DRV8353M: Reducing PWM Jitter on MOSFET Output – Causes & Solutions

Hello,

Please note that this part uses an internal oscillator for its internal spread spectrum clocking scheme for the synchronizer, which will inherently result in some jitter.

This being said, jitter can come from a variety of sources. I would need some more information to know exactly where it is coming from.  Here could be some possible sources of the jitter:

What does the input signal look like? Are the PWM frequency and duty cycle stable?  Have you tried multiple PWM frequencies?

Is the jitter the same 150ns on the rising edge of the outputs?

Is the power supply you are using for the board good quality? Do you have any decoupling caps to help filter out any noise from the power input?

Thank you,

Joseph

Hi Joseph,

Thank you for your response.

I would like to ask, what is the typical output jitter range for this driver?

Regarding your questions:

• The input signal is a PWM from an MCU, and I suspect the issue might be coming from there. This is input signal PWM.

• The measured jitter on the rising edge is around 88ns.

• As for the power supply, I don’t think there is an issue. I am using two 330µF/63V electrolytic capacitors in parallel for decoupling.

Looking forward to your insights.

Best regards,
Aji

Hello Joseph,

I’d like to add some clarification regarding the input signal above. The PWM signal is generated by an MCU, and the observed jitter appears to be caused by the torque (Q) and direct (D) current control in the FOC algorithm. In the idle state, the PWM signal seems stable without noticeable jitter.

I have adjusted some parameters:

• Mode: Independent PWM mode
• Switching Frequency: 20 kHz - 50 kHz
• Dead Time: 1 µs

Below is the PWM waveform without any control, just a basic PWM output:

And here is the waveform of the MOSFET’s rising and falling edges:

From these results, it appears that the PWM input itself is stable, but the PWM output still has some jitter, which seems to be causing noise in the motor. Do you have any suggestions on how to mitigate this?

Looking forward to your insights.

Best regards,
Aji

Hi Aji,

Thank you for providing more information. It would seem that the jitter from the input PWM signal is translating to the output. Fixing this input jitter would result in less output jitter when using the MCU. 

From your second measurements with no jitter on the input, the output jitter is behaving normally for this part. 

Here are some more questions I have:

1. I am wondering how you know that this jitter is certainly the cause of the motor noise. Are you sure that this is the only source of the noise? Would you also be able to describe the noise in more detail?

2. I see that you are using 12V and am curious about the peak output current for your application. Depending on this, you might want to look into one of our other parts (DRV8376) with integrated FET architecture which does not use this clock synchronization scheme. 

Thank you,

Joseph

Aji,

Another thing to check to potentially reduce motor noise would be to make sure your commutation style matches the winding style of the motor. For example, if you are using a trapezoidal commutation style on a sinusoidally wound motor, this could also be contributing to the motor noise.

Thanks,

Joseph

Hi Joseph,

Thank you for the answers.

I compared this with other ICs from different manufacturers using the same PWM program without any control adjustments. In those cases, the PWM jitter was around 20ns, and I was able to implement the FOC algorithm without experiencing any noise. The noise in this case sounds more like audio noise.

Regarding your second question, I plan to use a voltage range of around 24-60V with a maximum current of 30-40A in the future.

I will consider this further. Thank you for your help.

Best regards,
Aji

Josep,

I have 2 questions

1. Is the DRV8353 support for BLDC motor
2. Is this driver also can be used for FOC algorithm also?

thank you

Aji,

To answer your questions,

1. Yes, this driver's main purpose is to support BLDC motors. 

2. Yes, this driver can be used for the FOC algorithm.

This is how the part is expected to be used.

Thank you,

Joseph

Aji,

I am surprised that you can see 20ns jitter on the output with a different driver, when the MCU input you showed has 140ns jitter.

Are you using the TI EVM for this part to get your measurements, or are you using a custom board design get the measurements?

Thank you,

Joseph