LM3880: Some LM3880 works abnormally

Hi

Could you please check the EN threshold during the VCC ramping? EN shall not be over ENABLE threshold before VCC reaches 2.7V otherwise the behavior is unpredicted.

Regards

Yihe

Hi Yihe,

Thanks for your reply.

The startup waveform of the EN pin on lm3880 is as follows.

It takes about 20ms for EN to rise to ENABLE threshold(1.25V). And it takes about 1ms for VCC to rise to 2.7V.  There seems to be no problem here.

The startup waveform of the EN and VCC is as follows. The yellow waveform is EN. The blue waveform is VCC.

Best Regards,

Lin

Hello

What's the temperature in our case? What are the differences between 5 board or less 5boards in your system?

Is the problem only on that particular board when there are 5 boards in the chassis ?

Do you check the other side of the FLGA1, is it  possible to have  strong pull down somehow.

I do not think the drop would cause problem since it is above 2.7V.

Regards

Yihe

Hi Yihe,

The temperature in our case is about 20℃.

We can only find a difference is the startup waveform of the power supply has little difference between 5 board or less 5boards in the chassis.

We have tested about 80 boards. Only one board has the problem.

The other side of the FLGA1 is the EN pin of TPS548A20RVER. There is no other load connect  to FLAG1.

When there are five boards in the chassis, sometimes the abnormal board can be powered up. One time out of 15 times  the abnormal board can be powered up.  If we use power supply made by a other manufacturers to supply 5V, the abnormal board can be powered up always. But we cannot change the power supply in the chassis, and we want to find the reason of the problem.

We tested the power-on startup waveforms of 5V, EN of lm3880, FLAG1, FLAG2, FLAG3 when the abnormal board cannot be powered up. As follows.

(picture above.  cannot power up. yellow waveform is  EN of lm3880. blue waveform is FLAG1.)

(picture above.  cannot power up. yellow waveform is  EN of lm3880. blue waveform is FLAG2.)

(picture above.  cannot power up. yellow waveform is FLAG3 . blue waveform is EN of lm3880.)

(picture above. cannot power up. yellow waveform is EN of lm3880. blue waveform is VCC_5V.)

(picture above.  cannot power up. yellow waveform is  VCC_5V. blue waveform is FLAG2.)

We tested the power-on startup waveforms of 5V, EN of lm3880, FLAG1, FLAG2, FLAG3 when the abnormal board can be powered up. As follows.

(picture above. can power up. yellow waveform is FLAG1. blue waveform is EN of lm3880.)

(picture above. can power up. yellow waveform is  EN of lm3880. blue waveform is FLAG2.)

(picture above. can power up. yellow waveform is  FLAG3. blue waveform is EN of lm3880.)

(picture above. can power up. yellow waveform is  EN of lm3880. blue waveform is VCC_5V.)

(picture above. can power up. yellow waveform is  VCC_5V. blue waveform is FLAG2.)

Best Regards,

Lin

Hi

When there are 5 boards, does other 4 boards have the same voltage drop on the VCC as the suspected board or not?

When used power supply from other vendor, the suspected board is always working even there are 5 boards in the chassis. Is this right?

what are the differences between these two supplies, voltage drop on the VCC?

Do you know the cause of the VCC drop and can it be fixed?

Have you tried to swap the LM3880 for cross checking?

Regards

Yihe

Hi Yihe,

When there are 5 boards, other 4 boards have the same voltage drop on the VCC as the suspected board. All boards share a power supply to supply VCC_5V.

When used power supply from other vendor, the suspected board is always working when there are 5 boards in the chassis. Power supply from other vendor has longer ramp time and has no voltage drop. Startup waveform of the power supply from other vendor is as follows.

 The power supply vendor said voltage drop is inherent property of power supply, it cannot be fixed unless totally change power supply design.

We are planning to swap the LM3880 for cross checking next week. It cannot be done in recent days for some reason.

Best Regards,

Lin

Hi

Thank you for the update. Basically the slew rate is different between different power supply. please report back after the cross checking is done.

We haven't heard any similar issue.

Regards

Yihe

Hi Yihe,

We swapped the LM3880 on the normal board to the suspected board. The suspected board becomes can be powered up always. And the normal board becomes cannot be powered up.

What can I do to find the reason of the problem?

Best Regards,

Lin

Hi Lin

Thank you for the update. we will check and get back to you.

Since it is holiday at US, the response is expected after New Year.

Thank you for your patience.

Regards

Yihe

Hi

Have you tried to increase or decrease the slew rate of VCC? The chip is not expected that kind of drop from design point view.

Regards

Yihe