OMAP-L138 - IBIS Model - Cref and Vmeas

Milian

For question #1, these are just different loading capacitances for the timing measurement. However, they do not represent the max load capacitances.

For question #2, IBIS model is used to verify signal integrity, not for timing analysis.

Thanks

David

David,
I do not know what to make of your answers. Let me reword my questions:

Please provide a pair of Capacitance load and Voltage thresholds for which ALL timing specs in the datasheet are guaranteed.

Many thanks,
Milan

Milan

This info is covered in section 5.1 of the data sheet.

The IBIS model is used to simulate signal integrity, not for AC timing measurement.

Thanks

David

David,
1]
The datasheet section 5.1.1 reads as follows:
It is not necessary to add or subtract the transmission line delay (2 ns or longer) from the data sheet timings.
Q: How does the author of the document know how many devices are connected to the output in my design? How does he/she know the slew rate in my design?

2]
The datasheet section 5.1.1.1 reads as follows:
5.1.1.1 Signal Transition Levels
All input and output timing parameters are referenced to Vref for both "0" and "1" logic levels..
For 3.3 V I/O, Vref = 1.65 V.
Q: If you measure the timing from Vref to Vref, why do you say it is not necessary to add or subtract the transmission line delay (2 ns or longer) from the data sheet timings? I do not know what to make of this specification.

3] 
The IBIS model does  not appear to have been taken under test conditions. Why is that?

4] 
The OMAP-L138 IBIS model available on your website is not usable since models  are swapped with components. Please indicate when you are going to fix it.

Many thanks,
Milan

Please clarify.

Thanks,
Milan

Milian

Let me take a look at question 1, 2, and 3 and get back to you. For question 4, attached is the updated OMAPL138 IBIS model.

Thanks

David

Re 4 - I have been getting different OMAP-L138 IBIS files from different TI folks, most likely modified locally.

Being in Aerospace, this approach is totally unacceptable for me due strict workflow process and file version control requirements.

Please take appropriate measures towards releasing an official usable OMAP-138 IBIS file with an updated version, timestamp and revision history.

Many thanks,
Milan

Hi Milan,

The updated IBIS files are scheduled to go live sometime next week. I will inform you once that has happened by posting here again.

Regards,

Sunil Kamath

Milan

 1] The datasheet section 5.1.1 reads as follows:
It is not necessary to add or subtract the transmission line delay (2 ns or longer) from the data sheet timings.
Q: How does the author of the document know how many devices are connected to the output in my design? How does he/she know the slew rate in my design?

A: The datasheet is showing the datasheet timing reference model for the FACTORY pin electronics for the numbers spec’d.  The propagation delays of the transmission line are calibrated out of the measurements in the FACTORY. Therefore you do not have to account for the FACTORY transmission line delays, but you still must account for your own PCB transmission lines delays.

2]
The datasheet section 5.1.1.1 reads as follows:
5.1.1.1 Signal Transition Levels
All input and output timing parameters are referenced to Vref for both "0" and "1" logic levels..
For 3.3 V I/O, Vref = 1.65 V.
Q: If you measure the timing from Vref to Vref, why do you say it is not necessary to add or subtract the transmission line delay (2 ns or longer) from the data sheet timings? I do not know what to make of this specification.

Please see the answer to the first question. On your PCB, you must drive the signals to Vref within the times specified in the datasheet using a slew rate >= 4V/ns. The factory used an unspecified slew rate < 4V/ns.  So you should use a faster one in order to exceed factory test conditions.

 3] 
The IBIS model does  not appear to have been taken under test conditions. Why is that?

The test condition is used for AC timing measurement, but the IBIS model is used for signal integrity check. These are two separeate test setups.

Thanks

David

Hello David,

1] I see your point now.
    What is the factory transmisison line capacitance?

2] I conclude the documentation is unclear and should be improved.

3] Not sure why you keep telling me this.
IBIS models are used for Flight time simulations the output of which may have significant impact the timing analysis results and margins.That's why I need to know the exact test conditions (see bullet #1) for which datasheet specifications are guaranteed.

Milian

For question #3, yes, the IBIS simulations will help determine slew rates and propagation, but it can only be used to check the relative impact of PCB conditions with respect to the same signal. It does not comprehend absolute timing, so it can’t be used, for example, to simulate a clock signal and a data signal to make sure setup and hold are good. It CAN be used though to see if changing a PCB trace CHANGES the slew rate/propagation enough to increase timing windows. One should use more accurate model such as SPICE Netlist or timing lib for Verilog if timing is really a concern.

Thanks

David

David,
Thank you for having enlighten me on timing analysis process. I am afraid I disagree with you.

IBIS models describe IO timing based on precise lab measurement under worst case test conditions. Thus, IBIS models are ideal for accurate timing analysis on the condition that data in the model is valid.

I have been missing OMAP-L138 test load info as the datasheet does not seem to be clear in this asspect.

Many thanks,
Milan

Milan,

The updated IBIS files are on ti.com now.

Regards,

Sunil

Sunil,
Thanks for your notice.

1. I can't seem to locate IBIS models for ZWT and BGWTMEP devices.
2. May I ask you to indicate a test load for which OMAP-L138 datasheet timing specs are valid?

Many thanks,
Milan

Milan,

The link is hidden in the frame - you have to click "Show More" to see it. Here is the link anyway: http://www.ti.com/litv/zip/sprm378c.

Will get back to you on the test load. It should be 15pF but I will confirm.

Regards,

Sunil