S32K312 CAN sampling points points are not accurate

[Ali22]

Dears

We encountered the following problem: Our CAN sampling point calculation method is shown in the following figure. Currently, we are equipped with 13/15=86.6, but the actual sampling points measured by CANOE between 90.3% to 100%, with a significant deviation. May I ask if there is any problem with my calculation method?Or are there any other points to pay attention to?

[Senlent]

Hi@Ali22

Please tell me your flexcan's clock setting, and the baud rate , sample point you want get, and i will give you some reference.

[Ali22]

We want sample point is around 0.87

[Ali22]

Dears

After calculating with the tool you provided, the result is the same as before. What I want to know is that the sampling point I calculated was 86.666%, but the actual test value exceeded 90%. Is this deviation considered normal?

[Senlent]

Hi@Ali22

I think the calculation method here is correct. This may be caused by clock errors and test errors.
I suggest switching the FLEXCAN module clock to an external crystal clock, reconfiguring it, and then testing it again.（Because the error of the external crystal clock is usually very small）

[Senlent]

Hi@Ali221

Please refer to the below configuration.

https://community.nxp.com/t5/MPC5xxx-Knowledge-Base/MPC5xxx-S32Kxx-LPCxxxx-CAN-CAN-FD-bit-timing-cal...

[Ali22]

Dears

CAN clock is 60MHz, Bitrate is 500K, Thanks.