airgradienthq / arduino Goto Github PK

When in offline mode, do not show errors like WiFi NA etc.
This should be switched off on display as well as on LEDbar.

TMP_RH:

https://github.com/airgradienthq/arduino/blob/master/AirGradient.h#L135-L136

      float t;
      int rh;

periodicFetchData()

https://github.com/airgradienthq/arduino/blob/master/AirGradient.cpp#L300-L301

    sprintf(result.t_char,"%d", result.t);
    sprintf(result.rh_char,"%f", result.rh);

Are the format strings there not reversed? t is float and rh is int?

This library reads CO2 sensor data from pin D4 (GPIO2). Unfortunately, this pin is also connected to the ESP8266 onboard LED, causing it to blink whenever CO2 data is read. This is rather annoying when the sensor is someplace like a bedroom or nursery.

I can't see any way to fix this in code, but I'm putting this out there to see if anyone else has ideas. My workaround until then is black electrical tape 😄

The root LICENSE on the repo is MIT. The header on the files is CC BY-SA 4.0 Attribution-ShareAlike 4.0 International License. Is the intention that the example content is CC and the library is MIT?

getCO2_Raw simply looks for a value of 254 in the first byte to check if a packet is good. Then parses bytes in the array slots 3 and 4 into a value.
A standard response from the S8 is Array[] { 0xFE, 0x68, 0x02, 0xXX, 0xXX, 0xXX , 0xXX }

The data in the 4th and 5th bytes are the CO2 response value while the bytes in the 6th and 7th are the CRC.

If this data stream is out of sync the simple check for the “Any sensor” address in the first byte can cause you to report an incorrect value of 17410. I have seen this issue on my sensor multiple times. Implementing a CRC check would validate the data an prevent bad values from being sent.

Hi,

is there a way to enable or disable the automatic baseline calibration?

By default, is it enabled or disabled?

Add the display of current AQI, either calculated on the server or locally.

It appears that the LEDBar and Display are not completely off at 0% brightness but still visible. These should be completely off.

Hi! I see "CO2 Successfully Initialized. Heating up for 10s" message after "ag.CO2_Init" in my code. Where can i change this 10 sec value?

Hello. I'm a complete noob cf arduino but I managed to complete this nice build thanks to your instructions. I now have an airgradient basic package that sends readings to a Grafana dashboard. Very nice

However, like many users (from what I've read), I noticed that the SHT31 sensor that I'm using, is offset by 0.6-1 degrees celsius. From the investigations of some users on the airgradient forum, it seems that it is caused by the wemos d1 chip itself, transmitting its heat through conduction through the wires themselves.

I've tried mitigating this issue by wiring the sht31 using long cables to the airgradient pcb and secluding the sensor out of the enclosure and air-tightening it from the rest of the components. But there is still some 0.6-1 degrees too much.

So would it be possible to implement an offset to the temperature reading inside of the airgradient.h library?

I upgraded my AirGradient One V9 from 3.0.0 to 3.0.1 trough the www.airgradient.com/documentation/one-v9/ page and realized that I now don't get a noxIndex value from the API from airgradient.com. Then I found the www.airgradient.com/documentation/factory/ page where 3.0.2 is available, but flashing this also did not fix the missing nox value. It is shown on the sensor display.

• Where should I post this? Since I did not change anything else than the firmware, I think this has to be the reason.
• Where is the correct place to find the actual latest (stable) firmware?

Hello,

Thank you for a great product.

The new Prometheus /metrics endpoint is great, so I no longer need to POST periodically to a server. But the code requires successful POST to the server URL or it turns on an error LED.

Would it be possible to make periodic server POST reporting configurable in the server config? Then people can disable it if they don't need it.

Thanks! :)

Remove displayMode: "on" in firmware. Was replaced with brightness properties

Hi,

can you please split the libraries, so that there's a dedicated one for the SensAir S8 sensor?

