HTU21D shows 128.86 degrees

[akhabel]

Hi,
My question is if someone have experienced this issue with HTU21D I2C sensor that the measured humidity looks real while the temperature is 128.86 for 14 bit resolution and 128.90 for any other resolution no matter if I set slow I2C or not.
I'm using git version as I needed a custom build on ESP-01.
Thanks in advance.

BR,
AA

[Ath]

Does the sensor show up correctly in the I2C scan?

[akhabel]

yes, it does, at 0x40. And also tested the sensor with arduino htu21d example sketch and works well, both RH and temp.

[chromo23]

It´s probably defective. One of my five sensors had the same issue.

Edit:

And also tested the sensor with arduino htu21d example sketch and works well, both RH and temp.

Ubbs, missed that part.. sorry.
It´s strange....

[chromo23]

Just tested it with a recent build and you are right... it indeed shows 128,86
So i guess there is an issue with the plugin...

[chromo23]

The latest release (20220809) doesn’t seem to have this issue

Edit: I filed an issue on github...

[TD-er]

Do you have any log output?
Each log item regarding this plugin, will probably start with "SI70xx"

[akhabel]

Hi chromo23,
thanks for confirming, I came to the same conclusion by loading latest release. I will try to rollback to that git version and see

[TD-er]

It was changed in this PR: https://github.com/letscontrolit/ESPEasy/pull/4161

[TD-er]

I'm browsing through the changes and noticed the I2C address wasn't a configurable parameter before.

Can you check it is set to the previous default of 0x40?

[chromo23]

Can you check it is set to the previous default of 0x40?

Here it is set to default...
Log is inconspicuous:

Bildschirmfoto 2022-10-24 um 13.35.21.png (24.5 KiB) Viewed 7006 times

[TD-er]

But did you actually save the settings with the address set?

[chromo23]

I´m not sure what you mean. But i did set it to 0x40 and hit the submit button (changed it to 0x41 before just to check)...

[akhabel]

Hi
checked the same
in log see like: P014: Temperature: 128.86 Humidity: 46.60
i2c address is 0x40 (64) - (default)
just for curiosity switched to 0x41 but that case gives "nan" for both values and in log: SI70xx: Not available at address: 41

[TD-er]

OK, so it seems it is more of an interpretation issue of the temperature data.
The humidity does seem to be plausible and at least differing among both your setups.

[akhabel]

Hi,
basically yes. rolled back git revision and now works fine.
thanks for helping me out with this question you are doing a great job!

[TD-er]

Thanks for reporting the issue.
This needs to be fixed for sure before making a new build.

[fmuntean]

Hi,
basically yes. rolled back git revision and now works fine.
thanks for helping me out with this question you are doing a great job!

The sensor has a heater inside it. Is the sensor really at the 128 Celsius ? Can you check if the sensor is hot ?
If it is not hot then I recommend to remove the plugin then add the plugin back which should reset the internal variables as now I2C address can be selected and is stored internally with the resolution.
A screen shot of the plugin settings and some logs would help here.

[akhabel]

the chip is not hot.
if you provide what to check and how to check then I can do it and share
by the way I'm using gpio1/3 for i2c communication so I can't capture terminal logs.

[TD-er]

You can also see the logs via the web interface (tools -> log)
Just make sure to set the web log level on the tools->Advanced page.

[akhabel]

Hi there,
With info level the printout is the same as I sent earlier and chromo23 sent the screenshot as well.
With error level I don't get anything

[fmuntean]

open the log in the browser then on a separate tab go and configure the plugin.
Then from the log tab copy the entire logs.
I need to see the logs that are sent during initialization of the plugin.

[akhabel]

646798: EVENT: TaskExit#htu=2,0
646819: EVENT: TaskInit#htu=2,0
646839: SI7013 : sn=405ba4-32114854
646839: P014: chip_id=50
646973: P014: Temperature: 128.86 Humidity: 59.40
646976: EVENT: htu#temp=128.86
646977: EVENT: htu#humid=59.40

[fmuntean]

646798: EVENT: TaskExit#htu=2,0
646819: EVENT: TaskInit#htu=2,0
646839: SI7013 : sn=405ba4-32114854
646839: P014: chip_id=50
646973: P014: Temperature: 128.86 Humidity: 59.40
646976: EVENT: htu#temp=128.86
646977: EVENT: htu#humid=59.40

Thank you this is what I need.
This sensor seems to be a "clone" of the SI70xx and the datasheet is very vague on the implementation.
However I see that a change I made is to read the temperature from the humidity measurement.

This is a detailed info on the approach:
- in order to get better humidity response the temperature is required so when measuring humidity the temperature is also measured
- then for Si70xx to keep the sensor from heating up we do not need to start a new measurement for temperature but just need to read the previous value.
- the HTU21D does not seems to have that command implemented returning all 1s thus giving you the max temp always.

I can use the chip_id = 50 to switch the state machine into doing an extra temperature read.

The serial number command is not documented at all for HTU21D either so if you have multiple sensors please check the chip_id for all to see if it is the same.
If not then please provide the serial numbers to see if we can figure out the location of chip_id.

[TD-er]

[...]
- the HTU21D does not seems to have that command implemented returning all 1s thus giving you the max temp always.

I can use the chip_id = 50 to switch the state machine into doing an extra temperature read.
[...]

You could also check for the return value using that command.
This can be done in the init and if it returns only 1's then set a flag to not use this command.

[fmuntean]

I would prefer to use the chip_id if possible.
as the all 1s is also not documented. but that value of 128.86 is when the raw value is the max (all bits are 1)
I am preparing the pull request as we speak with the change.

[TD-er]

Great!
Let's hope people can then test it for a few days, so we will be making a new build on Sunday or Monday as this is (as far as I know) the last blocking one for the new build.

[akhabel]

Hi there,
I did the same test with all sensors I've got:

83465: EVENT: TaskExit#htu=2,0
83485: EVENT: TaskInit#htu=2,0
83503: SI7013 : sn=405dc7-32114854
83503: P014: chip_id=50
83638: P014: Temperature: 128.86 Humidity: 57.60

59643: EVENT: TaskInit#htu=2,0
59658: SI7013 : sn=6464c2-32114854
59659: P014: chip_id=50
59793: P014: Temperature: 128.86 Humidity: 46.70

46880: EVENT: TaskExit#htu=2,0
46903: EVENT: TaskInit#htu=2,0
46988: SI7013 : sn=6466c7-32114854
46988: P014: chip_id=50
47130: P014: Temperature: 128.86 Humidity: 56.00

[fmuntean]

excellent. I started the pull request https://github.com/letscontrolit/ESPEasy/pull/4321 with the changes.

[TD-er]

When the GitHub Actions build is ready, could someone try it?
https://github.com/letscontrolit/ESPEas ... 3345553088

[akhabel]

Hi,

built again from this commit:
commit cff4e44522261f9b312732b01225fd52faefdd1b (HEAD -> mega, origin/mega, origin/HEAD)
Merge: 22cc1fbbc ce04d1009
Author: TD-er <gijs.noorlander@gmail.com>
Date: Wed Nov 2 22:39:46 2022 +0100

Merge pull request #4321 from fmuntean/P014-SI70xx

[P014 si70xx] fixing support for htu21D

here is the log output:
306905: EVENT: TaskInit#htu=2,0
306924: SI7013 : sn=6466c7-32114854
306924: P014: chip_id=50
307165: P014: Temperature: 21.80 Humidity: 60.10

[TD-er]

Cool!
Thanks for the feedback.