DRV8889-Q1: EMC problem around 80MHz

Hello Daniel,

We preliminarily judged that the 80M EMC issue may be caused by the outward conduction of the charge pump, can you help confirm:
1. Which part of the DRV8889 is the 80M EMI likely to be generated from? There is no 80MHz fundamental frequency source in the device. 
2. What is the frequency of the Digital oscillator? Either 10 MHz or 16 MHz. Either cases 80 MHz is a harmonic frequency.
3. What is the switching frequency of the A B phase current chopper control? This depends on the smart tune decay mode used, VM voltage, IFS current and the stepper L and R values. Typically it will be in the 10's of kHz
4. What are the corresponding optimization and improvement measures, such as how to adjust the spread spectrum of the charge pump, reduce the output voltage and current of the charge pump, etc.?

- Enabling SSC (spread spectrum) should mitigate the digital oscillator harmonics.

- Reducing slew rate SR using CTRL1 register settings will mitigate any harmonics generated due to switching edges.

- Using ferrite beads in the CP caps usually between VM and the 220nF capacitor will mitigate EMC from the charge pump.

- Having capacitor 2.2nF to GND on AOUT1,2 and BOUT,2 could also help mitigate this harmonic frequency. 

Were there any other peaks in the multiples of 10 or 16 MHz fundamental frequency in the recorded spectrum? Thanks.

Regards, Murugavel

Hello Murugavel,

Thanks for your feedback. Can you help with below questions:

1. Enabling SSC (spread spectrum) should mitigate the digital oscillator harmonics.

How to active SSC on the digital oscillator and charge pump? I see there is no description how to spread spectrum in datasheet.

2. Were there any other peaks in the multiples of 10 or 16 MHz fundamental frequency in the recorded spectrum?

We will cheek and get back to you.

Hello Daniel,

1. I just verified and confirmed this is enabled permanently in this device. See page-13 of the datasheet.

2. Okay please let us know. Thanks.

Regards, Murugavel

Hello Murugavel,

1. Are there chances for customer to config the SSC of digital oscillator and charge pump? We want to operate with more or less spread spectrum to validate.

2. The statement here mentioned:

    

   Does that means the SSC is activated only when the chip is giving output?

   The EMC issue we met is that even without motor connection, just init the chip, then emission is a lot.

3. There is no fundamental frequency of 10M or 16M and related harmonic from onsite test data.

Thanks.

Hello Daniel,

1. Like I mentioned previously, the SSC in this device is a feature. This cannot be disabled.  

2. The statement you highlighted means, in addition to the SSC which is already active, the output slew rate SR must be configured to a slowest possible SR setting suitable for the application to minimize radiated emissions from the device. Fast SR will generate stronger radiated emissions.

3.  "There is no fundamental frequency of 10M or 16M and related harmonic from onsite test data.". This rules out the internal digital clock as a potential source for the 80MHz. 

If the customer does not have capacitors on AOUTx and BOUTx, they can try 3.3nF to GND on each output. This can reduce the emissions. Because you mentioned the emissions were even with no motor connection I've not discussed about the DECAY modes. With no load current chopping tON and tOFF may not be happening. Does the output have long cable harness connected? In this case the capacitors I mentioned will definitely help.   

Regards, Murugavel