HDC2010: Have offset of ~ +20% on some samples.

HI Martin,

How many sensors are reading +20% high, or approximately what percentage are reading high? Are all the sensors reading high from the same lot, or are there members of the lot that have no issue? Have you exposed these offset sensors to any chemicals or very high RH & temp conditions? have these parts been exposed to pink antistatic foam? Is the +20% offset constant across different ambient humidity levels, or does the offset vary as the humidity varies (above 80%, the sensor wont output more than 100% RH of course so it wouldn't present as exactly +20%RH in that range)? Have these sensors been scratched or damaged?

Thanks

-Alex Thompson

Hi Alex,

Thank you for the support!

I think we can say 100% of the sensors with +20% are from one lot and 100% without that offset are from another lot.

The sensors weren't exposed to chemicals or very hogh RH & temp conditions.

I need to check whether these sensors were exposed to pink antistatic foam. Can you meanwhile give me information about the influence of that foam on the sensor?

The offset is constant across different ambient humidity levels.

The sensors weren't scratched or damaged.

Any help is appreciated.

Best regards,

Martin Peevski

Martin,

Are there any HDC2010s from the lot that is have +20% errors that are good (no large error)? Or is it a 100% failure rate within that lot?

Given that this is isolated to a specific lot, could you please share the lot information (the label on the box)? I also think it would be wise to start the customer return process so we can get these parts on our bench and test them. The link to start this process is here: https://www.ti.com/support-quality/additional-information/customer-returns.html

Pink foam has been shown to cause RH errors because it off-gasses a chemical which interacts with the sensing polymer. These RH errors often show as positive RH errors, but +20% is very high, and I am skeptical that this is the cause. If you can determine if the problem lot was exposed to pink foam or any chemical that the working lot was not, that would be very helpful.

Thanks

-Alex Thompson

Hi Alex,

It looks like 100% failure for that lot.

You asked about the label on the box. I don't understand which box did you meant. Did you mean the HDC2010 box?

About the customer return process I think it will be hard to do it because we use manufacturer in China and I don't know where from he purchased the HDC2010's. But we can try sending to him 3 our devices if we decide doing that.

Regarding the pink foam - it's possible the manufacturer in China to put the assembled PCB's in such foam.

We'll try to figure out whether he exposed the problem lot to pink antistatic foam.

We put the ready devices in that cardboard box:

When I have more information will write to you again.

Meanwhile feel free to share with me any information you have.

Thank you and best regards,

Martin Peevski

Hello Alex,

That photo shows manufacture process of our device in the China factory.

But they use that process for all devices, and for the devices which are without humidity offset.

Do you think this can be the reason?

Do you think they should stop use that material? If yes, please give as advice what they should use instead.

Best regards,

Martin Peevski

Martin,

Thanks for all the additional information. I do not think pink foam exposure is the issue here. We do see positive RH errors induced by pink foam, but not +20%, and not this consistently. Furthermore, It would not cleanly isolate to one lot since all the others have gone through this process. The factory should not use pink foam for any of our humidity sensors, but I highly doubt that is the issue here.

By "label on the box", I meant the lot number on the HDC2010 box. We should be able to retrieve this ourselves once we can get our hands on some of these failing devices. I apologize, I should have been more clear.

You have some of these failing assemblies, correct? Would it be possible to have the manufacturer in China send you some parts from the failing lot, and then you can start the customer return process?

Regards

-Alex Thompson

Alex,

We have failing devices at our office, so may be we can start customer return process.

Can you advice what the manufacturer can use instead of pink antistatic foam?

Regards,

Martin Peevski

Martin,

The manufacturer should use black foam, not pink foam or LDPE foam/bags. Our understanding is that the additive that makes the foam pink is what outgasses and impacts the sensor. Here is an example of a box with black foam:

https://www.digikey.com/en/products/detail/conductive-containers-inc/PP2020/9922362

Please start the return process as soon as possible. Myself or another team member can then start testing these parts to see if we can verify the +20% errors, and hopefully identify root cause.

Regards

-Alex Thompson

Alex,

Thank you, we will instruct the manufacturer to not use pink foam or LDPE foam/bags anymore.

Regarding the Customer Return Process - what shall I do now - send 3 PCB's to the distributor whom our manufacturer purchased the HDC2010's?

Best regards,

Martin Peevski

Martin,

I think you should send the PCBs to TI, since I want to evaluate these and see if there is a lot issue. Normally you would have to remove the ICs from the PCB for a customer return, but that is much trickier with BGA parts so do your best.

Please provide the lot numbers of the bad and good lots if you can as well. If you cant, we should be able to determine that once we have the devices.

Could you please share a photo of the PCB? If the HDC2010 is sent still mounted to PCB, I want to ensure there are test points so that I can access the signals of the part without needing to interface with any MCU or custom communication software.

Regards

-Alex Thompson

Alex,

For me isn't problem to desolder the HDC2010's before send it to you. We have hot air and I'll be careful.

Can desolder 2 pieces and send one on PCB. I think you should not have a problem with our MCU if it's with erased memory. I hope you to have access to the I2C bus then.

Please, send me the information about the destination address.

Regards,

Martin Peevski

Martin,

Glad to hear you can remove the HDC2010s. here is what you should send:

• HDC2010s from failing lot, still in the reel unopened
  • you'll likely need to get these from your manufacturer, but this would be ideal for us to test since as I understand this lot has a 100% fail rate
• HDC2010 still mounted on your PCB. I have access to I2C bus, but will need to bypass your MCU. Please highlight a way to do so.
• HDC2010s you have seen fail, desoldered from your PCB
• controls

You will need to still do the proper TI customer return process, we want to avoid customers directly shipping to TIers. While that way is technically faster, there is no internal system tracking the return and failing devices can get lost or forgotten about. Please use this link to get started:

 https://www.ti.com/support-quality/additional-information/customer-returns.html

Thanks

-Alex Thompson

Hi Alex,

- I'm unable to send you unopened HDC2010s from that reel

- I'll send you two desoldered from their PCB's

- I'll send you two PCB's with desoldered MCU in order you to have access to the I2C bus.

- What do you meant by "controls"?

Best regards,

Martin Peevski

Martin,

Thank you for sending what you can. By "controls" I just mean working devices that you have, to compare against in case we need to perform FA on these parts. If that is not possible/practical, I have some HDC2010s I sampled from the TI store that could be used as a control device.

Thanks

-Alex Thompson