LM3150: Switching stopped for a long time after OVP detection

ACGUY

Part Number: LM3150

Hello E2E,

Our system has an USB port, that VBUS power from LM3150.
When connected the USB, occur the surge on VBUS and Vfb, LM3150 stop switching. Stop time is about 50usec. When during stop the operation, LG=High, HG=Low.
Therefor, Vout will be descend to about 3V, resume operation.

Question,
When Vfb=0.36V at Vout=3V. The voltage is lower than Vfb-ov.
Why doesn't device resume operation quickly?

Regards,
ACGUY

over 5 years ago

0 Samuel Jaffe over 5 years ago

TI__Mastermind 25590 points

ACGUY,

Thanks for sharing the waveforms. Can you also share this waveform while also probing the output current? We may be encountering OCP from a hot-plug.

Please also share the schematic and layout.

Thanks!
-Sam

0 ACGUY over 5 years ago in reply to Samuel Jaffe

Guru 11170 points

Hi Sam,

I attached schematic. As a matter of fact this system is customer's. I am an engineer of distributor in Japan. I received request for clarification of cause about this phenomenon from customer. They refused to submit the layout diagram.

About current waveform, I requested to them. They took it. But when using current probe with oscilloscope, not occur it. Please see attached png file.

I confirmed circuit value for around LM3150. I-lim point is bery high. I think that it is no problem.

LM3150 schem.pdf

0 Samuel Jaffe over 5 years ago in reply to ACGUY

TI__Mastermind 25590 points

ACGUY,

I considered SCP since it looks like FB has a spike below 60% of VREF but it recovers too quickly to trigger soft start. Still it would be good to confirm SS does not fall during this event. Please check with the customer.

What is happening in the waveform you show? Is this the current going from the output of the LM3150 to the USB? Why do we see a large negative spike, then many positive spikes? Does the USB have a voltage when it is connected? And you said the issue did not occur on this waveform? If so, please ask again to capture a waveform with the issue.

It will be difficult to diagnose the issue if we do not have the layout. Please check if they would send it over email instead of on the public forum.

-Sam

0 ACGUY over 5 years ago in reply to Samuel Jaffe

Guru 11170 points

Hi,

This measured point is between source to GND of Low-Side-FET. SS-port voltage was not below. I think that cause is not over current. Switching operation restarted when Vfb lowering to 0.36V.

Question;
Does LM3150 have mode change (or reset) threshold at Vfb=0.36V?

Regards,
ACGUY

0 Samuel Jaffe over 5 years ago in reply to ACGUY

TI__Mastermind 25590 points

ACGUY,

I am checking with the design team to address your question. In the meantime you can change the output capacitance to see if this is triggered by VFB threshold or by a certain amount of time.

-Sam

0 Samuel Jaffe over 5 years ago in reply to Samuel Jaffe

TI__Mastermind 25590 points

Please also check the input voltage during this event as well. It would be a great help to see the input voltage and output current during this event.

Thanks,
-Sam

0 ACGUY over 5 years ago in reply to Samuel Jaffe

Guru 11170 points

Hello sam,

Thank you for your cooperation. Vin is stable. Serge voltage was not propagate from Vout to Vin.

Best regards,
ACGUY

0 Samuel Jaffe over 5 years ago in reply to ACGUY

TI__Mastermind 25590 points

ACGUY,

We see a spike on FB from your waveform. This spike may trigger the over-voltage comparator. Try removing C72 to reduce the coupling. That may fix the issue.

-Sam

0 ACGUY over 5 years ago in reply to Samuel Jaffe

Guru 11170 points

Hello sam,

Our concerned point is not it.
1) Vfb was raised up by USB inserted serge noise,
2) LM3150 stop operation,
3) Vfb below to 0.72V, LM3150 was not restart operation. <= Why???
SS port voltage was not decrea. => It is not by current limit function.
This cause is definitely OVP. However, restart point is Vfb=0.36V.
We want to know why the restart point is not Vfb = 0.72V.

