Feature request: Add support for Sensirion SEN5x

[erdinguma]

Describe the problem you have/What new integration you would like

I'd like you to add support for the newest line of PM sensors from Sensirion - the SEN5x series.

Please describe your use case for this integration and alternatives you've tried:

Sensirion has released their newest line of air quality sensors, the SEN5x series. These sensors combine the best features of many of their other sensors, including the SPS30, SHT30, and more.

These sensors are also cheaper than Sensirion's SPS30 PM sensor (~$34 for the SEN54 vs. $55 for the SPS30 on DigiKey). I suspect most users will ultimately gravitate to the the SEN5x series as it offers more bang for the buck.

Adding this sensor would be a great addition to Tasmota!

Additional context

Relevant Links:
SEN55 - (PM, RH/T, VOC, NOx) - https://sensirion.com/products/catalog/SEN55/
SEN54 - (PM, RH/T, VOC) - https://sensirion.com/products/catalog/SEN54/
SEN50 - (PM) - https://sensirion.com/products/catalog/SEN50/

[sfromis]

It would be welcomed if "someone" submits a PR to include support for the SEN5x series.

[DouglasRudd]

I am still new with tasmota. so It could be a config issue.
sensor is the sen 55 no other sensors on this mcu
mcu is a esp8266ex.
00:00:00.002 HDW: ESP8266EX
00:00:00.050 CFG: Loaded from flash at F4, Count 51
00:00:00.056 QPC: Count 1
00:00:00.070 ROT: Mode 1
00:00:00.071 SRC: Restart
00:00:00.073 Project tasmota - Lab _Sen55 Version 12.3.1.6(tasmota)-2_7_4_9(2023-02-08T22:18:25)
00:00:01.368 I2C: SEN5X found at 0x69
19:01:07.939 Running readMeasuredValues for SEN5X...
19:01:07.964 SEN5x readings:-
19:01:07.966 MassConcentrationPm1p0:
19:01:07.968 MassConcentrationPm2p5:
19:01:07.970 MassConcentrationPm4p0:
19:01:07.971 MassConcentrationPm10p0:
19:01:07.973 AmbientHumidity:
19:01:07.975 AmbientTemperature:
19:01:07.977 VocIndex:
19:01:07.979 NoxIndex: n/a
19:01:07.981 AbsoluteHumidity:

[tyeth]

Aha, thanks for the extra info, fascinating! I'm also relatively new to it, and haven't tried an esp8266 yet. I wonder if I need to do something a little different, but in theory the Wire.begin should still work, which means the sensor connected on the first i2c bus. There is only support for Wire1 (second i2c bus) if ESP32 is defined. Only other thought is data types, which are floats and int16's as far as I'm aware which I naively assume is fine on esp8266. Possibly the backend sensirion library maybe doesn't support esp8266. I'll consider tasmoterising one of my smart-plugs to test the esp8266 over the weekend. Also your console isn't showing the extra debug info, I think you have to do a build with the extra #define DEBUG_TASMOTA_SENSOR which may then show a specific failure in the console.

[DouglasRudd]

I can try with an esp32, I checked the debug, and no errors.. I also flashed it with the sample Arduino code just the verify the sensor is working correctly

[tyeth]

You should see an additional message if the DEBUG_TASMOTA_SENSOR is defined, it will mention i2c bus number 0, either as a failure or success (Line 109 or 114 depending on error/success), may require setting logging to debug or more-debug in the console.
https://github.com/arendst/Tasmota/blob/development/tasmota/tasmota_xsns_sensor/xsns_103_sen5x.ino#L109-L114

[DouglasRudd]

I do have the debug set and the output you can see in my previous post, the sensor is detected on the correct address.

00:00:00.073 Project tasmota - Lab _Sen55 Version 12.3.1.6(tasmota)-2_7_4_9(2023-02-08T22:18:25)
00:00:01.368 I2C: SEN5X found at 0x69
19:01:07.939 Running readMeasuredValues for SEN5X...

both option 3 and for (debug and debug more) to not show any errors.

[tyeth]

Sorry for the lack of explanation on my part, the line you highlight is tasmota executing line 113, not debug sensor output, and tasmota doesn't log the bus number to console (although it is supplied to the I2cSetActive function), and the lines I mentioned (109 or 114) do mention a bus, so there should be a second similar message ending with "bus 0" logged to console.

