LM5176-Q1: schematic review

Hi Stefan,

My customer got the board today. There are 2pcs LM5176-Q1 in their board. After they power on(LM5176-Q1's condition, VIN=12.5V, Vout=15V and load is open), one LM5176-Q1's output is abnormal(the output waveform is same with SW waveform) while another LM5176-Q1 works normally. There two device have same schematic, they will further check whether the chip is damaged.

In parallel, they want to know whether Vgs will be larger than 20V at the moment of power up.

Thanks & Best Regards,

Sherry

Hi Sherry,

VGS (Voltage between SW1 and HDRV1) is generated with a small charge pump (build with the Diode and Cap on BOOT1). So this Voltage can only be up to the Vcc voltage.

Best regards,

 Stefan

Hi Stefan,

The customer has new discovery. In their board, there are two LM5176-Q1, one LM5176-Q1 powers SOC1, and another powers SOC2. These two LM5176-Q1's schematics are identical, and these two devices' load are also same. Currently the load is open. 

In their board, the LM5176-Q1 that powers SOC1 can't output correctly while the another LM5176-Q1 that powers SOC2 outputs correctly. They test another board, the result is same. This phenomenon is very strange because these two LM5176-Q1 have the same schematic. 

Their test condition is VIN=12V, VOUT=15.5V, load is open. We tested some waveforms as attached. Our support needed are as below, 

1. Could you please help provide some debug ideas for this fault phenomenon?

2. When VIN>VOUT, our device works at buck mode; if VIN<VOUT, our device works at boost mode. It seems that the SOC1 LM5176-Q1 can't enter into boost mode correctly. Could you please help clarify how does our device determine which mode should be entered at power-on?

1321.waveform.zip

SOC1_LM5176.pdf

Thanks & Best Regards,

Sherry

Hi Sherry,

can you add some notes what the various scope plots in the zip file does show.

Thanks,

 Stefan

Hi Sherry,

as there are two version of the hardware - it would be good to have a scope shot of each version (working and none working) with the always the some conditions (Vin / Load)

Is the layout for both versions also identical ?

Thanks,

 Stefan

Hi Stefan,

The name of waveform shows the test point. For example, the third picture is ''ch3_t10008_s   ch4_t10008_g''.

t10008 is the name of mosfet as the second picture. Ch3 means that t10008's source voltage and ch4 is t10008's gate voltage.

For the working LM5176-Q1, T10007 is always high signal, T10008 is off, T10009 and T10010 are switching. The layout maybe different, I need to check further.

Thanks & Best Regards,

Sherry

Hi Sherry,

i am not sure on how the measurement setup was but it looks like there was high noise pickup. So they might have used a very large ground loop for the scope probes (Note: the default ground connector already forms a very big ground loop).

Based on that can you please check the measurements the lower two FETs again.

Can you also zoom in to see a few clock cycles.

Can we compare the layout of the two systems - if they are not identical - esp. on how the layout for the sense resistor is done.

Best regards,

 Stefan

Hi Stefan,

Yes, you are correct. The root cause is the sense resistor. In the layout, this sense resistor doesn't connect to GND directly, so the sense resistor acts as a large resistor.

Just curious, could you please help clarify how does our device determine which mode(buck, boost or buck-boost) should be entered at power-on?

Thanks & Best Regards,

sherry

Hi Sherry,

I have not seen an update on this thread for the last 11 days.

If the issue is solved it would be great to press the is solved button.

If there is still something open, just let me know with the new details.

Best regards,

 Stefan