In the code the CO2_init is taking:

void AirGradient::CO2_Init(){
  CO2_Init(D4,D3);
}

CO2_Init(int rx_pin,int tx_pin){}

void AirGradient::CO2_Init(int rx_pin,int tx_pin){
  CO2_Init(rx_pin,tx_pin,9600);
}

However in schematics the rx/tx is connected as such:
D4 --- TX
D3 --- RX

which seems like contradicting with the code where rx_pin is passed in D4, tx_pin is passed in D3.

Did I miss something?

I'm not sure if this is most appropriate here or in the Forums, since it seems to be more related to the Dashboard page.

Going to the Hardware page I can see all of my devices and when they were last seen:
https://app.airgradient.com/settings/hardware

I see a column for "CO2 Calibration Requested" with a checkbox, but you cannot check the box here to request calibration. I think this is confusing as an end user, so if this is not supposed to be clickable, I think changing it to a Yes/No or True/False string would be better to understand.

I did find after poking around that if I click the three dot menu to the far right of a Monitor, I could select "Edit Location" and see another "Request CO2 calibration", which I checked and clicked Save. In the Hardware page, I briefly saw it checked, but clicking the blue Refresh button cleared the checkbox. After the first time I did this nothing updated on my sensor that I could tell, even though the checkbox had cleared. I still had incorrect numbers for CO2.

I tried it again about 5 minutes later, and maybe it did eventually work from the first attempt, but after requesting calibration a second time and going back to see the current values, it was updated.

Summary:

• Checkbox in the hardware page is misleading since users can't check it and must Edit Location to request it
• In the Hardware page, a Yes/No value for "CO2 Calibration Requested" would be more clear
• In Edit Location, I think the c in calibration should be capitalized to match the Hardware page where it is capitalized. Consistency is key
• If possible, once calibration is requested, the checkbox/true on the Hardware page should stay until the calibration is completed, rather than resetting immediately
• I don't think "Edit Location" is an intuitive place to find the option to request CO2 calibration. Perhaps its own menu under the 3 dots in the Admin column, so it had "Edit Location", "Show Monitor Information", and "Request CO2 Calibration"

In my fork of this library for the 4.2 pro, I've made these improvements:

• Switch to unicode font
  • CO2 -> CO₂
  • PM -> PM₂₅
• NOX -> NOx
• Limit temperature to 1 instead of 2 digits precision
• Optional compile time flag for minimum brightness
• Vertically align the values correctly
• Fix version comment at the top of the file (said 3.7 previously)

Should I create PRs for (some of) these changes?

Result (full brightness):

My sensor currently is on FW 0.1.8 and the target FW supposedly is 9.3.0 but if I do a browser flash it keeps saying 0.1.8.

Prometheus/OpenMetrics was supported in earlier versions. This was removed and would be great if it was added again.
@dechamps originally performed the updated.

I'm not sure if this is most appropriate here or in the Forums, since it seems to be more related to the Dashboard page.

Going to the Hardware page I can see all of my devices and when they were last seen:
https://app.airgradient.com/settings/hardware

The "Current Boot Count" is not incrementing on my ONE that I recently upgraded to 3.1.1. I certainly see in the logs that it is sending a value for "bootCount" that is incrementing, but the Dashboard only shows 0.
I suspect it may be an incorrect value name, since this link shows that it should be called boot and not bootCount
https://github.com/airgradienthq/arduino/blob/master/docs/local-server.md

The instructions here mention selecting a "Lolin (Wemos) D1 Mini board", but current versions of the ESP8266 toolchain for Arduino have multiple sub-variants with different amount of flash memory.

The D1 mini (clone) appears to match the board that comes with the kit (4MB flash).

Property CamelCase to be fixed.

First up - the code compiles fine in the usual arduino IDE.

But - in platformio - where it builds with certain warnings enabled - in this case -Werror=return-type - we get:

AirGradient.cpp: In member function 'TMP_RH AirGradient::periodicFetchData()':
AirGradient.cpp:306:16: error: control reaches end of non-void function [-Werror=return-type]
  306 |     returnError(error);

I can fix it locally by just changing line 306 to:

return returnError(error);

I guess that the arduino IDE doesn't enable that warning.

