MM145453V NOPB failures

Hi Mike,

Welcome to the E2E forums!

I'm working on tracking down the appropriate team who supports this device. I believe questions for this device belong in the high speed interface forum. I will move this thread accordingly.

Thanks!

Alex

Hi Mike,

I would like to try to understand why your MM145453V is failing. Do you have a schematic, or can you try to describe your system and what you have connected to what and what your inputs are? Any additional info may be helpful. Thanks.

 I tried inserting a PDF file of the schematic below but it didn't work. How do I attach a PDF file?

The chip is used in a UV meter that OAI manufactures.

0306-152-01H.pdf

Mike,

We don't see anything obvious as to why the device is failing. Have you tried multiple MM145453V devices and they each fail? If so, this might be a design issue. If it is just this one part that failed, it might be a bad part. What is your VDD (VCC) voltage and what is the clock frequency? We will continue to look into this problem.

Hi Paul,

1. The failure rate is about 3 % overall.
2. We started shipping the product in 2009 and we ship about 200 units a year. We have seen failures as early as 2010 but there were not many failures until the end of last year and this year. The some of the failures we have seen this year date back to units shipped in 2009 and some were shipped a month ago.
3. The board worked after installing a new device and no other parts were damaged. After the board was fixed, we placed the failed device back in the board and it still failed.
4. The room temperature was in  the 20 to 25 degree C range. The meter was turned on and the display came up with all the segments on. The last time it was used it worked fine then it was turned off.
5. The device failed as soon as it was turned on as far as I know.

Hi Paul,

Yes For the one failure OAI had in house, the meter was working one day and failed the next time it was turn on a week later. As far as we know the customers in the field are seeing the same failure mode.

Hi Michael,

 I talked to the software engineer and he said the SPI clock going to the part is 375K Hz.

Hi Mike_M,

Sorry for delay, I've been out of the office for a couple of days.

I think the time is coming to return unit(s) to TI for Failure Analysis. I suspect we'll see EOS damage but if it is located to certain pin(s) it may help focus where to look closer on the board.

How many bad units do you have on hand and are they similar failure modes? Several samples are preferred, 3-5 ideal.

Where did you get the units from? It is normal to return them via the distribution channel they were purchased from.

What is your location?

Regards, Paul

Hi Paul,

I have three parts that failed the exact same way. The parts failed with all the segments turned on.

OAI bought the parts from Mouser Electronics.

I am located in San Jose, California.

Do I contact Mouser or you to see about returning the parts for failure analysis?

Thanks Mike. I'm just double checking with QA on best return flow.   Paul

Hi Mike, the best procedure is to return all 3 units via Mouser. That way we can keep a clean document trail. There will be a form to complete but I think you already have the answers.

I expect to see clear damage on some pin(s) which will hopefully allow us to target a closer look at your board and test (power-up/power-down) procedures.

Keep me informed per your submission. Since we are getting into specifics I think it best to use email in future.

regards,    Paul    paul.dolan@ti.com

Hi Mike,

I've looked at the report which shows EOS damage on pin 25 (data). I've asked the Applications Engineer to review your schematic again & focus on p.25.

I've looked and I can't see anything which looks like an obvious risk.

Please provide more details on the 6V protection diode since I don't see anything helpful when I google the number "4300-105-01". Are any of these being damaged?

"

Hi Paul,

4300-105-01 is OAI's internal part number for Littelfuse SA6.0CA TVS diode. None of the TVS diodes have been damaged.

Hi Mike,

I am the app engineer who used to support this device. I looked into the report and couldn't find anything obvious as well. My concerns were the integrity of the data line that connects to the DATA pin which is the trouble pin 25. Also, why the need for the protection diodes? Are high voltage spikes expected on the data line and if so, why? And I would like to get more information about what is connecting to the connector on the other side of the protection diodes. Lastly, can customer scope probe the data line to see what voltage spike/signal is causing the damage? I look forward to your response. Thanks.

-Michael

Hi Mike,

The protection diode was added to the board after this problem showed up. The meter is used to measure the intensity of a UV light source. These light source use a high voltage spark to start the UV lamps, so I added the diode in case the meters were near the high voltage wires when the light source was started. The diode didn't help because we were getting a 2 to 3 percent failure rate before adding the diode and after adding the diode. Also after the diode was added we had a meter fail while it was sitting on the self not near any light sources.

On the other side of the connector is a SPI EEPROM and one or two photo diodes.

Monitoring the data lines isn't practical since there are many customers and most of the time there is no problems.

Mike Manley

Hi Paul and Michael,

 Would the damage on pin 25 more likely be caused by a voltage on pin 25 greater than VDD or a voltage less than ground?

Mike Manley

Hi Paul and Michael,

 Would the damage on pin 25 more likely be caused by a voltage on pin 25 greater than VDD or a voltage less than ground?

Mike Manley

Hi Mike,

Sorry for the late reply, but I don't know the answer. Doing some experiments may help. There was an internal reorg in Jan and the device got moved to another group. I'be be happy to continue support as far as possible but we may have come to a crossroads and need new input.

Mike DeSando - do you have any more input or can you help identify some applications support in the new group?

regards,   Paul

Hi Paul,

 When this meter doesn't see a signal larger than a threshold value it goes into a low power mode. It goes into the low power mode by turning off the positive voltage supply to all the chips except the micro controller and it puts the micro controller in  a low power mode with the controllers outputs turned off. If the positive supply voltage for the MM145453 was turned off and pin 25 got a normal voltage signal from the micro controller, would this cause the damage on pin 25?

Hi Mike,

Interesting scenario you've discovered. Do you know what voltage will be applied to the data pin 25 by the uC in this situation?

regards,   Paul