TXS0102: Different IC markings have different characteristics

Allen yang

Part Number: TXS0102

Hi Team,

The factory reported that the TXS0102DCTR received material has two different marks.
When the NG mark material is used, the TXS0102DCTR fails to function properly.

OK marker: 35UT NG: NFE
631X 5CR

1 month ago

0 Clemens Ladisch 1 month ago

Guru 325030 points

The TXS0102DCTR is made at two fab sites with different dies. All the guaranteed datasheet parameters are the same; it is likely that you circuit tries to exceed these parameters. Please specify exactly what fails; show an oscilloscope trace or voltage measurements.

0 Allen yang 1 month ago in reply to Clemens Ladisch

Expert 1485 points

Hi Clemens,

The test waveform is as follows

Architecture diagram

Schematic

3617.Schematic.pdf

0 Clemens Ladisch 1 month ago in reply to Allen yang

Guru 325030 points

What is the problem with that waveform?

0 Allen yang 1 month ago in reply to Clemens Ladisch

Expert 1485 points

Hi Clemens,

Are there differences in characteristics between products from different regions?

This sample comes from two countries: US and DE.

0 Clemens Ladisch 1 month ago in reply to Allen yang

Guru 325030 points

Typical characteristics might be different. But as long as your circuit stays inside the recommended operating conditions, their behaviour will be the same.

What exactly is wrong with the test waveform?

0 Allen yang 1 month ago in reply to Clemens Ladisch

Expert 1485 points

Hi Clemens,

wrong waveform

0 Clemens Ladisch 1 month ago in reply to Allen yang

Guru 325030 points

What is R249? What signals are measured with the yellow and green waveforms?

0 Allen yang 1 month ago in reply to Clemens Ladisch

Expert 1485 points

Hi Clemens,

Please refer to Schematic

Original RC249=22R, RC250=22R, RC232=0R,

RC249=0R, Test pass

However, the signal overshoot occurred.

RC232=22R, can improve overshoot

MDIO is being routed online, but R232 has been changed to 22R and RC250 to 0R.

Final solution : RC249=0R, RC250=0R, RC232=22R, What problems might arise with this design?

signal overshoot

Schematic

5504.Schematic.pdf

0 Jack Guan 1 month ago in reply to Allen yang

TI__Guru 58131 points

Hi Allen,

I did not see the mentioned R249/ R250 in the provided schematic. From the waveform below it seems that the knee is the concern? Please confirm. Adding an additional dampening resistor could help to reduce the transient edges on falling/ rising edges to be less harsh, and it seems that it is an acceptable solution in your case. It is OK to use these dampening resistors at the outputs of the TXS with no concern in this application.

Regards,

Jack

0 Allen yang 1 month ago in reply to Jack Guan

Expert 1485 points

Hi Clemens,

Sorry R249/ R250 = RC249/ RC250

Does the TXS0102 support a frequency of 6.25MHz?

0 Jack Guan 1 month ago in reply to Allen yang

TI__Guru 58131 points

Hi Allen,

Yes, 6.25MHz/ 12.5Mbps is supported. Please refer to the switching ch. table in datasheet for more details:

Regards,

Jack

Logic

Logic forum

TXS0102: Different IC markings have different characteristics