TPS61199: Driver burns immediately with VIN 24V and EN only switched by 5V

Hi everybody,

just an short update: 

I did the 5V for Enable Pin test on the eval board and the TPS61199 died immediately (removed the jumper cap from JP2, applied the +5V from my laboratory PSU to the EN Pin on the Jumper JP2). 

There is a 10k series resistor between JP2 and EN Pin of the TPS61199... what is the problem with the EN pin?

Best regards,

Stefan

Hi Stefan,

There seems to be multiple things happening here.

"After finding this out I disconnected the microcontroller on our LED driver from the EN pin and inserted the resistor divider as it is on the eval board -> result is the driver does not burn and works."

The Absolute Max rating on EN and PWM pins is 20V as per the datasheet. When input is 24V and with resistive divider, the voltage EN sees is less than Abs max, so it didn't burn the device.

However, it doesn't make sense why +5V from microcontroller would kill the unit. Is your ground at the proper voltage. Is there a change that a value beyond 20V is getting applied at EN pin?

Thank you,

-Harini

Hi Harini,

thank you for the reply.

Ok, understand the max. rating of EN and PWM Pin and also the resitor divider to reduce the VIN to max. 20V on the Eval Board.

On our LED driver a microcontroller (powered by 5V generated with a LDO from the 24V VIN, GND is the same as for the TPS61199) is connected over a 10k series resistor to the EN (and PWM) Pin of the TPS61199. As long as the microcontroller pulls the EN Pin low everything is fine. But as soon as the microcontroller switches the EN Pin to 5V the TPS61199 gets killed. This does not make sense to me either since the EN should be a logical input. 

However to exclude an error on our LED driver I did the test with the Eval Board, see picture attached for the setup. Even in this configuration the TPS61199 burns immediately after connecting the 5V on the EN Pin.

The same test but with the jumper cap on the EN jumper pins 1 and 2 (so using the resistor divider for the EN voltage) the TPS61199 works fine.

Any idea why this is happening? Any suggestion how it could still be done to enable the TPS61199 by the 5V microcontroller? 

Thank you,

Stefan

 

Hi Stefan,

Please add a 100k resistor between the gpio of the mcu and the enable pin of the tps. Try again grabbing the waveform on the 2 terminals of the resistor.

I dread the gpio is too fast that cause ov/uv on the pin and destroys the part.

Also would be interesting to know where the damage is (en pin shorted to gnd or the output fet)

KR

Vincenzo

Hi Stefan,

Is the issue you noticed resolved? Is there anything we can help with?

Best,

-Harini

Hi Vincenzo, Hi Harini,

sorry for my late reply, needed some time to test your suggestion.

The 100k work only sometimes. I did some turn on/off tests and from time to time the driver still gets killed.

Attached are some oscilloscope pictures from the turn on waveform.

CH1 is the waveform on the 100k series resistor for the EN-Pin (the MCU GPIO side).

CH2 is the waveform on the 100k series resistor for the EN-Pin (the TPS61199 EN Pin side).

CH3 is the VDD output of TPS61199

CH4 is the supply voltage of TPS61199

Unfortunately I did not capture the waveform when the driver gets destroyed, I will try to get it captured.

After destroying the IC the VDD pin (output of the internal voltage regulator) is shorted to GND. In some cases also VIN is shorted to GND.

With the 100k the EN signal is rising "slowly" so the ov/uv should not be the problem? Do you have other ideas what there is happening?

I will try to capture the startup waveform when the driver gets destroyed.

Best regards,

Stefan

ok I see,

as I suspected, when you rise the EN the internal LDO turns on and then burns out, becasue of the too high VIN-VDD gap.

this doesn't happen at 12Vin but at 24V the gap is much higher... I can't say exactly why the internal LDO dies, but it is something I already saw many times (tipically when EN is not biased and VIN ramps up quickly...)

what you can do is to reduce the stress on the LDO:

- add a little R in series between the VIN rail and the VIN pin of the IC, 10 Ohm likely

- reduce C3, down to 1uF

- improve the layout (EVM works fine)

hope this helps a bit

KR

Vincenzo

Hi Vincenzo,

thank you very much for your quick reply and your suggestions.
I will test your suggestions, but this needs some time, since I'm currently involved in other projects.
I will reply when I have found the time to do the tests.

Best regards,
Stefan