Hoi i have with success build this version of mySQM. But temperature
looks strang compair with equipment i have in garden
this equipment display temp 6.7 and humidity 89% while mySQM see below
i used the same sensor as in the pdf is displayed
Any idea which can explain this different
while the pressure seams seams correct
Ayy suggestions are welkom
Chris
TEMPERATURE
Air Temp 13.490
Dewpoint 4.530
Sky Temp [A] 17.13
Sky Temp [IR] 3.67
DARK SKY
Dud sensor, i had that in mind as well, it looks like more and more
I flashed it again, dubble check the libarries attached in the firmware, a clean setup. I had there still the version of mySQM without wifi. to be sure that its not related to that.
In de meantime i order some other new sensor as well expected tomorrow.
Nice things to play with in the weekend.
I need to find where to collect the logs
Also connected the sensor to J8 BME280 as described on the print
Chris
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
i like to inform you that, it looks like i found the problem, by following the suggestion.
By this i like to share my findings so it helps others.
I have very small case, which is about the size of the pcb.
although i have drilled some holes in the case for air flow it looks like that the LM voltage regulator was the cause of higher temperature value. 1cm distance between LM and sensor. Tip place the BME far enough from the LM so it can't influence de temp anymore ore even out site the case which is what i will do although still protect again have rain. with the suggested idea which can be found in the pdf
below graph illustrated the differents
Left from from the peak is the wrong value cause by the LM
Right the free flow of air
The blue line is from the sqm while the orange /yellow one is from an internet weather station.
Unless one wanted to get an industry grade type sensor at a sigificant cost, the BME280 is actually a good sensor to use.
The PDF on the BME280 sensor says
"Please note that exposure to high humidity (such as being covered in water or snow) can lead to break-down of the sensor over time. Care should be taken in enclosing the sensor in a suitable protective container as indicated above. The sensor needs to be mounted external to the case, and should not be mounted internally otherwise the sensor could be affected more by the humidity/temperature inside the case compared to the external environment. Mounting the sensor on the underside of the case to prevent dew forming on the sensor is also a good idea."
At the end of the day. users are free to mount the sensors any which way. The PDf does try to give reasons about where the best place would be to mount it.
R
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
From earlier experience with the BME280 in another project I noticed that the temperature sensor of this device tells about 3°C more than any other sensors of commercial weather stations (consumer quality).
I tried 3 different thermometer and in comparison to them always 3°C away.
In my simple old project I have set a fixed offset but unfortunately the behaviour is only linear within a range of +5°C to 25°C so far I can remember. Outside this range it becomes non-linear. So it would be an idea to implement a correction table and a manual procedure to fill up this table. I can imagine that all BME280 temperature sensors will have the more or less same behaviour that an code implemented "standard" correction table would maybe help for other users.
In the web I found additionally more informations.
First one is that the embedded sensor is more or less used to compensate temperature impact on the barometer and not officially to be used as a thermometer.
Second one is to reduce the read-out sequence to low values because of every read-out warms up the sensor. Here in consequence I have lowered the read out interval from 4000ms to 16000ms in the code.
Regards,
Hardy
PS: for a future relase but unfortunately with lot of impact on the PCB, sketch and more or less all parts of the project the usage of a Dallas DS18B20 sensor would be a great improvement. This sensor is very reliable!
Last edit: elythomaslumber 2021-12-19
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I think there are likely to be different opinions on this.
As long as the bem280 sensor is not fitted to the pc board, mounted inside the case or outside the case next to other components, it does give good performance. Performance will suffer if it is being affected by the heat generated from other components. This is why the mounting suggest was inside a small case as shown in the pdf, and mounted by itself on the underside of the case (so it does not get exposed to rain or snow).
And I think it is more than accurate for the way we are using it. I would say that having used ds18b20 in other projects that I have never had two probes measure the same temperature, which is why in those projects there is always an offset available in the firmware to correct the reading.
I have not found this to be the case with the BME280. It is more than probable that the coefficients being used in the code of other projects are incorrectly setting up the calibration of the sensor, leading to different results.
