TLV62595: Jitter and ringing problems

Hi Rhea,

You can try adding 1 more 22 uF output cap to the output to see if the jitter is reduced.  The 2x 22uF caps may not give enough effective capacitance.  Can you also share the part number of those caps?

I did not see the layout in the attachment.  Did you intend to attach it?

Does the EVM work fine in the same test?

Thanks,

Chris

Hi Chris,

Sorry，to add attached document.

I need to communicate with customer for other problems,thanks!

TLV62569DBV-20230327.docx

Thank you for sending the layout.

The return path from the input caps to the IC's GND pin is very long (through vias) and much longer than on the EVM.  Can they try adding a 22uF capacitor between the IC and inductor, right across the IC's VIN and GND pins?  This is how the layout is done on the EVM.

Chris

Hi Chris,

Thanks for your support!

I have sent it to the customer for change test.

Test device and method are added in the attachment.

TLV62569DBV-20230328.docx

Thanks a lot! 

Thanks for sharing their setup.  Does the EVM work fine in the same test?

Can they try adding a 22uF capacitor between the IC and inductor, right across the IC's VIN and GND pins? 

Chris

Hi Chris,

The customer does not have TLV62595DMQR EVM,but has TPS62827 EVM. The two chips are pin to pin compatible.

At the same time, TLV62569DBV on the product plate was measured for comparison.

The attached documentation provides relevant test diagrams.

TLV62569DBV-20230331.docx

Thanks for your support~

Thank you for measuring the TPS62827 EVM.  Are those results acceptable to the customer?

So, we need to identify what differences the EVM has compared to their board.  One difference is the input cap placement.  They should add an input cap on their board right next to the IC and see if this improves the jitter.

Another possible difference is the amount of total output cap, which is not present in their EVM test setup.  How much additional capacitance is on the 0.9V rail on other pages of the schematic?

Chris

Hi Chris，

When the TPS62827EVM-036 EVM is connected to the customer's product board, SW jitter is reduced.

Please help to see what solutions are available. Updated the attachment.

TLV62569DBV-20230406.docx

Thanks!

Thank you for powering their board with the EVM.  Since that behaves better, the PCB layout is responsible for the difference mostly.  They should re-layout their PCB, with the input cap next to the IC's pins, as done on the EVM.

Chris