Re: BIRD #62 -- Enhanced Specification of Receiver Thresholds

Hello Al:

   I will answer your third question first as it has a direct bearing on
your other two questions.

  The Vth parameter is speced as Vth, Vth_min and Vth_max so that
variation in threshold voltage due to differing gate thresholds
ratios between the P and N channel devices could be isolated from the
change in threshold due to supply variations. To put it another way,
Vth_min and Vth_max represent the variation in threshold voltage under
typical operating (VCC and temp) conditions only. In IBIS, the min and max
columns include effects from both operating point (VCC and temp) as well
as process. This is not what the authors wanted, so there had to be
three separate parameters. Note that one can calculate the traditional
min and max value for Vth by factoring in the the variations in supply
voltage (or reference voltage as the case may be) as shown in the BIRD.

  The above also explains why I choose to use a separate [Receiver Thresholds]
keyword for these parameters. The parameters of the current [Model Spec]
keyword use only the traditional three column typ/min/max format. I was not
comfortable adding to this keyword parameters that use the parameter = value
format, especially with the Vth_min and Vth_max parameters. No other keyword
mixes these two formats for single valued (non-table) parameters.

  As far as the relationship between this keyword and other keywords go, the
bird states that if the [Receiver Threshold] keyword is present the
Vin*_ac and Vin*_dc parameters override any other input threshold parameters
given elsewhere. However, I do agree that the bird should call out the
affected parameters specifically. (Note that the [Receiver Threshold]
keyword does not effect any of the overshoot limit parameters in [Model
Spec]). I will clarify this the next time the bird is reissued.

Thanks for the input.

   Regards,
   Stephen Peters
   Intel Corp.

> 1. How does this relate to existing receiver related info in ibis (Model Spec
> stuff, thresholds, etc.) ?
>
> Please be specific. If some keywords are the same, or have some fixed
> relationship, say so.
>
> I am not asking for a personal reply. This should be part of the spec.
>
> 2. Since we have some keywords, and the Model Spec group, this seems to add
> another group that is related in that it provides related info, yet blocked out
> separately, with a form of modularity that seems orthogonal to what it seems
> ought to be grouped. Shouldn't the existing thresholds and the new added info
> be in the same group, so a new user can understand it without following it
> historically? Why not just add the new keys under Model Spec?
>
> 3. Why do some parameters have typ/min/max syntax, while others have 3 separate
> parameters: Vth, Vth_min, and Vth_max?
>
>
> I have no problem with the new info provided about the circuit, just its
> presentation.
Received on Thu Sep 16 13:03:14 1999