…

On Sat, 11 Feb 2023 at 21:20, DouglasRudd ***@***.***> wrote: I do have the debug set and the output you can see in my previous post, the sensor is detected on the correct address. 00:00:00.073 Project tasmota - Lab _Sen55 Version 12.3.1.6(tasmota)-2_7_4_9(2023-02-08T22:18:25) *00:00:01.368 I2C: SEN5X found at 0x69* 19:01:07.939 Running readMeasuredValues for SEN5X... both option 3 and for (debug and debug more) to not show any errors. — Reply to this email directly, view it on GitHub <#15463 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABTBZ477L4OSL4QADKUCEG3WW77CPANCNFSM5UKG5ITQ> . You are receiving this because you commented.Message ID: ***@***.***>

[DouglasRudd]

I think we are looking in the wrong direction, I think it has something to do with the float conversion to string/char in the PTRS().
I have found that it will not output with a %f, I also found in the docs reference to %_f (https://tasmota.github.io/docs/Sensor-API), and looking at some of the other code, I found the use of dtosrtfd() to convert floats to char and then using %s for output. I am testing this theory now.
what compliler are you using? I am using gitpod.

[DouglasRudd]

FYI here is the output from debug RE the Bus info
00:00:00.073 Project tasmota - Sen55 Version 12.3.1.6(tasmota)-2_7_4_9(2023-02-12T01:36:50)
00:00:01.368 I2C: SEN5X found at 0x69
00:00:01.369 ************ Sensirion SEN5X found, i2c bus 0

[arendst]

Tasmota does not support %f as it's library is way too large. We have our own implementation by using %*_f where * may be an option or a fixed number of decimals. The variable also needs to be given as a pointer (superceded with &).

@tyeth some changes recommended:

• I suggest you take a look at any other driver searching for %*_f to fix.
• While going over the driver also revisit the use of hardcoded unit names. They should also be replaced by standard language specific unit names.
• Replace the fixed celsius unit by the temperature conversion unit making it available for fahrenheit too. Again search the code for examples of how to do this.

[tyeth]

You'll be relieved to hear it's over. Pull request #17961
Testing and feedback gratefully appreciated @DouglasRudd @raptordemon , but locally tested using tasmota32s3cdc.

[raptordemon]

not sure if it has been mentioned but esphome supports this sensor, so there might be something that can be reviewed there: https://esphome.io/components/sensor/sen5x.html i have one on my desk

[tyeth]

Sorry it's taking me longer than desired to do the additional changes. I've done the units and float changes as requested and I've got it displaying in console, but my c/c++ sucks, I'm getting LoadProhibited reboots after about 4reads, whereas before the %*_f changes it was working nicely for me. I'll muddle through, but my local environment is a mess and I'm not getting line numbers anymore, just corrupted stack message instead.

[tyeth]

Ah thank goodness. I deleted my users\tyeth\platformio folder (I'd pip installed previously) and have a clean enviroment. I still can't set breakpoints that work inside the driver, but it breaks on startup, and more importantly the exception has been caught, I can now progress easily having spotted my mistake.

[tyeth]

It's merged into development Thanks everyone for your contributions, many eyes appreciated 🦉

[bimac]

Thanks a lot for the effort! I just set two of these up on M5stamps - works beautifully.

[tyeth]

Very glad to hear it worked! Funnily enough its the new Ikea Air quality sensor dropping in April, SEN54, supporting matter with their hub via future update. Nice to know we've bagged a good'un sensor-wise

[bimac]

Uuh, thanks for the hint about the Ikea thingy - it looks quite promising! I'll definitely look into this. I set up two SEN55 for the office, but a zigbee version of the SEN54 would be great for home use.

[raptordemon]

Glad to hear you were able to make it work! Nicely done!

[bimac]

Should the "PM2.5" field perhaps be renamed to something like "PM2_5"? I think the dot could make parsing and handling a bit problematic in some cases.

[tyeth]

Fair point, no one likes dots in their javascript field names...
Stricly speaking there is a less than symbol I wished to place beforehand, and I'm not a fan of PM for particulate matter, as it is easily confused for PPM when the unit is often ppb. The parsing and handling should be done in json-land not scraped from front page, and it's a legal fieldname so all is well if a little unpleasant for handling. I do half agree though, so I'm sorry for playing devils advocate.

[barbudor]

Other particule sensors also use the dot
It's indeed far from perfect for parsing but can be handled as xxx['PM2.5']

[Jason2866]

Yeah, remember had to do some trickery to get the dot parsed correctly in OpenHab

[noppingen]

Does anybody have a solution for this:

Tried numerous build options, always the same result. PM values start at around 1000 after power on and then increase linear. The SEN54 is obviously detected ok.

A buffer overrun maybe? Sensirions SPS30 needed an increased i2c buffer size earlier, is that somehow related maybe?

[noppingen]

Note to myself: Using an ESP32S2 with 240MHz and setting Sleep 100 fixed the issue. I guess my WebQuery rules left too little ressources for the loop.

[Jason2866]

Indeed it is stressing the one core S2 a lot "answering" WebQuerys. There is a reason why mqtt is the heart of Tasmota.