Regards,
ACGUY

0 Samuel Jaffe over 5 years ago in reply to ACGUY

TI__Mastermind 25590 points

ACGUY,

I am talking with the team. We will respond in the coming days.

Thanks,
-Sam

0 Samuel Jaffe over 5 years ago in reply to Samuel Jaffe

TI__Mastermind 25590 points

ACGUY,

We are still looking into the 50us delay. Were you able to resolve the issue by removing C72?

-Sam

0 Samuel Jaffe over 5 years ago in reply to Samuel Jaffe

TI__Mastermind 25590 points

ACGUY,

Please also send a waveform of SS during this issue. We'd like to see if it drops at all.

-Sam

0 Samuel Jaffe over 5 years ago in reply to Samuel Jaffe

TI__Mastermind 25590 points

ACGUY,

Were you able to perform this test?

-Sam

0 ACGUY over 5 years ago in reply to Samuel Jaffe

Guru 11170 points

am,

I tried C72 removed. The symptoms have improved somewhat, but the device has stopped working. The stop time has been shortened a little.The voltage at the SS terminal does not change at all.

Regards,
ACGUY

0 Samuel Jaffe over 5 years ago in reply to ACGUY

TI__Mastermind 25590 points

ACGUY,

I will talk with the team and come up with some more theories. I will respond in a day or two.

-Sam

0 Samuel Jaffe over 5 years ago in reply to Samuel Jaffe

TI__Mastermind 25590 points

ACGUY,

You said removing C72 helped but the VOUT waveform you shared is almost identical to the first one you sent. What improvement were you seeing?

Please share the FB waveform now that you've removed C72. There should be less noise - I'd like to confirm.

You can also place a very small filter cap from FB to GND. Use 20pF. This should also reduce the noise.

The longer we work at this the more I think this is an issue with layout. This will be very difficult to solve without seeing the layout. Can you please try again to get the layout.

Thanks,
-Sam

0 ACGUY over 5 years ago in reply to Samuel Jaffe

Guru 11170 points

Sam,

Please answer to question for my first post.
Our question is not about alleviation technic for noise. The recovery condition from OVP does not match the content of the data sheet. Please comment on that.

Question,
When Vfb=0.36V at Vout=3V. The voltage is lower than Vfb-ov.
Why doesn't device resume operation quickly?

Regards,
ACGUY

0 Samuel Jaffe over 5 years ago in reply to ACGUY

TI__Mastermind 25590 points

ACGUY,

I understand that this is frustrating that we cannot quickly find out what is causing this issue.

I am not suggesting noise mitigation for the sake of noise mitigation. I am trying to find clues that help us find out what is happening inside the part. There are many conditions that may cause us to see this issue but still not violate the datasheet and expected operation of the part. But in order to find what condition we are in I will need to see results to some tests and see the layout.

-Sam

0 Samuel Jaffe over 5 years ago in reply to Samuel Jaffe

TI__Mastermind 25590 points

ACGUY,

I have not heard back in 11 days. This thread will lock. Please post again if you are still having issues.

Thanks,
-Sam

0 ACGUY over 5 years ago in reply to Samuel Jaffe

Guru 11170 points

Sam,

I obtained the pattern layout from customer.

Could you check it up?

Regards,

ACGUY

0 Samuel Jaffe over 5 years ago in reply to ACGUY

TI__Mastermind 25590 points

ACGUY,

This only shows me the component placement. Please also share the layout which shows the traces. I need to see how everything connects as well.

Thanks,
-Sam

0 ACGUY over 5 years ago in reply to Samuel Jaffe

Guru 11170 points

It's Reproduced on EVM.
https://e2e.ti.com/support/power-management/f/196/p/804309/2979328#2979328

Thanks,
ACGUY

Power management

Power management forum

LM3150: Switching stopped for a long time after OVP detection