TPS7A39: TPS7A39: Doesn't regulate properly

Hi Kheng

This looks a lot like a damaged device. The min and max values for Vnr/ss are tested at final test so when it leaves our fab site we know that it is between 1.172V and 1.208V.

1. At any point did you ever see the negative side fail? It is not tied to the internal bandgap Vnr/ss pin, so if it never failed we know the failure point is on the internal reference.

Kheng Lim said:

The two ICs built with different batch were tested in the same se-tup.

2. When you say different batch of IC's do you mean the exact same device TPS7A39 from a different wafer lot?

Kheng Lim said:

I repeated the same process to a second IC (externally fed 1.2V to the pin NR_SS), the IC outputs the correct voltages for 10mins. Then the positive output voltage rose to 11V but the negative still stayed at -1V.

3. The positive output was still being externally driven with 1.2V and the device was not regulating to 6V?

The high Vnr/ss of 4.4V would cause the output to be set roughly 4x as large at around 22V, so with an input of ~11V, the pass device would be fully on. The error is with the Vnr/ss internal bandgap reference. the device was trying to regulate to 22V but was not being given enough input to get there so it was driven fully on.

It looks like in the second case the pass device failed as output is just following input, meaning the pass device has most likely been damaged.

Regards,

John

1. At any point did you ever see the negative side fail?

When the TPS7A39 (5 different ICs witht the same production code TI93J P49D) were first powered, the positive output was 11V and the negative output was -11V and the Vnr/ss was measured 4.4V. 

2. When you say different batch of IC's do you mean the exact same device TPS7A39 from a different wafer lot?

Same device TPS7A39, production code TI93J P49D

3. The positive output was still being externally driven with 1.2V and the device was not regulating to 6V?

The IC was externally driven to 1.2V. The positive side regulating to 6V at first, after 10mins the voltage shot up to 11V and stayed there. The negative side remained at -1V.

Hi Kheng,

Yes I misremembered the function of the Buffer pin, it does have a connection to the Vref, that is buffered and compared to GND using the low side feedback network.

In the condition where you had the device operate at 11Voutp, and -1Voutn. ie. the negative side was operational and the positive side was not functioning, What was the Vbuf at? There is an internal spec for the offset voltage from Vbuf to Vnr/ss and it is on the order of <|10mV|. if the Voutn is regulating but Voutp is not, I suspect that this buffer voltage is out of spec and differs from what is seen when the part was tested before being shipped out.

Could you share the schematic?

Regards,

John

Hi John,

The Vbuf had almost the same voltage as Vnr/ss which was 1.22V.

Hi Kheng,

If Vbuff and Vnr/ss are both at the correct voltage, but only the negative rail is regulating, then the pass device is not operational. This would exhibit as a failure as suddenly VOUT stops regulating and follows VIN as you've seen.

Looking at your schematic, I don't see any glaring concerns, which makes sense as you already used this schematic successfully.

At any point during startup is there a sequence or test procedure that is run that may tie the INP to a >12V rail?

Or is the trace for nr/ss close to a voltage rail that may exceed VINP during testing or handling of the device?

I will talk to management to see if this is a full lot failure issue, as there does not seem to currently be a problem with your schematic or setup.

Regards,

John

Hi John,

Thank you so much for your explanation.

+12V is the highest voltage available on the board. Will the inrush current damage the IC as there is no series resistance to limit the current?

My main concern is of all the boards that we bult, all the regulators exhibited the Vnr/ss outside of that voltage tolerance listed in the datasheet. Therefore, we cannot safely use the regulator in our application as it is.

I can send you some of the regulators that were removed from our boards if you like and want to perform IC characterization or further tests.

Hi Kheng,

You have 10nF on NR/SS defining a softstart functionality, and with 10uF on your output side capacitors, I don't think that inrush is a concern with blowing the part.

Lets do two things in parallel.

Kheng Lim said:

I can send you some of the regulators that were removed from our boards if you like and want to perform IC characterization or further tests.

1) Begin returns process using the following links:

https://www.ti.com/support-quality/additional-information/customer-returns.html

https://www.ti.com/productreturns/docs/createReturn.tsp

2) If able, it would be great to see an ABA swap on the current and old boards. I know the schematic was copied over exactly, but if you could take a known good device and take it off a known good board, and place it on the new failing board to test whether it still operates, then place it back on the original board to make sure it still operates. Simultaneously take the failing device and place it on the known good board, see if it regulates, then place it back on the failing board to see if it regulates or fails regulation again.

Regards,

John