TPS63024: VOUT becomes unstable as PCB heats up

Capture below shows VOUT and the FB pins when the output becomes unstable

Capture below shows the VOUT and VIN pins. Note a 3.23V DC offset was applied on the scope for VIN to zoom in.

Hi  Tony,

could you help to capture voltage of EN same time? 

Regards

Tao

Hi Tao, Below are some captures of the EN pin together with VOUT.

First one below shows triggering on EN pin and showing VOUT at initial startup.

Capture below shows triggering on EN pin and showing VOUT on shutdown

Capture below shows trying to tigger on EN pin falling edge, while showing VOUT. The scope never triggers, so the EN pin is always HIGH while VOUT becomes unstable.

Following shows triggering on VOUT while unstable and showing EN. I never see EN pin go low.

As a hardware guy, I was hoping it was a firmware bug causing EN pin to glitch low, but that isn't the case

Thanks for the help,

Tony

I also want to add that today I noticed if I shutdown as much as possible on the PCB and enable this power supply the output remains stable. If I then apply heat on the PCB with a hot air gun, the output transitions to unstable. Once I remove the hot air, the output transitions back to being stable.

Hi   Tony,

thanks for your detailed description, that is really good.

so  i think OTP had been triggered, there is no other logic can make device stop switching if EN is stay high.

but i am really confused why you have a 47degree in your test, it should be much higher,

Regards

Tao

Below is a video with a thermocouple placed on the case of TPS64024 with some thermal paste. At 47C the instability begins. We have a few devices that exhibit this problem, so it isn't just one device.

Here is a top side view of device. Perhaps the date/batch code can tell you something?

Here is BGA side of a device removed today from a PCB exhibiting the same problem. Two balls on top left probably got damaged after removal.

Hi   Tony,

is it possible that during the manufacturing process, the chip mounting ball is not in close contact with the PCB, resulting in an increase in the thermal resistance of the chip. then, although the temperature of case is 47, but the junction temperature is over ABS one?

for the topmark, we got some info below.

This is Assembled in TI Clark Philbin Data code 2020 Feb 

Regards

Tao

Hi  Tony,

may be you can find something different between bad and good board by X-ray.

Regards

Tao

Hi Tao,

I'll ask for XRAY and see if that shows anything.

If the part isn't soldered down good, I don't think it would change the junction to case thermal resistance?

The datasheet specs the Junction to case thermal resistance as 0.5C/W. With a 47C case temperature we would need to be drawing 186Amps to hit the 140C junction temperature. This is not possible. Something doesn't add up.

Tony

Hi   Tony,

actually, bad connection between chip and borad will impact the thermal resistance. you can take the picture below as ref.

yes correct, but consumption are not only conduction loss but also some switching loss.

Regards

Tao

Hi Tao,

I captured some thermal images of the switching chip U14, and also of the inductor L3 at the time VOUT becomes unstable.

First image below shows the top side of the PCB. Image on left is visible spectrum, image on the right is FLIR thermal image. I had to blank out remaining area of PCB that is proprietary.

SP1 points to the case temperature of U14 which is 53.3C.

Bx1 has a red triangle that shows the location of the max temperature at 55.0C in the boxed area around U14. The box contains the input & output capacitors, and also the PCB temperature near U14.

Bx2 is the box around the whole PCB with location of max temperature at 65.3C.

Second image below shows the bottom side of the PCB. Image on left is visible spectrum, image on the right is FLIR thermal image. 

SP1 points to the case temperature of the inductor L3 which is 40.9C. It is directly on the opposite side of U14.

Bx1 has a red triangle that shows the location of the max temperature at 51.7C in the boxed area around inductor L3 and also the PCB temperature in that area.

Bx2 is the box around the whole bottom of the PCB with max temperature location at 61.2C.

Based on these thermal images, I'm having a hard time believing the junction temperature of U14 is at 140C. Given we have seen this problem on multiple PCBs we believe we have a bad batch of parts. 

We plan on sending this PCBA to get U14 replaced, but need to understand what is happening to prevent possible problems in the future.

Thanks,

Tony

Hi Tony,

There is almost no chance that the product will have batch consistency issues about your phenomenon.

also,  i am not 100% sure OTP was triggered, but seems no other logic can make device stop switching. anyway, root cause is also important to us. Hope this issue will go after chip replaced.

Regards

Tao