how to deal with bidirection port in function ATPG?

[lty]

I am doing the function ATPG for a chip.
there are many bidirection ports for the chip, these ports is not the scan port.
How should I deal with them?
Could I ignore their direction, and dont assign a fixed value to its trimode EN control signal?
Or If I want to used them in our ATPG mode,and I assign a fix direction for them,Then when in ATPG mode, I use input ports for stimulus and output ports for result capture, Is that right?

Thank you for your any comments.

 

[rameshsuthapalli]

Hi Ity,

There are two solution to deal with the biriections pins in a design.

1. fix them in the I/p direction which will reduce the coverage by little much since u can't detect the faults of the logic rthat is driving pad when it is in o/p mode.but by placing i/p direction u can controll more signal that was driven by this so covrage will be more than when it is fixed in the O/P mode.
2.don't fix the bidirectional pads.which will reduce the little covegare on the I2C and C2O paths. since the main coverage is only because of the scan chains than others.

regards,
ramesh.s

 

[dft_guy]

Any ATPG tool will automatically detect the direction of the bidirectional pin when creating patterns so that it will not drive the pin when it is behaving as an output. So theoretically, you don't have to do anything to these pins, and you will get good coverage.

If you had used one of these bidi's as a scan input or output, then care needs to be taken and the ATPG tool needs to know how to treat it. You'd have to look at the docs for the given tool.

The only other complication would arise if the bidi has a pullup/down resistor - then you might lose some fault coverage, because the tool might not be able to simulate it correctly.

Hope that helps,
John
DFT Digest

 

[shiv_emf]

make all bidi ports as inputs...in scan mode