LP8863-Q1: I2C fail

A Bokma

Part Number: LP8863-Q1

Dear team,

I have made several prototypes with the LP8863 without any problem. Now the first production run is in and I have a problem with the I2C bus.

Of a bigger batch (which I have to test ) the first 4 out of 5 do not function. With correct levels, timing both with and without leds, I get a NACK in return. In the end I replaced the LP8863 and the board works fine.

With a second one, also.

Testing is done with both own SW as via the I2C of the evaluation kit.

Any idea what the cause of the failure can be? Production is done in a professional assembly factory. ( LP8863 AQ1 - 02TG4 - AQYV )

Kind regards,

Anne

over 5 years ago

0 Charles-Chen over 5 years ago

TI__Genius 11836 points

Hi Anne,

It is better to provide the waveforms for analysis better with your power on sequence.

0 A Bokma over 5 years ago in reply to Charles-Chen

Prodigy 125 points

Hi Charles,

Which waveforms do you want to see?

Vin (12V) , VDD (5V), VDDio ( 3V3) and enable ( 5V) ?

First the 12V is applied, feeding a TLE4476 5V and 3V3 regulator. Then the enable is set.

Kind regards,

Anne

0 Charles-Chen over 5 years ago in reply to A Bokma

TI__Genius 11836 points

Hi Sir,

The waveforms of SCL and SDA when NACK happens. What's the time between the power on of device and the start of your failed I2C communication?

0 A Bokma over 4 years ago in reply to Charles-Chen

Prodigy 125 points

Hi Charles,

In the picture the first frames from start. As you can see the first 4 frames get an ack, but the 5th not. The 3th line is the enable, the 4th line the 3V3 power.

The time between start up and the first byte is about 40ms. All frames sent after the 5th frames also generate a NACK. Frames sending is 0x58.

One remark. Due to behavior of the uC the enable is logic high during startup. Then after 25ms logic low. But even when I disconnect the enable so it is low the same behavior occur.

In the meantime I have tested some more boards and about the half is functioning right, the other part is not.

Kind regards, Anne

0 Charles-Chen over 4 years ago in reply to A Bokma

TI__Genius 11836 points

Hi Anne,

Are you trying to read the register, right? After the read command, the 9th bit of 5th frame are supposed to be pulled down by I2C master rather than backlight LED driver when you want to read the register. Besides, there are glitches in SDA line. Are the glitches real or just caused by measurement issues?

0 A Bokma over 4 years ago in reply to Charles-Chen

Prodigy 125 points

Hi Charles,

Something went wrong with the picture. The bytes you see is from another device. So I included the right picture.

Picture 1: power on. As you can see the first byte will be send after about 100ms.

Picture 2. This is a zoom in to the first byte of picture 1. This is a write of the address so I suppose the ack/nack is from the LP8863. I have changed the clock to 50kb instead of 400kb. The pull-up resistors are 1k2 but I also test it with 10k, which makes no difference. As you can see there are no glitches in this measurement.

Kind regards,

Anne

0 Charles-Chen over 4 years ago in reply to A Bokma

TI__Genius 11836 points

Hi Anne,

What if you resend the failed message when your got NACK or toggle VDD? Please also check the VDD and VDDIO of device to ensure the amplitude has no problem.

0 A Bokma over 4 years ago in reply to Charles-Chen

Prodigy 125 points

Hi Charles,

Unfortunately no progress.

- Resending: same result ( NACK )

- Toggle VDD: NACK

- Toggle VDDIO: NACK

- Checking amplitude: stable, about 50mV noise.

Any ideas? I also checked functioning and non functioning boards agains resistance / shorts etc.

Is it an idea I send the schematics + board layout? ( email, no post on the forum )

Kind regards,

Anne

0 Charles-Chen over 4 years ago in reply to A Bokma

TI__Genius 11836 points

Hi Anne,

It's OK to send me the emails. Is it possible to be related to the soldering issues?

0 Charles-Chen over 4 years ago in reply to Charles-Chen

TI__Genius 11836 points

Hi Anne,

Please update your latest feedback.

0 A Bokma over 4 years ago in reply to Charles-Chen

Prodigy 125 points

Hi Charles,

Maybe a stupid question but how can I send you an email? I did make a request to be a friend of mine.

Kind regards,

Anne

0 Charles-Chen over 4 years ago in reply to A Bokma

TI__Genius 11836 points

Hi Anna,

I have accepted the request.

0 Charles-Chen over 4 years ago in reply to Charles-Chen

TI__Genius 11836 points

Hi Anna,

We recommend to use scope to capture the actual waveform of power supply, SCL and SDA to check whether there are glitches or not.

0 Charles-Chen over 4 years ago in reply to Charles-Chen

TI__Genius 11836 points

Hi Anna,

I am still waiting for your update.

0 A Bokma over 4 years ago in reply to Charles-Chen

Prodigy 125 points

Hi Charles,

October 1st I have sent you an email with additional information.

In between the problem seems to be solved. Due to a component assembly error and the subsequent testing afterwards seems to broke the IC.

When testing another batch different these boards are all ok.

If the other boards of the batch are also ok then the problem is solved and the IC works fine.

I'll let you know if so, in that case this question can be closed.

Thanks for the help.

Kind regards,

Anne

0 Charles-Chen over 4 years ago in reply to A Bokma

TI__Genius 11836 points

Hi Anne,

Good to hear that. I will close the threads temporarily. The thread will be opened again once you ask any questions here.

Power management

Power management forum

LP8863-Q1: I2C fail