TJA1042T/3 CAN Transceiver doesn't work

[ldj979]

Dear NPX users,
nice to meet you.

I am a beginner designer using NXP products.

I designed the CAN Transceiver circuit by looking at the data sheet and application note of the NXP High Speed ​​CAN Transceiver TJA1042T/3 product suitable for 3.3V MCU.

Based on the completed circuit, a problem occurred during hardware debugging of the board and I have a question.

I know that high-speed CAN communication requires 3.5V and 1.5V of CAN H and CAN L signals, respectively, based on 2.5V. 

However, in the board I designed, although the 2.5V reference point is the same, there is a problem that CAN H and CAN L output 3.9V and 1.1V, respectively.

Currently, I set 500k/bps and 87.5% of Sample Point through PEAK System's CAN communication equipment.

I checked the communication, but found the problem that communication was not working.
 
I am judging that communication is not possible due to the above problem. 

I would like to hear from users if the circuit I designed is a badly designed circuit.

[kef2]

With 500kbps shortest time between two adjacent edges should be 2us. On your waveform it is something like 5-10us. Please check your CAN bus timing. 

[ldj979]

Thanks for the advice. As additional information Using STM32F407VG6 chip, 
APB1 Clock was set to 24Mhz. 

Prescaler was set to 3, Time Quantum to 125.0ns, Bit Segment 1 to 13 Times, 
and Bit Segment2 to 2 Times. As a result, we set the conditions of 500k bit/s, 
sample point 87.5%.

The PCAN settings were also set the same as above. Maybe my settings are wrong?

[kef2]

I'm not familiar with STM32. Having oscilloscope you always can confirm CAN bit timing. Waveform you presented doesn't look like 500kbps CAN.

[ldj979]

As a result of checking with your advice, 
the external crystal frequency did not match and the bit timing seemed to be strange. 

I modified the crystal frequency and it works fine as a result.
thank you

[ldj979]

Additionally
CAN equipment is also equipped with a Termination Resistor.