Re: BIRD 58

Regarding the c_comp issue discussed in the Mar 26 IBIS meeting,
related to BIRD 58, I would like to register my opinion in favor
of allowing for the use of separate c_comp values attributed to
each Driver Schedule stage.

Although c_comp may be named as such because of the original
intention that it encapsulate an overall component level capacitance,
the practical matter is that Driver Schedule stages are most
likely to correspond to actual transistors in an IO section. Each
transistor does have it's own effective die capacitance. The
designer of the stage buffers may (should) have validated their
behavior individually, with c_comp intact. To then use them in a
scenario where the c_comp values are not used may not make sense.
(insert discussion of the aggregate capacitance for a cluster of
transistors vs. sum of individually calculated capacitances here).

Furthermore, IBIS could gain maximum flexibility by stating that
the c_comp values of individual stages are used in addition to
the c_comp that might be specified for the top-level model. In
this way, model designers can specify c_comp=0 for any portion
that is not to be used. This part of the proposal is debatable in
my own mind, simply because I have some difficulty producing an
elegant explanation of what the top-level capacitance represents,
when the`"transistor" capacitance is already accounted for. But it
offers flexibility. And it eliminates the requirement that simulators
ignore c_comp data present in the Model, which users may not be aware
of.

Mike LaBonte

Muranyi, Arpad wrote:
>
> ****************************************************************************
> **
> ****************************************************************************
> **
>
> BIRD ID#: 58
> ISSUE TITLE: Driver Schedule keyword clarification
> REQUESTER: Arpad Muranyi, Intel Corporation
> DATE SUBMITTED: 3/2/99
> DATE ACCEPTED BY IBIS OPEN FORUM: pending
>
> ****************************************************************************
> **
> ****************************************************************************
> **
>
> STATEMENT OF THE ISSUE:
>
> The [Driver Schedule] section of the latest IBIS specification (v3.2) does
> not provide enough information to the model maker and EDA tool vendor on
> essential details of this keyword. The spec. can be interpreted in several
> different ways which can lead to serious discrepancies and consequences.
>
> ****************************************************************************
> **
>
> STATEMENT OF THE RESOLVED SPECIFICATIONS:
>
> One paragraph from BIRD35.3 is included in the [Driver Schedule] usage rules
> section of the specification. This portion seems to have been
> inadvertently left out from the spec. and explains the gray areas.
>
> The additional words do not change the original intent behind this keyword.
> It is only added to spell out how the keyword must be used so that the model
> maker and EDA tool vendor could be in agreement on how the model works.
>
> The new section is marked with a "*" character at the beginning of each
> line.
>
> |===========================================================================
> ==
> | Keyword: [Driver Schedule]
> | Required: No
> | Description: Describes the relative model switching sequence for
> referenced
> | models to produce a multi-staged driver.
> | Usage Rules: The [Driver schedule] keyword establishes a hierarchical
> order
> | between models and should be placed under the [Model] which
> | acts as the top-level model. The scheduled models are then
> | referenced from the top-level [Model] by the [Driver
> Schedule]
> | keyword.
> |
> |* The [Driver Schedule] keyword overrides any existing
> [Pullup],
> |* [Pulldown] and [Ramp] or [Rising Waveform] and [Falling
> |* Waveform] data for that model. However, the required tables
> |* for [Model] are still required for backwards compatibility
> |* reasons for simulators which do not support multi-staged
> |* switching. This information can be used as a composite
> model
> |* cross-check, although the resulting model may not perform in
> |* the same manner as the multi-stage model.
> |
> | The [Driver Schedule] table consists of five columns. The
> | first column contains the model names of other models that
> | exists in the .ibs file. The remaining four columns
> describe
> | delays: Rise_on_dly, Rise_off_dly, Fall_on_dly, and
> | Falling_off_dly. All values are referenced to 0 seconds for
> | the start of the rising transition and 0 seconds for the
> start
> | of the falling transition. All delays must be equal to or
> | greater than 0.
> |
> | The Rise_on_dly entry gives the beginning of the low-to-high
>
> | transition. The Rise_off_dly entry may be given to end the
> | low-to-high transition and initiate a high-to-low transition
>
> | during the rising cycle. Similarly, the Fall_on_dly gives
