TPS62913EVM-077: hardware setup for noisy measure method

Hi Fulvio,

When you measure the output ripple on C8, you need to use a tip and barrel approach and directly across C8 to instead of a flying ground lead. This will reduce the amount of radiated noise picked up (the flying ground lead acts a like an antenna). You may add a high frequency capacitor 470pF at C17 location. Then you should see same result on J5. SMA J5 is just a probe socket to help reduce the measurement loop area and its positive point does connect to Vout side of C19. Please refer to the link as below for detail ripple measurement.

https://training.ti.com/power-tips-measuring-vout-ripple-dcdc-converters?keyMatch=OUTPUT%20RIPPLE%20MEASUREMENT

Thanks,

Nancy

Hi Nancy,

Thanks for your kind answer but I am looking for someone who know the issue, I am not looking for comments without real experience.

I have not discussed the method because it is clear to expert person that must not be used flying ground lead.

The method I used is soldering a microcoax connector near C8 capacitor and the hardware connection is at the state of the art.

The real topic is not how to connect to C8 capacitor but if TI has followed some rules regarding how to place J5 SMA connector (track lenght and 470pF capacitor) to pick up the noise.

In my opinion the trick (cheat) used by TI to pick up the noise signal help them to obtain good result but it is not the real measurement and it is NOT the real result for the customer.

Best regards,

Fulvio

Hi Fulvio,

I talk with original EVM designer. Here is his input. The goal is to measure what the actual load will see rather than the voltage across C8.

Since the SMA is connected to C8 using a not fully shielded PCB trace we have high frequency electromagnetic coupling into this trace. Therefore we filter this with a 470pF high frequency cap. This is an exact representation what the real load will see, like a high speed ADC having several high frequency bypass capacitors connected right at their supply pins in addition to the parasitic high frequency cap of the high speed ADC package itself.

If you have further question, please email me offline: nancy_zhang@ti.com. I can loop the EVM designer directly for further discussion. Close this post now.

Thanks,

Nancy

Fulvio,

In addition to the EVM, we have implemented the TPS62913 in AFE and ADC applications, and taken measurements to verify the performance in these real applications.  The filter network shown by Nancy is similar to the implementation on these ADC and AFE boards.  Please refer to the following: 

Powering the AFE7920 with the TPS62913 Low-Ripple and Low-Noise Buck Converter

Powering Sensitive Noise ADC Designs with the TPS62913 Low-Noise Buck Converter

Demystifying and Mitigating Power Supply Ripple and Noise Implication on AFE8092 (Rev. A)