SN65LVDS31: Supported simulation method for SN65LVDS31 single-waveform IBIS model

Part Number: SN65LVDS31
Other Parts Discussed in Thread: DS90LV048A,

Hello,

We need to model an SN65LVDS31 clock transmitter connected through PCB traces and board connectors to a DS90LV048A receiver.

Each differential pair connects to two input channels of the same receiver. There is one shared 100 ohm termination at the receiver end, not two parallel terminations.

BCLK is 12.288 MHz or 6.144 MHz. The intended supply range is 3.0 to 3.6 V and ambient temperature is -20 to +85 degrees C.

The SLLC012 download provides an SN65LVDS31D SOIC model with typical data and one rising and one falling waveform. The waveform fixtures use 50 ohms, 1.23 V and 10 pF. We need TSSOP package data and will not substitute SOIC package parasitics.

Please clarify:

1. What supported simulation or extraction method can use this single-waveform-per-edge, current-mode three-state model? Is a supported SPICE equivalent or another documented method available without HyperLynx? Our development environment is macOS. The open-source converters we examined require two waveform fixtures per edge.

2. Is a newer model available with minimum and maximum electrical data, additional waveform fixtures, and SN65LVDS31PW package data?

3. What load and input-transition conditions are necessary to apply the published transmitter and receiver propagation-delay limits when each differential pair connects to two receiver input channels?

4. How should we compare the DS90LV048A timing fixture's 0 to 100 percent input-transition requirement with simulated waveform results measured from 10 to 90 percent?

We need a supported model method and applicable timing conditions. We will keep typical simulation results separate from guaranteed voltage and temperature limits.

Kind regards,
Jonathan Reichardt

  • Hi Jonathan,

    4. How should we compare the DS90LV048A timing fixture's 0 to 100 percent input-transition requirement with simulated waveform results measured from 10 to 90 percent?

    We couldn't provide much recommendation in terms of which simulation tools to use. We generally use ADS or Ansys here internally. The IBIS specification does not seem to require more than the Pullup and Pulldown I-V curves are required for the 3-state model that the the driver uses. (see SLYT388 table 2). 

    4. How should we compare the DS90LV048A timing fixture's 0 to 100 percent input-transition requirement with simulated waveform results measured from 10 to 90 percent?

    No, no newer model is available. The D package model is all we have at this time. There also does not appear to be a great substitute model that i can find in this package. 

    4. How should we compare the DS90LV048A timing fixture's 0 to 100 percent input-transition requirement with simulated waveform results measured from 10 to 90 percent?

    The transmitter load conditions are given in the datasheet, RL=100Ohm, and CL = 10pF, arranged as in figure 8-2 of the datasheet. The 100 ohm resistor should be placed at the receiver side. The rise/fall time of the test circuit is 1ns.

    4. How should we compare the DS90LV048A timing fixture's 0 to 100 percent input-transition requirement with simulated waveform results measured from 10 to 90 percent?

    You should be able to measure the ~0-100% timing in simulation from the output plots. As long as its under 3ns you are within the DS90LV048A's generator timing. Its hard to convert from 10-90% to 0-100% directly since its not likely to be linear. Just remember if you are checking the propagation time at the receiver end its measured from the 0 crossing point of the differential voltage to the 50% output voltage of ROUT. 

    Best Regards,
    Brandon Fisher