TPS62810-Q1: Abnormal SW Waveform

Wayne Liu

Part Number: TPS62810-Q1

My customer sent me a SW waveform, the spikes are so big, could you please help analyze why? The red circle is the position of inductor. Thanks.

over 3 years ago

0 Dorian Brillet de Cande over 3 years ago

TI__Genius 14866 points

Hello Wayne,

Can you share the layout of this board ?

What is the load condition for this behavior? I see that you you convert 5.0 VIN to 1.8Vout.

Thank you very much!

Regards,

Dorian

0 Wayne Liu over 3 years ago in reply to Dorian Brillet de Cande

TI__Expert 8605 points

Hi Dorian,

Thank you so much for your quick support.

The load is 1.8V 1A, neither inductive nor capacitive. The layout is as below

0 Dorian Brillet de Cande over 3 years ago in reply to Wayne Liu

TI__Genius 14866 points

Hello Wayne,

My first impression would be that it is related to the VIA which is between the SW pad and the inductor. As the SW pin will switch at high speeds with steep steps, having a via to link other components from the other side of the the board will only induce that this noise will couple through the layers of the board.

If you want to improve this behavior you should have al the components connected to the SW node on the same side of the board avoid the use of via on the SW pin.

Let me know if you have any other questions.

Thanks a lot!

Regards,
Dorian

0 Dorian Brillet de Cande over 3 years ago in reply to Dorian Brillet de Cande

TI__Genius 14866 points

Hello Wayne,

I haven't heard back from you.

Do you have an y further questions here ?

Let me know.

Thanks a lot!

Regards,
Dorian

0 Wayne Liu over 3 years ago in reply to Dorian Brillet de Cande

TI__Expert 8605 points

Dorian,

You are right, customer has verified this issue is related with the via, they are optimizing it, thanks for your support.

0 Dorian Brillet de Cande over 3 years ago in reply to Wayne Liu

TI__Genius 14866 points

Hello Wayne,

If you need anymore help once the layout is optimized, let me know!

Thank you,

Dorian

0 Wayne Liu over 3 years ago in reply to Dorian Brillet de Cande

TI__Expert 8605 points

Hi Dorian,

Let's go back to this issue.

Yesterday customer tested the SW waveforms on our EVM, they got similar spikes waveforms as the above one, the highest voltage is 8V and width is 2ns, is this still normal? Thanks.

0 Dorian Brillet de Cande over 3 years ago in reply to Wayne Liu

TI__Genius 14866 points

Hello Wayne,

Do you know how customer is sensing the SW waveform?

If the ground loop from your measurement method is long they you will pick some noise which will appear as a pike like customer observers.

The following is a good read on this topic:

Look at topic 3: proper measurement technique.

Let me know if you have any questions here.

Thanks,

Dorian

0 Dorian Brillet de Cande over 3 years ago in reply to Dorian Brillet de Cande

TI__Genius 14866 points

Hello Wayne,

I haven't heard back from you.

Do you have any follow up questions here ?

Did customer use the described technique to minimize the switching peaks ? Did you get feedback from customer?

Let me know.

Thanks a lot!

Regards,

Dorian

0 Wayne Liu over 3 years ago in reply to Dorian Brillet de Cande

TI__Expert 8605 points

Dorian

Thanks for your support, customer has solved this issue with your suggestions.

Power management

Power management forum

TPS62810-Q1: Abnormal SW Waveform