TPS54334: Chip failure causes output (pin 3) to be shorted to ground

Gajadhar Sewnath

Part Number: TPS54334

I have a circuit that was built for the TPS54336ADDAR chip (first picture). Due to supply issues, the TPS54336ADDAR you see in the first picture was substituted with the TPS54334DDAR regulator by the CM. We are seeing a high failure rate of this chip in the field (20%). The load for this circuit is less than 200mA. The scope traces of the output vs input voltages on the board does not show anything unusual. However, TI's WEBENCH tool shows different components for the compensation network, switching inductor and pullup on PGOOD. Can anyone tell me if these component differences are enough to cause this chip to fail?

My circuit:

WEBENCH's circuit:

over 3 years ago

0 Zixu Wang over 3 years ago

TI__Expert 8100 points

Hi Gajadhar,

Could you please share what is the phenomenon for the failure?

BRs

Zixu

0 Ruby Li over 3 years ago

TI__Expert 7520 points

Hi Gajadhar,

Would you please tell me the LOT code of the devices? Is there any load transient in the application?

Is there any waveform like PH of the burned device?

Thanks.

Ruby

0 Ruby Li over 3 years ago in reply to Ruby Li

TI__Expert 7520 points

Hi Gajadhar,

Add one more comment here. The inductor is a little bit large for the application. Would suggest to use 6.8uH. Hope this helps.

Ruby

0 Gajadhar Sewnath over 3 years ago in reply to Zixu Wang

Prodigy 10 points

Hello Zixu,

I have 2 types on my desk:

1. The TPS54334 overheats to the point that the solder mask around the device bubbles.

2. There is a 1mm hole created in the package that is quite visible from the top.

Both these chips have shorts between pin3 and ground

0 Gajadhar Sewnath over 3 years ago in reply to Ruby Li

Prodigy 10 points

Hello Ruby,

The LOT code of the 2 devices is the same and is: TI 7AA (line2) P3PQG3 (line 3).

What do you mean by waveform PH? The device is dead already. I will not be able to get any waveform out of it now.

0 Gajadhar Sewnath over 3 years ago in reply to Ruby Li

Prodigy 10 points

Hi Ruby,

I understand the inductor is large, but will this large inductor be the cause of the TPS54334 burning out?

0 Gajadhar Sewnath over 3 years ago in reply to Gajadhar Sewnath

Prodigy 10 points

Also, I thought that this chip had thermal protection. One of the chips looks like it went through thermal meltdown. How is this possible with the built in overload and thermal protection? Under what circumstances will this protection not work?

0 Ruby Li over 3 years ago in reply to Gajadhar Sewnath

TI__Expert 7520 points

Hi Gajadhar,

The lot code seems not correct. Here is one example below. Would you please do me a favor to check again?

I was meant the waveform of the good devices. Would you please do me a favor to share the waveform of the PH, inductor current and output voltage together?

Thanks.

Ruby

0 Ruby Li over 3 years ago in reply to Gajadhar Sewnath

TI__Expert 7520 points

When the inductor is large, it may be unstable for the application. Hope this helps.

Ruby

0 Ruby Li over 3 years ago in reply to Gajadhar Sewnath

TI__Expert 7520 points

Hi Gajadhar,

I would like to confirm more info here:

1. The TPS56336 was used for the application before and did not have any burn out. When changing to TPS54334, there was 20% failure?

2. Under which condition the chip burned out? Steady state or power on off?

3. How many devices were burned out out of total?

4. Where does the input voltage come from? Would you please also do me a favor to check the waveform of the input voltage so that I would check whether there was any spike?

Many thanks.

Ruby

0 Gajadhar Sewnath over 3 years ago in reply to Ruby Li

Prodigy 10 points

Hello Ruby,

1. TPS56336 was never used, because of sourcing issues. We only used the TPS54334 in a circuit that was designed for the TPS54336.

2. The chip burned out on power up.

3. We had 2 devices burn out from a total of about 10 systems.

4. Input voltage of 24 volts comes from an OTS medical grade supply: Delta: model#: IMA-S1000-24-YYPLI (24V-42A). I have attached the input and output waveforms from a good board for you to analyze. Yellow is input supply, green is 5V output from TPS54334, Orange is 3.3V supply derived from 5V supply onboard.

0 Ruby Li over 3 years ago in reply to Gajadhar Sewnath

TI__Expert 7520 points

Hi Gajadhar

Thanks for your quick feedback. The phenomenon of the damage seems like there was over input voltage when powering on. Is there any bandwidth setting in the scope? And there is glitch when powering on, could the slew rate of the input voltage be adjust? If yes, please change that to 10V/ms. The length of the input wire should be reduced and adding input cap may help here.

What's more, would you please send me the layout if it is possible?

Thanks.

Ruby

0 Gajadhar Sewnath over 3 years ago in reply to Ruby Li

Prodigy 10 points

Hi Ruby,

I am out till next week Tuesday and will do the experiment when I return. I am sure the input to the scope was not bandwidth limited. I will try setting the slew rate as you recommended and get back to you next week.