It is true that the internal temperature sensor is used - italics - to compensate for the humidity and pressure values of the semsor - italics - as one would expect it to be, Because those values are indeed affected by temperature, so need to be compensated for differences in temperature
From the data sheet I can see that the for the bme280 sensor,temperature has an accuracy of +-0.5 of a degree at a temperature of 25 degrees, and an accuracy of +-1.0 degree over 0-65 degrees
The DS18b20 is slightly better, but, and this is important, the ds18b20 probe - the one with the metal case on the end - its accuracy depends upon how the sensor is encased and the thermal connections thatexist between the sensor and the metal end piece. That is why there is a large variance between ds18b20 probes.
Of course, it would be a simple matter to attach a ds18b20 to a esp32 daughter board and provide the reading that way, if one felt so inclined.
Regards
Robert
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi Robert,
thanks a lot for your fast response and explanation!
I will take your advice when designing my housing for the for the device. I will not use the PCB but a breadboard with copper strips. So I will place the BME280 in a separate box with good ventilation.
I'm excited about the next major relase you've anounced and hope my IR-sensor will arrive soon from China.
Best regards and Merry Christmas to you and yours...
Hardy
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hoi i have with success build this version of mySQM. But temperature
looks strang compair with equipment i have in garden
this equipment display temp 6.7 and humidity 89% while mySQM see below
i used the same sensor as in the pdf is displayed
Any idea which can explain this different
while the pressure seams seams correct
Ayy suggestions are welkom
Chris
TEMPERATURE
Air Temp 13.490
Dewpoint 4.530
Sky Temp [A] 17.13
Sky Temp [IR] 3.67
DARK SKY
SQM 15.71338
NELM 1.92
Lux 0.05598
SkyState CLOUDY
Cloud Cover 72.44
WEATHER
Humidity 54.6
Pressure 1018.73 hPa
Last edit: Chris Alberts 2021-11-24
A dud sensor?
Which header on the pcb did you connect the BME280 sensor?
A log file might also help.
Oh, I nearly forgot, the promised update arrives tommorrow,
u might want to upgrade to that first.
Hoi thanks for the quick replay
Dud sensor, i had that in mind as well, it looks like more and more
I flashed it again, dubble check the libarries attached in the firmware, a clean setup. I had there still the version of mySQM without wifi. to be sure that its not related to that.
In de meantime i order some other new sensor as well expected tomorrow.
Nice things to play with in the weekend.
I need to find where to collect the logs
Also connected the sensor to J8 BME280 as described on the print
Chris
Hi Chris
The more I know the better the answer I can give you,
Did you build the unit with pcb or did you create your own or just wire some components up
I need to know more about the hardware and how it was built. Are you running Linux or something else.
I think the problem is likely to be the hardware, than firmware, as the firmware has been around for some time and is pretty stable,
So I would probably have a few questions about the hardware, can you provide any pictures of your controller showing connections etc.
I am happy for you to email me those direct to my email which wou will be able to find on the front page of the document pdf
I would probably start with the hardware first
I woul appreciate it you can send the firmware folder (you can edit out any ssid/password) that you may have edited in the files.
The purpose of that is that I can then recreate your setup here and I have found that helps.
If you are running LOCALSERIAL it is way too hard to get any meaningful logs from the controller.
That is why the logging functions only work with windows application, linux application or the purpose built logger tools.
The new firmware etc
Shortly. It is worth the wait.
R
Hoi
i like to inform you that, it looks like i found the problem, by following the suggestion.
By this i like to share my findings so it helps others.
I have very small case, which is about the size of the pcb.
although i have drilled some holes in the case for air flow it looks like that the LM voltage regulator was the cause of higher temperature value. 1cm distance between LM and sensor. Tip place the BME far enough from the LM so it can't influence de temp anymore ore even out site the case which is what i will do although still protect again have rain. with the suggested idea which can be found in the pdf
below graph illustrated the differents
Left from from the peak is the wrong value cause by the LM
Right the free flow of air
The blue line is from the sqm while the orange /yellow one is from an internet weather station.
