Re: s2ibis2 question?

Dear IBIS gurus:

I'd like to share a similar experience working with s2ibis program,
which Arthur has posted earlier.

1. For 3-state I/O buffer, the enable pin (in the pin list) needs to be
   connected to a resistor and tie to VDD or GND, depending if you have
   active-high or active-low buffer. s2ibis program does not do this
   automatically. In addition, declaring [Enable] Active-High or
   Active-Low in .s2i file does not guarantee that pullup/pulldown
   output disable simulation results are correct, unless the proper path
   is provided to VDD or GND by the users as mentioned above.

2. This is a related question to the above.
   When s2ibis generates Power clamp table and Pullup table for 3-state I/O,
   s2ibis sweeps a voltage range VDD->2*VDD for Power Clamp simulation.
   I think the program does this according to IBIS Spec:
   
   Curve Low Voltage High Voltage
   ----- ----------- ------------
   [POWER Clamp] POWER POWER + POWER

   But when n-channel is used as pullup device, we only get insignificant
   current flow because the voltage range is not in the active region of the
   device. Has anyone have similar experience?

   To generate pullup table, only diode data [power clamp] in the active
   region [in the voltage range where the diode is active] is subtracted
   from from pullup data to generate pullup table. So then is it OK to sweep
   -VDD->VDD for n-channel pullup device in Power clamp simulation? because
   in this region the diode is active. It's contrary to IBIS Spec outlined
   above. Please advise.

Best regards,

Hoang Nguyen
Mitsubishi Semiconductor
Received on Tue Oct 15 08:35:56 1996