LP3982: LDO giving 3v3 instead of 3v but no loading effect or heating

Hi Swapniil,

Swapniil Kanwar said:

1. Why is this happening if datasheet says 2% error in output voltage? I have purchased this IC from ti.com -- official TI website

This particular device has a level 3 MSL rating, which implies that the device has some susceptibility to humidity. Do you know the rough humidity and temperature while in storage? Have you tried baking a board to test if the output recovers to 3V?

Swapniil Kanwar said:

2. Now if it is giving 3.3V and rest of the circuit is working fine (since circuit has 3.6V max limit) can I proceed forward for integrating this PCB in my product? I am dicey about the longevity of this IC, will it survive at least 1yr provided there is no direct contact with human on circuit in final device? 

If there is not some sort of coating while this device is in storage or not operational, moisture can get into the package and can cause issues if the board is not baked before using. The part likely isn't damaged, so I wouldn't say that it won't survive, but as you are seeing, moisture can affect how the device operates. 

Of the boards that show 3.3V on the output, are they all exactly 3.3V? In general if humidity were to get into a package and change the electrical behavior of a device, I wouldn't necessarily expect that they would all show the same incorrect characteristics or that they would regulate to a common voltage (e.g. 3.3V) vs. some random voltage. Had the boards that show 3.3V been verified before going into storage for the correct operation? Is it possible that they are actually 3.3V fixed out versions?

Regards,

Nick

Hi Nick,

1. The concerned place is near Delhi where relative  humidity varies from 42% in the summer to 82% in the monsoon. So what you said about level 3 MSL makes sense.
Can you please tell the details of baking the board. Since LiPo battery is attached to this PCB is it advised to do it? I was thinking to keep all the PCBs in an oven at 35 degree C for let's say 2-3Hours, will that work? 

2. There was no coating done on the device, just packed in a ESD safe plastic pouches, so moisture could've creep in. As far as the output voltage is concerned, only 3/25 showed 3V rest 22 devices are giving 3.28-3.31V without getting heating or any voltage drop due to loading effect (load current is ~300mA). No voltage fluctuations seen in the PCB. I have crosschecked the P/N of the LDO is LP3982ILD-3.0/NOPB which is a 3V LDO. As per your last paragraph comment, would you still think that moisture can cause a 3.3V shift instead of 3V? 

Hi Swapniil,

The timing/temperature for baking can be found in IPC/JEDEC J-STD-033C. The baking recommendation will depend on the time that the device was exposed to humidity and the RH. I can't speak to the baking parameters for the LiPo battery; you may need to reach out to the supplier for that question.

Swapniil Kanwar said:

As per your last paragraph comment, would you still think that moisture can cause a 3.3V shift instead of 3V? 

I normally would not expect that such a large percentage of devices would shift to the same voltage, but I suppose it is possible. If you bake the boards and see that the devices recover to 3V then that would suggest that they are in fact just being affected the same way by humidity, but it's difficult to say without doing the baking test. 

Regards,

Nick

Hi Nick, 

I tried baking PCBs but still the same issue. Baking was done at 100degC for 4 hours. I am still getting 3.3V instead of 3V. 

It was also observed that 12 out of 50 PCBs are experiencing current fluctuation when load current wants to draw ~200mA as controller goes in high power state. Rest other 38PCBs are working fine at all possible load currents. 

So with this experiment, our baking test got failed. 

Do you know any other solution to my problem? 

Hi Swapniil,

That is quite odd. Can you tell me what the top marking is for devices that have a 3V output and for those that have a 3.3V output? Are they different? The Datasheet says that the 3V devices should have "LTB". 

Swapniil Kanwar said:

It was also observed that 12 out of 50 PCBs are experiencing current fluctuation when load current wants to draw ~200mA as controller goes in high power state. Rest other 38PCBs are working fine at all possible load currents. 

Do you have any waveforms to show me what you mean by current fluctuation? Can you also share the schematic for this device?

Regards,

Nick

Hi Nick, 

During research and development phase it was measured that when device is switched on and is in Bluetooth advertising phase, the circuit consumes 60-65mA. But when device is in pairing process with a host, it consumes ~150-200mA (fluctuating). So as per my above query, 12 out of 50 pcb are showing a brownout reset of the controller when bluetooth pairing is performed, but rest other circuit are working fine even at 3.3V. 

I am unable to understand this. Since this is causing too much confusion and delay in manufacturing our devices, I am thinking of replacing IC to a new one (which gave me 2.99-3V output). 

As an electronics engineer I am still curious to know what can be the cause, why baking didn't affect the output, what can be done to avoid it and lastly if I continue using this how long should I give estimation of the circuit life to customers. 

I know this query is going on for long but this curiosity will help me making better circuits and strategy for future use. 

Thanks

Hi Swapniil,

Have you saved any waveforms of the fluctuating output you described? Also, do you expect the load to fluctuate when the Bluetooth device is in the advertising phase? With the information you've posted here thus far I do not have an explanation for the behavior you are seeing, but with that said, without seeing waveforms showing the behavior it's not easy to see a clear picture of what is happening. 

Did you end up looking at the top marking of the devices to make sure you have the correct and expected devices? I would agree that I'm curious as well why you are seeing 3V outputs on what are supposed to be 3.3V since this device is available in a 3V output and you have described that all of the failed devices are showing a pretty precise 3V output. If humidity was the cause of this I would expect that you would see more variance than you have described, and I would also expect baking to fix the devices. 

Would you like any help finding a different device? I would normally steer away from LPxxxx devices because these are older devices made on old technology processes, and unless there is a good reason for using an LP device (e.g. if there is some specific requirement that we don't have a better device that can meet it) I would suggest using a newer device. 

Regards,

Nick