TPS56837H: SW pin overshoot problem

Hi 

1.Please make sure measure the SW voltage correctly, please use small loop as Figure3 in this apps note.

2.Can you please share your schematic and layout?

Hi

I tried, but still couldn't solve the problem of downrushing to -8V. Could you please help me look at this problem again

Hi

Please check

We tried a SW increased continuation diode that reduced downstroke by 0.5V, but still had -2V,Still more than the recommended pin value

Hi Liyu

The layout part number seems not same with the schematic . But I saw from the layout that Cin is not placed close to device VIn and GND pin. Will you be able to add a 0.1uF Cin close to VIn and GND as close as possible? This should be helpful to reduce the voltage spike.

Thanks and best regards.

Hi

I tried to input the capacitor close to the chip and increase it by 0.1uF, but the downkick problem did not improve

Hi Li yu

Do you mean you improved the SW overshoot by adding RC through BOOT and SW?  The undershoot spike cannot be improved? What's the load for this test? Is it a stable load? 

Can you please share the correct schematic and layout ? And please also share the part number of inductor L7 and C157, C159.

Where did you place the Cin on the board, can you please attach a image?

Thank you.

Hi

Yes, adding RC can solve the upwash of SW, but the downwash is not solved, see the figure below

Sorry, our board has multiple DCDC, the schematic diagram and PCB position number do not correspond, see the figure below, the capacitor is GYA1E221MCQ1GS, the inductance is CSBL1265-4R7M, see the figure below for participation

Hi Li yu

One thing I want to check is that did the converter operate stably? There is CCM&DCM mode in the SW waveform. Is the load changing? 

Can you adjust the feedback resistors and the Cout according to the recommend component selection in the datasheet? Please use ceramic capacitors , not using electrolytic capacitors.

 BR.

Hi

The load is constant, electronic load 1A. The feedback resistance has not been changed, the output capacitance is 5*22UF according to the specification, and there is still -2V overshoot

Hi Li yu

Can you measure the SW, Vo with larger time scale and see if the circuit operates stable? Thank you

Hi li yu

The circuit seems operate unstable. can you place the C29? also did you use small test loop? can you please share a image of your test environment?

Thanks and best regards.

Hi

  I got the TPS56837H EVM, the load is 3A, SW,and there is also -3.65V overrush, which exceeds the recommended value in the manual. Please give me a reasonable explanation, thank you

Hi li yu

The spike is caused by the parasitic inductance and the fast transition of the switching current. There are parasitic inductance can be induced by IC package pins and pcb traces,  there should be no risk to the IC. Attach a related apps note: Understanding the Absolute Maximum Ratings of the SW Node

 The measurement is also important. Sorry that I don’t have the TPS56837H EVM on hand, I tested on TPS56837EVM, with Vin=24V, Vo=5V, Io=3A. Below is the test result under two different probe ways:

Thanks and best regards.

Hi

1.The test SW is tested using a ground ring, which is consistent.

2.Your SW test also has -2.9V, and the SW in TPS56837 below does not exceed -1V, which obviously exceeds the requirement, doesn't it? Or does the current SW exceeding requirements not affect DCDC functionality and performance?

For the second question,thank you

Hi LI,

Bella will reply to  you next week.

Colin

Hi  liyu

Please refer to the absolute maximum ratings in Table5-1, the -2V is DC value limit, the -5V,10ns is transient limit. The -1V in recommended operating conditions is DC value limit. 

The dc rating means that the part could be damaged if a constant voltage below –2 V is applied at the SW node from a source other than the output inductor during switching operation. there could be infinite current through the low-side MOSFET body diode, damaging the device.

It's normal to see a short-time negative voltage spike that can be lower than -2V during each switching period, this is not applied by another source and as said before, is caused by the parasitic inductance and the fast transition of the switching current. There are parasitic inductance can be induced by IC package pins and pcb traces,  there should be no risk to the IC. 

Thanks and best regards.

Hi

1. the -2V is DC value limit, the -5V,10ns is transient limit,On my side of the picture -3.56V is is transient limit,So the maximum value should be the value of -5V(<10ns), not -2V, Am I right in my understanding?

2.But I -3.56V should be  > 10ns,Confirm that there is no risk to IC?

These are my two questions, please answer, thank you

Hi li yu

1. Yes;

2. There is no risk to IC. the negative spike didn't even hit the -5V threshold.

Thanks and best regards.

Hi

I used the probe of x1 to test, and there is -6V at the foot of SW. I would like to confirm again that using the probe of X1 to test is more correct, right? And is there any risk to IC from the -6V downstroke? thank you

Hi liyu

yes you can use️1 ，full bandwidth.

Did you test evm? It should be ubder-6v using small probe loop from my experience，can you share the measurement environment photo? But as we stated before, there should be no risk.

BR

HI

Yes,  EVM test , the ground ring minimum path test, X1, full bandwidth test, -6V, I am worried about the risk

Hi

Thanks for looking at this problem

Hi,

Bella is OoO today, and she will back to office tomorrow. Will give you feedback later. Thanks for your patience.

Aurora

Hi Liyu

There should be no risk. We have never seen any issue caused from this negative spike .

I see that you probe on the SW-node of inductor, can you probe on the SW pin of IC? The spike should be smaller then.

Thanks and best regards.

Hi

SW pin of IC,The test is also around -6V, which is no risk to the circuit and IC？ 

Hi Liyu

Please let me test on the EVM when I got time. 

Thanks and best regards.

Ok, please let me know if you have the test result. I will pay close attention to it, thank you

Hi Li Yu

I test on TPS56837HEVM ,Vin=24V,Vo=15V, Io=3A. Probe on EVM the same as yours, the undershoot amplitude <2V. But I use a ×10 probe as I don't have a ×1 probe.

Can you please check with another probe?

Thank you