• State TI Thinks Resolved
  • Locked Locked
  • Replies 5 replies
  • Answers 2 answers
  • Subscribers 31 subscribers
  • Views 1125 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

HDC1080: accuracy issues

Helmut Kraus
Prodigy 25 points
Part Number: HDC1080

Hello,
we are using the HDC1080 and see a worse accuracy as given in the datasheet.

 The measured temperature differs by about 1°C and the humidity by about 10% between the sensors (at nominal 20°C / 40%).

We have 7 sensors located in close proximity (with one SCD41 from competitor) in a climate camber with ventialtion. They are all soldered on flexible PCBs with about 30mm distance to the mainboard for reading. So due to the fantilation and distance to heat sources we should not have any heating transfer to the sensors caused by some other devices. We already connected all sensors FPC's to the next mainboard and see that the error is moving with the sensor. So there's no relation to the mainboard or position in the chamber

What else could have an impact on the sensors which could result in such bad accuracy?

Thenks,

Helmut

  • 0
    TI__Genius 12395 points

    Hi Helmut,

    Do you have any sort of calibrated reference to compare the HDC1080s and the SCD41 against? Are you seeing all the HDC1080's having a large RH deviation from the SCD41, or just a single (or a few) HDC1080s with this problem? I would recommend using some sort of calibrated reference when doing any RH validation for any humidity sensor. If you are comparing against other sensors, we dont know which one has the real error.

    An RH error that large would come either from high temp & RH exposure for an extended period of time, or more likely some sort of contamination in the assembly process or in how its used. For the devices that show the large RH error, was there any difference in how they were assembled?

    Thanks

    -Alex Thompson

  • 0 in reply to Alex Thompson
    Prodigy 25 points

    Hi Alex,
    the datasheet states +/- 0.2°C temperature accuracy and +/- 2% RH accuracy.

    We measure values between 20.6 up to 21.8°C. The SCD41 20° (the climate chamber is set to 20)
    For the humidity we measure between 38% and 49%, the SCD41 shows 37. The chamber is set to 40.

    As I maybe understand the problem for the humidity to be suszeptile for contaminations I do not understand that for the temperature.

    All sensors have been assembled the same way on a professional SMT line.

    Thanks,
    Helmut

  • 0 in reply to Helmut Kraus
    TI__Genius 12395 points

    Helmut,

    Is the chamber you are using calibrated? Of thee 7 HDC1080s, how many have RH error greater than +3%? For the devices with the large RH error, a quick bake at 100C for about 2 hours should remove any excess moisture in the HDC1080 polymer and improve their RH performance. 

    Regards

    -Alex Thompson

  • 0 in reply to Alex Thompson
    Prodigy 25 points

    Alex,
    it will not make a difference if the chamber is calibrated, oder not. We are comparing the values from several HDC1080 with each other and the SCD41.







    both - temperature and humidity accuracy is worse as expected.
    Regards
    Helmut

  • 0 in reply to Helmut Kraus
    TI__Genius 12395 points

    Looks like most of the devices, including the SCD41, are reading too low in RH at the 45% setpoint. I know you have said its not important, but I really do recommend evaluating any humidity sensor against a calibrated reference, not another humidity sensor. Because I cannot tell which devices are reading incorrectly. As I mentioned, you should take the HDC1080s that are reading too high and do a 100C bake for 2 hours, and they should read like the rest of the devices. Humidity sensors have their accuracy drift over time or due to environmental exposure so baking can help bring affected parts back to an expected RH reading.

    -Alex Thompson