TPS54335A: Potential sporadic failure in invertering design

Steven Shenker

Part Number: TPS54335A

Hi there,

I have used the TSP54335A to design a 3.75V 2A negative output from a 9V supply, using a SW freq of 600KHZ. I have based the design around the TI inverting design guide for the same part. See: SLVA910A.

I also referred to the layout guide and believe my layout to be very close to the example in SLVUB76A.

I have done all calculations as per the guide and have also copied the layouts as per the guide. I have attached this below. The problem is that we just made 4 units of this device and 1 of the 4 units is outputting +0.5V instead of -3.75V. I am not sure where to start looking for failure debugging. I noted that the Webench calculated values for Rcomp and Ccomp1 abd Ccomp2 are different to mine, however I calculated these values as per the guide and this should be correct. (I didnt use Webench for the calculations as I am not sure whether these calculations would work for an inverting case).

Schematic:

Layout:

This is the V- output voltage as measured on the 1/4 incorrect device:

Voltage measured at PH pin 2:

Voltage measured at BOOT pin 9 (there is a small AC amplitude above of the DC signal if you look carefully):

It is strange because it is outputting a value in the opposite polarity. IE if this was not being using an inverting applicating it would be outputting -0.5V which seems implausible. Please let me know where can I investigate to fix this issue.

Many thanks!

over 2 years ago

0 Zixu Wang over 2 years ago

TI__Expert 8100 points

Hi Steven,

My colleague will give you reply by next week!

BRs

Zixu

0 Ruby Li over 2 years ago

TI__Expert 7520 points

Hi Steven

The webench works for the inverting buck boost, just change the input voltage to be 9+3.75V.

I have checked your design with Webench, it seems that the phase margin is 33degree.

Would you please share me the test method? What is the reference and what is the test point?

Hope this helps.

Ruby

0 Steven Shenker over 2 years ago in reply to Ruby Li

Intellectual 930 points

Hi Ruby,

Many thanks for your response.

Is this an adequate phase margin? Did you check with my design or the Webench components? I am unsure whether to change to the Webench recommended compensation components or not?

Regarding the test method, I assume you mean the scope tests points? The test points for the above are all listed but the reference for the measurements is DGND (IE the output of the inductor).

0 Nancy Zhang over 2 years ago in reply to Steven Shenker

TI__Mastermind 49370 points

Hi Steven,

Please see my slide as attached. Hope this helps.

TPS54335A IBB application.pptx

Thanks,

Nancy

0 Steven Shenker over 2 years ago in reply to Nancy Zhang

Intellectual 930 points

Hi Nancy,

Many thanks, I now see the schottky diode in the TI design I was working from to protect against this condition - although I will remove the CAPS as they are not recommended in your document. I am relieved to have found a possible cause of failure.

Final question to yourself or Ruby as I am not an expert in this area, can you comment on the phase margin based on my current R1 C1 C2 compensation setup as per Rubys comment?

Thanks again,

0 Nancy Zhang over 2 years ago in reply to Steven Shenker

TI__Mastermind 49370 points

Hi Steven,

Thanks for your information. If the input caps GND connected to -Vout, you may see a positive voltage (0.5V) sitting on the negative rail. Do you mean you found the possible cause of failure? Could you share with us?

Regarding the phase margin improvement, Ruby will comment on that.

Thanks,

Nancy

0 Steven Shenker over 2 years ago in reply to Nancy Zhang

Intellectual 930 points

No I thought that your explanation was a possible cause of failure. I have done some reading and in the inverting design guide for the TPS54335A it does require the bypass cap as I have it (see C3 and C4 in the TI design guide):

www.ti.com/.../slva910a.pdf

Could you comment on this? I wonder if this IC is different to yours in this respect?

+1 Nancy Zhang over 2 years ago in reply to Steven Shenker

TI__Mastermind 49370 points

Hi Steven,

1. I am not quite sure why TPS54335A needs the input caps placed between Vin and -Vout (device's GND). But it does add a Schottky diode D1 on the schematic. You can try to add a Schottky diode while keep C24/C39 no changing.

2. For lower phase margin, you need to add more output caps to decrease crossover frequency due to right half plane (RHP) in IBB topology. Also need to adjust your loop compensation values. Please refer to my last slide as attached. Hope this helps.

7380.TPS54335A IBB application.pptx

Thanks,

Nancy

0 Steven Shenker over 2 years ago in reply to Nancy Zhang

Intellectual 930 points

Thanks Nancy, I will give this a try. I appreciate the comprehensive responses.

0 Steven Shenker over 2 years ago in reply to Nancy Zhang

Intellectual 930 points

Hi Nancy, quick question on #2 above. I noticed that you have not selected my frequency of 600khz in webench. When I do this my values are slightly different to that of yours:

Can you confirm that I should use my values in the red square as per Webench at 600khz. Thanks

+1 Nancy Zhang over 2 years ago in reply to Steven Shenker

TI__Mastermind 49370 points

Hi Steven,

Your value looks good since you selected 2x47uF output caps. Please let us know if these solve your issue.

Thanks,

Nancy

Power management

Power management forum

TPS54335A: Potential sporadic failure in invertering design