TPS7A4001: Component overheating

Hello Fabien,
Based on the described operating conditions the power dissipated in the TPS7A4001 package is 0.7172W
Pd = ( ( 49V - 5V ) x 16.3mA) = 0.7172W

IF the ambient temperature is 25C, and R(theta)ja is at the expected 66.7C/W, then the junction temperature would be 72.8C.
Tj = (25C + (66.7C/W x 0.7172W) = 72.8C

My guess would be that the thermal pad on the bottom of the package is not soldered to an adequate copper area, or there are not enough thermal vias, to remove heat from the package. This will cause the actual R(theta)ja value to be much higher than expected.

Dear Daddio,

Thank you for your quick reply and your advice concerning power dissipation.

If we have a look on the PCB layout below, we have thermal pads under the component (U4) accompanied by thermal vias (N° 9). We also have the two resistors at the output (R7 and R6), the output capacitor (C9) and the input capacitor (C8). Below the component (at the PCB bottom) we have a complete ground plane for thermal dissipation. In spite of this, the temperature rises up to 126°C for some of the products.

The products come from the same production line, the TPS7A4001 from the same rolled package. Is it possible to have so many differences among the TPS7A4001? Have you ever had these bottlenecks concerning the TPS7A4001?

Thank you again,

Kind regards,

Fabien Bibi

Hi Fabien,

None of your images are properly attaching to your posts.  Please review the following to ensure that we are able to better respond to your specific application:

There should not be a wide variation on how individual TPS7A4001 devices perform thermally when used in the same application (same input voltage, output voltage, output current, layout, and ambient conditions such as air flow and temperature).  As Don suggested how well the thermal pad is attached to the PCB will directly impact thermal performance.

Very Respectfully,

Ryan

Hello Fabien,

Unable to view any PCB layout, but your description leads me to believe that it would be adequate.

The number of variables to sometimes cause an unexpected increased temperature is limited.

Since the silicon die is attached directly (i.e. glued) to the exposed thermal pad, the likelihood that there is any wide variation in the thermal resistance ( R(theta)ja ) package to package is not likely.

Solder coverage between the exposed thermal pad and the PCB copper could probably be as low as 60% before R(theta)ja would be negatively affected in any significant manner. Of course, 100% solder coverage would be preferred.

Oscillation of the TPS7A4001 output could possibly cause overheating, but this would probably affect the output DC voltage. In any case this possibly should be checked with oscilloscope. If the output is oscillating, increasing the output capacitance value should help.

If the load current is higher than expected, that will cause dissipation to increase, and temperature to increase. If it is possible, confirming the load current on overheating devices would be useful.

Hi guys,

Thank you very much for your answers. We'll take your advices into consideration and strengthen our PCB layout on the thermal pad under the component . In addition to this, we'll use thermal dissipators to increase the cooling of the device which I guess will help a lot.

I think that these two modifications should solve the problem.

Thank you again for your help.

Kind regards,

Fabien