LM22678-Q1 Output Shrt to GND

Moving to the Simple Switcher forum...

Hi,

When the device enter current limit, the current is constant at current limit. It could be the part gets overheat.
Do you know the temperature of the device during short?

Thanks
-Arief

Dear, Arief-san. 

Thank you for your reply. 

Ambient is room temp because it was happened at Labo area. 

So, 20 ~ 25 deg.C. 

This device has frequency current foldback & TSD. The concern is 

why it was broken within quite short time. 

Best Regards, 

H. Sakai

Do you have scope waveform of the short or maybe schematic and layout?

THanks
-Arief

Dear, Arief-san.

Unfortunately, I don't have the waveform, schematic, layout.
Your advice like assumption is helpful for us.

Best Regards,
H. Sakai

Is the device get damage during the short or is it get damage after the short is released or initially applied?
You can also check which pin get damaged (an ohmmeter measurement from the pin to GND).

As this point i can see that thermal might be an issue because you are at 7A current limit. Therefore your inductor might saturate also if it cannot handle 7A. I hope the DAP on the back of the IC is soldered on as a heatsink

Knowing the damaged pin also helps to troubleshoot what happen

Thanks
-Arief

Dear, Arief-san. 

Thank you for your reply and sorry to disturb your job again. 

1. The device got damage during the short to GND. It was broken within quite short time. 

2. VIN & SW was shorted to GND. 

I am expecting your advice. 

Best Regards, 

H. Sakai

Is there any spike on the SW node or VIN node during this short?
Also do you know if the back metal is soldered down?

Thanks
-Arief

Dear, Arief-san. 

Again, They didn't capture the waveform. 

The back metal is no problem. 

We just know LM22678-Q1 has a kind of weakness against output short to GND

compared with other TI products. 

Best Regards, 

H. Sakai

Dear, Yang-san. 

Thank you so much for your valuable information. 

i) You pointed out the ringing on Vout, SW due to the short farther away from the output. 

2) LM22678-Q1 has the current limit & Low frequency current fold-back function. 

I wonder there is a possibility that 1) symptom become worse due to 2)?

Best Regards, 

H. Sakai

The symptom is better with the current limit protection and frequency foldback function. The peak current is limited by the protection and the DC current is reduced by frequency foldback.

Dear, Yang-san. 

Thank you so much fir your teachings. 

I checked LM22678-Q1 Maximum ratings on SW node. 

It is -5V to Vin. 

The other side, SOA defines "Input Voltage". 

If "Input Voltage" meant the voltage between Vin and SW, 

It was reasonable that IC was broken to be over SOA because

the customer used under Vin=18V and SW was reached -5V due to

the ringing at output hard short to GND occurrence. 

I wonder above my assumption is correct?

Best Regards, 

H. Sakai

It can be understood that way. There's also ringing on the ground pin, so the difference between VIN and GND could also damage the IC quickly. -Yang