TPS25944A: Reverse polarity - high input current and overheat ev

Arnaud Puret

Part Number: TPS25944A
Other Parts Discussed in Thread: TPS25947

Hi,

I have implemented a reverse polarity protection for the TPS25944A as described in its datasheet and previously discussed in this post.

When tested in normal conditions, the eFuse is working well, no overheat.

When tested with input polarity reversed, the eFuse doesn't block its output:

-0.5 V was read at the eFuse output (maybe linked to the voltage drop of the blocking diode?),
a lot of current was drawn by the load, up to the limit set to the power supply (1 A),
the eFuse overheats (the thermal snapshot was taken 3 s after power on, heat was still rising after that).

What is the problem? From the datasheet, the "TPS25944 can be used as an eFuse or ideal diode" and during this situation the "internal FET blocks the reverse path" (from 10.3.1.4, p. 44).

Thank for your help.

Best regards,

Arnaud Puret

over 4 years ago

0 Avishek Pal over 4 years ago

TI__Mastermind 39550 points

Hi Arnaud,

Thanks for reaching out to us. We are looking into your query and will get back to you by end of the day today.

0 Avishek Pal over 4 years ago in reply to Avishek Pal

TI__Mastermind 39550 points

Hi Arnaud,

Please could share the following experimental waveforms during input reverse polarity test keeping the switch S1 open.

1) Input current, 2) Output voltage, 3) Output current, and 4) Voltage across the diode D4 (Anode to Cathode).

0 Arnaud Puret over 4 years ago in reply to Avishek Pal

Prodigy 80 points

Hi Avishek,

Thanks for your quick answer.

Here there are some of the requested waveforms. I can't provide output current, the eFuse output is connected to numerous parts through the PCB internal layer.

Important note: the input current is limited to 1 A by the power supply (input voltage reduce to ~3.5 V in this case). With a higher input curent limit, the same waveforms occur.

May this help!

Best regards,

Arnaud Puret

0 Avishek Pal over 4 years ago in reply to Arnaud Puret

TI__Mastermind 39550 points

Hi Arnaud,

Thanks for sharing the waveforms. Initially it is working fine. But concern is what causes this transient and getting all things messed. Do you have any idea? I am also thinking.

0 Avishek Pal over 4 years ago in reply to Avishek Pal

TI__Mastermind 39550 points

Hi Arnaud,

Please could you perform the same test after removing D3. Kindly zoom and measure the initial input current.

0 Arnaud Puret over 4 years ago in reply to Avishek Pal

Prodigy 80 points

Hi Avishek,

Initially (before the trigger), the power supply is not connected. Vin is measured at the supply, before the connector of the board (used to start the test).

On the waveforms, the -12 V of Vin and the ~0 A of Iin is OK in your red area (nevertheless, the ~0.3 V of Vd in questions me...).

I will see what I can do with D3, I must not break it during the unsoldering process and I have not the required tools...

May this helps!

Best regards,

Arnaud Puret

0 Avishek Pal over 4 years ago in reply to Arnaud Puret

TI__Mastermind 39550 points

Hi Arnaud,

Please share the waveforms if you are able to perform the test after removing D3. Actually, I am not finding anything wrong with the schematic what you have shared.

0 Arnaud Puret over 4 years ago in reply to Avishek Pal

Prodigy 80 points

Hi Avishek,

I couldn't properly unsolder D3 so I broke it to perform the test...

Unfortunately, the problem remains and the waveforms are the same. Vout seems to getting worse with -1V instead of -0.5 V with D3!

What I can do/test from now?

Best regards,

Arnaud Puret

0 Avishek Pal over 4 years ago in reply to Arnaud Puret

TI__Mastermind 39550 points

Hi Arnaud,

Extremely sorry for the inconvenience. Honestly, I unable to find what is the issue here. Is it possible to replace this device with an another sample of it, if you have? May I know that do you have TPS25944X635EVM? Then, I will try to explore if it's possible to test in EVM.

0 Arnaud Puret over 4 years ago in reply to Avishek Pal

Prodigy 80 points

Hi Avishek,

I have 10 prototypes which can be tested and all have the same problem. On the other hand, I haven't any TPS25944X635EVM to experiment with...

Best regards,

Arnaud Puret

0 Avishek Pal over 4 years ago in reply to Arnaud Puret

TI__Mastermind 39550 points

Hi Arnaud,

May I know that where are applying negative voltage? I think that it's between the nodes +12V-IN/2.1A and +12V-RTN/4.3A. If I am correct, then +12V-RTN/4.3A net should be connected to cathode of D4, not the device ground.

0 Arnaud Puret over 4 years ago in reply to Avishek Pal

Prodigy 80 points

Hi Avishek,

Negative voltage is applied between +12V-IN and GND (from the power connector of the board).

+12V-RTN is used the measure the D4 voltage drop and compute the right current which go through the eFuse, from +12-MON voltage.

May this missing precision helps you to find my mistake!

Best regards,

Arnaud Puret

+1 Avishek Pal over 4 years ago in reply to Arnaud Puret

TI__Mastermind 39550 points

Hi Arnaud,

May be 12V-RTN is being held at 0 V or some positive voltage in the system. Please leave this node completely open or disconnected from device ground and try.

0 Arnaud Puret over 4 years ago in reply to Avishek Pal

Prodigy 80 points

Hi Avishek,

Good point! Sometime, mistakes yell at you and you don't wand to listen...

+12V-RTN is connected to an ADC but when the power supply has reverse polarity, the ADC isn't working and its internal structure makes probably its inputs connected to GND...

Glad you investigate this problem until now. Thanks a lot!

Best regards,

Arnaud Puret

0 Avishek Pal over 4 years ago in reply to Arnaud Puret

TI__Mastermind 39550 points

Hi Arnaud,

Sorry for the inconvenience. We have an integrated reverse polarity protection solution, which is TPS25947. Please look at once this device. If your project schedule allows, kindly use TPS25947.

Power management

Power management forum

TPS25944A: Reverse polarity - high input current and overheat ev