TPS54239 overheats and self-destructs with 12 volt input

Neil Stanton

Other Parts Discussed in Thread: TPS54239

This chip is on a step-down circuit board very close to TI's reference design and I'm having sporadic failures of the chip (4 failures in approx 30 boards.)

The boards were assembled by a manufacturer I have used with no previous problems. This failure is serious because boards test okay before I ship and customers are finding the problem when things begin to smoke.

It's possible the solder-pad on the under-side of the chip is not bonding reliably to the circuit board. Any suggestions to fix this problem?

Neil Stanton.

over 7 years ago

0 JohnTucker over 7 years ago

TI__Guru**** 191445 points

You may need to have one or two boards x-rayed to check for solder coverage. What size thermal vias did you use?

0 Neil Stanton over 7 years ago in reply to JohnTucker

Prodigy 50 points

There are six 15 mil vias under the chip. It's a 2-layer 0.032" board. ground plane on back side of board. Good boards do not get hot under load, which is only about 500 mA at 4.5 volts output voltage.

0 Eric Guo1 over 7 years ago in reply to Neil Stanton

TI__Intellectual 2590 points

Just those 4 failure parts has problem and the other board can work well? TI will test every part before ship them to customer.
would you please replace the IC with new one to see whether the board can work well?
Soldering problem is greatest possible for this problem.
thanks!

0 Neil Stanton over 7 years ago in reply to Eric Guo1

Prodigy 50 points

The question I really want answered is:
Can TI verify that improperly grounded pad on back of the the TPS54239 cause chip failure as soon as input voltage is applied?

0 JohnTucker over 7 years ago in reply to Neil Stanton

TI__Guru**** 191445 points

So far as I know there is no electrical requirement for the pad to be soldered, only thermal. In the event of over heating, the device should be protected by the internal OTP. If there is some other solder defect that causes pins to be unintentionally shorted, there may be damage to the device. SW to GND and VIN to VBST in particular may cause the device to fail.

0 Neil Stanton over 7 years ago in reply to JohnTucker

Prodigy 50 points

If grounding the chip pad is not critical (except thermal), my only explanation for the failures is not correct. Seems like a fairly simple test to deliberately isolate the pad grounding and see what happens. It's obvious from the failure that the internal OTP did not protect the chip. The failure is within a second or 2 and hot enough to cause plastic to melt. All the characteristics of a hard short-circuit in the IC.

0 JohnTucker over 7 years ago in reply to Neil Stanton

TI__Guru**** 191445 points

Yes it sounds like EOS to me. Are you hot plugging the input supply?

0 Neil Stanton over 7 years ago in reply to JohnTucker

Prodigy 50 points

Depends what you mean by hot plugging. Input is controlled by a switch and voltage can be anything from 5 to 16 volts DC depending on customer application. I have no control over quality of customers' power supply except documentation requiring DC between 5 and 16 volts. As a precaution, input goes to a bridge rectifier and filter capacitor before connection to IC Vin.

0 JohnTucker over 7 years ago in reply to Neil Stanton

TI__Guru**** 191445 points

Can you monitor the input voltage to look for over voltage transients at start up? TPS54239 is not very tolerant of over voltage conditions. There is some margin in the absolute maximum rating but not much.

Power management

Power management forum

TPS54239 overheats and self-destructs with 12 volt input