P3T1755 SDA Drive Levels ?

[smiller2]

I attempting to read the temperature via I3C at 12.5MHZ  from an FPGA implementation of the I3C controller.   The first byte is partially correct, but the send byte always reads as 0xFF.   It seems that on the transition of the read SDA line from High to Low, it takes too long and the P3T1755 detects a bogus Restart and places SDA into HiZ.     If I reduce the SCL frequency to 10MHZ with a 60nS clock low time, then it works as expected. 

Looking at the scope traces, the SDA line drive during reads seems quite weak when compared to some other devices.    I will attempt to attach scope shots of the working and non-working cases.  

Is there a spec on this drive strength?    Is there a command to increase it? 

Thanks.

[smiller2]

The schematic goes across several boards, so it is not easy to share.   The SDA and SCL lines are FPGA pins with 1K pullups at the FPGA.  The signals route through several boards and have additional 1K pullups close to the P3T1755 sensor.     I am doing a Private Read command to retrieve the temperature.    I will attach some logic analyzer screen shots.  In the fail case, the logic analyzer is detecting that the SDA line transitioned HIGH while the SCL was already HIGH and is decoding as a STOP condition.    The P3T1755 has control of the SDA line at this time, so the issues are the prop delay in the SDA level change and the drive strength of the SDA line.    The datasheet does not seem to address the drive strength.  

[guoweisun]

HI Smiller

What's the communicate distance between sensor and controller?

If the long distance also can't support the maximum speed.

[smiller2]

The FPGA is on a separate board.   There is probably 24 inches of PCB trace and two connector sets between the FPGA and the sensor.

[guoweisun]

Hi Smiller

For the I3C,the maximum wire length would be a function of speed, as all the reflections and Bus turnaround must complete within one cycle.

Larger distances can be achieved at the lower speeds than at the higher ones.

For example, at 1 meter (between Controller and Target), the maximum effective speed is around 6 MHz for read, to allow for clock propagation time to Target and SDA return time to Controller.

So in your case consider about 60cm length communication distance plus two connectors which should speed down to guarantee the communication quality.

Hope this could help you!

[guoweisun]

HI Stephen 

Could you please share with the test waveforms and your schematic here?

Thanks!