But - it does look like a return on that line would be a good idea?

Not entirely sure - but thought I'd mention it.

Hi all,

I'm putting the Wemos in deep sleep mode, but the sensors are still functioning. How can I disable them as well when in deepsleep?

Looking at logs after installing the 3.1.1 firmware, I see a message that I was not expecting:

[Configure] Info: Parse configure: {"country":"US","pmStandard":"ugm3","co2CalibrationRequested":false,"ledBarTestRequested":false,"ledBarMode":"co2","displayMode":"on","placeOpen":false,"abcDays":8,"tvocLearningOffset":12,"noxLearningOffset":12,"temperatureUnit":"c","targetFirmware":"3.1.1","ledBarBrightness":10,"displayBrightness":0,"model":"I-9PSL-DE"}
[Configure] Info: Parse configure success
[Configure] Info: Update ignored due to local unofficial changes

I haven't attempted to make any local changes with the API or anything, just flashed the firmware and then opened the logs.

Is this message misleading or what would cause it to see local unofficial changes?

First boot of a device after clean install of the 3.1.1 firmware. This was while I was connecting to the hotspot on my phone and configuring my Wifi information.

Note the request handler not found and No more processes errors

I did not have this unit setup in the Dashboard, so I also got some Error 400 but I don't think that is related to the errors mentioned, unless it is due to not being able to retrieve configuration from the Dashboard.

*wm:-2 networks found
*wm:2 networks found
[ 61179][E][WebServer.cpp:649] _handleRequest(): request handler not found
[ 61185][E][WebServer.cpp:649] _handleRequest(): request handler not found
[ 61240][E][WebServer.cpp:649] _handleRequest(): request handler not found
*wm:Connecting to NEW AP: NachoWifi
*wm:Connect to new AP [SUCCESS] 
*wm:Got IP Address: 
*wm:192.168.2.244 
[ 69910][E][WebServer.cpp:649] _handleRequest(): request handler not found
[WifiConnector] Info: Setting postToAirGradient set from True to True successful
[Configure] Info: postDataToAirGradient: Ignored set to 1
[LocalServer] Info: Init: airgradient_xxxxxxx.local
Mqtt uri is empty
Connect to MQTT broker failed
[ApiClient] Info: Post uri: http://hw.airgradient.com/sensors/airgradient:xxxxxxx/measures
[ApiClient] Info: Post data: {"wifi":-57,"boot":0}
[ 73868][E][WiFiClient.cpp:517] flush(): fail on fd 49, errno: 11, "No more processes"
[ 88502][E][WiFiClient.cpp:517] flush(): fail on fd 49, errno: 11, "No more processes"
[104879][E][WiFiClient.cpp:517] flush(): fail on fd 49, errno: 11, "No more processes"
[ApiClient] Info: Post uri: http://hw.airgradient.com/sensors/airgradient:xxxxxxx/measures
[ApiClient] Info: Post data: {"wifi":-57,"pm01":-1,"pm02":-1,"pm10":-1,"pm003Count":-1,"atmp":-1000.99,"rhum":-1,"bootCount":0}
[105800][E][WiFiClient.cpp:517] flush(): fail on fd 49, errno: 11, "No more processes"

I have two indoor AirGradient ONE's running the 3.1.1 firmware.

Today, following a brief wifi network interruption one of them disassociated from the wifi network.

I don't have it connected to a computer, so I don't have any log output to share.

Hey guys,
I really appreciate the work you are doing! I'm in the midst of modifying the firmware of my AirGradient One. I came across an issue when compiling the 3.0.0 release:

Air Quality Sensor/src/bsp/BoardDef.cpp:323:6: error: narrowing conversion of '-1' from 'int' to 'uint8_t' {aka 'unsigned char'}

I'm on really thin ice here as I'm an absolute C++ noob, but I suspect the issue is that the resetPin in BoardDef.h is declared as uint8_t but assigned the value -1 in BoardDef.cpp?

