 97223

 This meeting was conducted in accordance with the EIA Legal Guides and EIA
 Manual of Organization and Procedure.
 
 The following e-mail addresses are used:

   ibis-request@vhdl.org
       To join, change, or drop from either the IBIS Open Forum Reflector
       (ibis@vhdl.org), the IBIS Users' Group Reflector (ibis-users@vhdl.org)
       or both. State your request.

   ibis-info@vhdl.org
       To obtain general information about IBIS, to ask specific questions
       for individual response, and to inquire about joining the EIA-IBIS
       Open Forum as a full Member.

   ibis@vhdl.org
       To send a message to the general IBIS Open Forum Reflector. This
       is used mostly for IBIS Standardization business and future IBIS
       technical enhancements.

   ibis-users@vhdl.org
       To send a message to the IBIS Users' Group Reflector. This is
       used mostly for IBIS clarification, current modeling issues, and
       general user concerns.

   ibischk-bug@vhdl.org
       To report ibischk2 parser bugs. The Bug Report Form Resides on
       vhdl.org in /pub/ibis/bugs/bugform.txt along with reported bugs.

 Information on IBIS technical contents, IBIS participants, and actual
 IBIS models are available on the IBIS Home page:

   http://www.eia.org/eig/ibis/ibis.htm

 Check the pub/ibis directory on vhdl.org for more information on previous
 discussions and results. You can get on via ftp anonymous, "guest" login from
 telnet or dial-in (415-335-0110), or send an e-mail request to the automatic
 archive server, archive@vhdl.org.
 
 "IBIS Spoken Here" placards are available from Jon Powell (jonp@qdt.com) for
 use at trade shows.
 ==============================================================================
Received on Mon Feb 5 11:33:11 1996

This archive was generated by hypermail 2.1.8 : Fri Jun 03 2011 - 09:52:29 PDT