Thanks for the support
Chris
I would like to talk about my experience.
Premise:
BME280 in not the most suitable sensor for outdoor use.
Measure change a lot depending in where is located.
For a great measuere, You cannot place it near a wall, but minimun at 5meters
Hi Stefano
Unless one wanted to get an industry grade type sensor at a sigificant cost, the BME280 is actually a good sensor to use.
The PDF on the BME280 sensor says
"Please note that exposure to high humidity (such as being covered in water or snow) can lead to break-down of the sensor over time. Care should be taken in enclosing the sensor in a suitable protective container as indicated above. The sensor needs to be mounted external to the case, and should not be mounted internally otherwise the sensor could be affected more by the humidity/temperature inside the case compared to the external environment. Mounting the sensor on the underside of the case to prevent dew forming on the sensor is also a good idea."
At the end of the day. users are free to mount the sensors any which way. The PDf does try to give reasons about where the best place would be to mount it.
R
From earlier experience with the BME280 in another project I noticed that the temperature sensor of this device tells about 3°C more than any other sensors of commercial weather stations (consumer quality).
I tried 3 different thermometer and in comparison to them always 3°C away.
In my simple old project I have set a fixed offset but unfortunately the behaviour is only linear within a range of +5°C to 25°C so far I can remember. Outside this range it becomes non-linear. So it would be an idea to implement a correction table and a manual procedure to fill up this table. I can imagine that all BME280 temperature sensors will have the more or less same behaviour that an code implemented "standard" correction table would maybe help for other users.
In the web I found additionally more informations.
First one is that the embedded sensor is more or less used to compensate temperature impact on the barometer and not officially to be used as a thermometer.
Second one is to reduce the read-out sequence to low values because of every read-out warms up the sensor. Here in consequence I have lowered the read out interval from 4000ms to 16000ms in the code.
Regards,
Hardy
PS: for a future relase but unfortunately with lot of impact on the PCB, sketch and more or less all parts of the project the usage of a Dallas DS18B20 sensor would be a great improvement. This sensor is very reliable!
Last edit: elythomaslumber 2021-12-19
Hi Hardy
I think there are likely to be different opinions on this.
As long as the bem280 sensor is not fitted to the pc board, mounted inside the case or outside the case next to other components, it does give good performance. Performance will suffer if it is being affected by the heat generated from other components. This is why the mounting suggest was inside a small case as shown in the pdf, and mounted by itself on the underside of the case (so it does not get exposed to rain or snow).
And I think it is more than accurate for the way we are using it. I would say that having used ds18b20 in other projects that I have never had two probes measure the same temperature, which is why in those projects there is always an offset available in the firmware to correct the reading.
I have not found this to be the case with the BME280. It is more than probable that the coefficients being used in the code of other projects are incorrectly setting up the calibration of the sensor, leading to different results.
It is true that the internal temperature sensor is used - italics - to compensate for the humidity and pressure values of the semsor - italics - as one would expect it to be, Because those values are indeed affected by temperature, so need to be compensated for differences in temperature
From the data sheet I can see that the for the bme280 sensor,temperature has an accuracy of +-0.5 of a degree at a temperature of 25 degrees, and an accuracy of +-1.0 degree over 0-65 degrees
The DS18b20 is slightly better, but, and this is important, the ds18b20 probe - the one with the metal case on the end - its accuracy depends upon how the sensor is encased and the thermal connections thatexist between the sensor and the metal end piece. That is why there is a large variance between ds18b20 probes.
Of course, it would be a simple matter to attach a ds18b20 to a esp32 daughter board and provide the reading that way, if one felt so inclined.
Regards
Robert
Hi Robert,
thanks a lot for your fast response and explanation!
I will take your advice when designing my housing for the for the device. I will not use the PCB but a breadboard with copper strips. So I will place the BME280 in a separate box with good ventilation.
I'm excited about the next major relase you've anounced and hope my IR-sensor will arrive soon from China.
Best regards and Merry Christmas to you and yours...
Hardy