I'm lacking the knowledge to make an actual fix, I was able to get around it with the compiler flag -Wno-narrowing.

After flashing the latest 3.1.1 firmware, I had the display and LED working at normal levels, and now the display has gone completely off, and the LEDs are at a low brightness level

I assume this is some kind of night mode kicking in, but I haven't configured a night/sleep mode in the Dashboard, nor can I even find where that may be an option to enable or configure.

In my time zone it is only 8:00 pm and not when I would want the unit to be in a sleep mode, but the end user should

1. Opt-in for sleep mode
2. Be able to configure the sleep time window in the Dashboard

This is all assuming this isn't just a bug.

Put device in offline mode
Do factory reset with long press
It appears factory reset does not delete offline mode

Factory reset should put the device in the same state like a complere new flash with erase of all partitions.

Is the hardware / software combination able to handle IPv6? If so it would be nice to have it as an option in the Config secion…

https://www.airgradient.com/open-airgradient/instructions/diy-pro-v42/

If you use Chrome to upload the firmware from the website, everything works fine. If you follow the instructions and manually compile and flash the "Example" file provided (DIY_PRO_V4_2) in Arduino IDE there are several issues:

• The display is upside down.
• The device boot loops unless you press the hardware button to enter the config.

Does anyone have the actual sketch that was used to generate the web flash bin file?

Add support for displaying PM/CO2/etc. quality or color, instead of or alongside raw values.
Would require safely parsing the returned data.

Like other sensor names, TVOC should be in all caps, not lower case.

It is an acronym and should be capitalized, as well as matching the formatting of all of the other displayed readings.

Also, I believe it should be VOC instead of TVOC, as that is what the manufacturer refers to it at:
https://sensirion.com/products/catalog/SGP40/

Attempting to compile and flash from source, the ONE_V9.ino seems to be missing some of the dependent libraries. I run into missing headers PMS.h and s8_uart.h - it looks like the latter is from this library but I can't figure out what the former is from - I found this library but it has an incompatible DATA declaration:

/home/stevekuznetsov/Arduino/libraries/AirGradient_Air_Quality_Sensor/examples/ONE_V9/ONE_V9.ino: In function 'void updatePm()':
ONE_V9:290:26: error: 'struct PMS::DATA' has no member named 'PM_RAW_0_3'
       pm03PCount = data1.PM_RAW_0_3;
                          ^~~~~~~~~~
Multiple libraries were found for "WiFiClient.h"
 Used: /home/stevekuznetsov/.arduino15/packages/esp32/hardware/esp32/2.0.11/libraries/WiFi
 Not used: /app/Arduino/libraries/WiFi
exit status 1
'struct PMS::DATA' has no member named 'PM_RAW_0_3'

There seems to be no message when hardware watchdog is reset, at least in offline mode.
OR
Hardware watchdog is actually not reset.

I think it might make sense to harmonize the JSON structure sent to the server. Would make implementation for Home Assistant easier.

See
airgradienthq/python-airgradient#61

@nick-4711 What do you think?

Merged
#129
due to time limits. Comment in #129 still needs to be addressed.

Currently the colors for the LED bar are not defined globally so if we e.g. want to change a color slightly, we need to go through a lot of code. Propose to centralise it, e.g. RGB code for red, green etc and then call the color name.

The AirGradient DIY Air Quality Sensor (Pro Version PCB Version < 3.7) ceased showing voc or tvoc with SGP41 sensor

1% is still quite bright on the monitor. I don't think its 1%. This only affects the LedBar. Display looks ok.

local config shows:
http://airgradient_xxx.local/config

But from server comes:
...,"ledbarBrightness":1,"displayBrightness":18,"...}

The PM2.5 window in the middle of the display looks oddly squished, especially since there is so much blank space to the far right of tvoc:

I also wonder if PM2.5 would properly handle 3 digits, as it looks to only have sufficient space to look good for up to 2 digits.

What if it needed to display 288? Feels like it may not fit within the bars, or at least would be very cramped looking.