> | the beginning of the high-to-low transition. The
> Fall_off_dly
> | may be given to end the high-to-low transition and initiate
> a
> | low-to-high transition.
> |
> | Use 'NA' when no transition is applicable. For each model,
> | the transition sequence must be complete, i.e., it must
> start
> | and end at the same state.
> |
> | Only the [Pulldown] and [Pullup] tables and transition data
> | [Ramp] or [Rising Waveform] and [Falling Waveform] data are
> | used from each model that is referenced. The [Model]
> keyword
> | provides the specification information, [GND Clamp] and
> [POWER
> | Clamp], and C_comp, regardless of information contained in
> | the referenced models.
> |
> | It is recommended that a "golden waveform" for the device
> | consisting of a [Rising Waveform] table and a [Falling
> | Waveform] table be supplied under the [Model] keyword to
> | serve as a reference for validation.
> |
> | No [Driver Schedule] may reference a model which itself has
> | within it a [Driver Schedule] table keyword.
> |
> | Other Notes: The added models typically consist of Open_sink (Open_drain)
>
> | or Open_source models to provide sequentially increased
> drive
> | strengths. The added drive may be removed within the same
> | transition for a momentary boost or during the opposite
> | transition.
> |
> | The syntax also allows for reducing the drive strength.
> |
> | Note that the Rise_on_dly, Rise_off_dly, Fall_on_dly,
> | Fall_off_dly parameters are single value parameters, so
> | typical, minimum and maximum conditions cannot be described
> | with them directly. In order to account for those effects,
> | one can refer to the fastest waveform table with the delay
> | number and then insert an appropriate amount of horizontal
> | lead in section in those waveforms which need more delay.
> |
> | Note: In a future release, the [Driver Schedule] keyword may
> | be replaced by a newer method of specification that is
> | consistent with some other planned extensions. However, the
> | [Driver Schedule] syntax will continue to be supported.
> |---------------------------------------------------------------------------
> --
> [Driver Schedule]
> | Model_name Rise_on_dly Rise_off_dly Fall_on_dly Fall_off_dly
> MODEL_OUT 0.0ns NA 0.0ns NA
> |
> | Examples of added multi-staged
> transitions
> M_O_SOURCE1 0.5ns NA 0.5ns NA
> | low (high-Z) to high high to low (high-Z)
> M_O_SOURCE2 0.5n 1.5n NA NA
> | low to high to low low (high-Z)
> M_O_DRAIN1 1.0n NA 1.5n NA
> | low to high (high-Z) high (high-Z) to low
> M_O_DRAIN2 NA NA 1.5n 2.0n
> | high (high-Z) high to low to high
> |
> |===========================================================================
> ==
>
> ****************************************************************************
> **
>
> ANALYSIS PATH/DATA THAT LED TO SPECIFICATION:
>
> An attempt to build an IBIS 3.2 buffer model for a three-stage buffer using
> the [Driver Schedule] keyword failed. The initial testing of this model
> with the tools of two EDA vendors showed only two of the three stages
> working. This happened because the first stage of the buffer was placed
> into the "main" [Model] section of the IBIS model due to a personal
> interpretation of the IBIS 3.2 spec. After consulting with several people,
> it became apparent that the specification does not mention some of the key
> information that is necessary to build a correct IBIS model for this kind of
> a buffer.
>
> The following paragraph in BIRD35.3 contains the missing information which
> is necessary for making correct IBIS models using the [Driver Schedule]
> keyword. This section seems to have been accidentally left out from the
> specification, and should be pasted into the appropriate location as shown
> above. (Note that the first and second sentences are deleted because they
> do not provide new information in the context above).
>
> "The new keyword [Driver Schedule] provides a syntax for sequencing other
> models. This is given as a keyword under the [Model] keyword. It overrides
> any existing [Pullup], [Pulldown] and [Ramp] or [Rising Waveform] and
> [Falling Waveform] data for that model. However, the required tables for
> [Model] are still required for backwards compatibility reasons for
> simulators
> which do not support multi-staged switching. This information can be used
> as a composite model cross-check, although the resulting model may not
> perform in the same manner as the multi-stage model."
>
> ****************************************************************************
> **
>
> ANY OTHER BACKGROUND INFORMATION:
>
> Telephone conversation with Bob Ross on 2-26-99 confirmed that the IBIS
> specification is missing the description of some of the usage rules which
> were documented in BIRD35.3 and was implemented accordingly in ICX.
>
> ****************************************************************************
> **
Received on Fri Mar 26 11:16:30 1999