Which dust sensor do you use and why?

[Nca78]

As said by @gohan you can't measure air quality with only 2 parameters, you should measure :

• dust and specifically PM 2.5 which are the most harmful, there are now good and relatively low priced sensors for that. It mainly comes from outside (cars, factories, ...) but activities like cooking can also generate a lot of PM.
• formaldehydes which are harmful, irritating and are exhaled a lot by furniture. I've not seen any calibrated sensors available but they must exist as some companies like holdpeak sell some (but maybe they're lying about the specs ?)
• CO2 which concentration increases with human breathing when you close doors/windows. There are sensors but they are expensive for the good quality ones, and if CO2 affects cognitive capacity it is completely and rapidly reversible so it's not a big deal
• VOC which can be a bit similar to formaldehydes in their effects. There is only one calibrated sensor available at a reasonnable price which is BME680. It is calibrated for VOC that match CO2 so it will give you an "eCO2" index.

So my choice went to BME680 + laser scattering PM sensor, and I'll keep an eye open for any good formaldehydes sensor but I don't have too much hope as all gaz sensors need a good an regular calibration, so unless some company makes a new design and supply an advance software suite like Bosch does for BME680

For PM sensors, I am about to test a few sensors but you can buy a Plantower PMS5003 without risks, I don't really believe their claimed accuracy of 10% or their values for PM sizes other than PM2.5 (I think they are calculated/estimated based on PM2.5) but they are clearly accurate enough for a personal/home sensor, they have a good design that manages to keep led and light sensor away from dust buildup and they are very well priced at around US$15. Commercial pm sensors like Laser Egg use Plantower sensors.
If you are ready to invest US$30 then you can have a look at the Honeywell HPMA115S0-XXX, it will give you only PM2.5 but it's fully calibrated and you have a +/-15% accuracy warranted by Honeywell for 20 000 hours.

[vladimir]

@nca78 said in Which dust sensor do you use and why?:

If you are ready to invest US$30 then you can have a look at the Honeywell HPMA115S0-XXX, it will give you only PM2.5 but it's fully calibrated and you have a +/-15% accuracy warranted by Honeywell for 20 000 hours.

@gohan tell me please, do I need a special sketch to use it? Or they all work on the same principle?

[Nca78]

@vladimir said in Which dust sensor do you use and why?:

@gohan tell me please, do I need a special sketch to use it? Or they all work on the same principle?

It's working the same way than Plantower sensors, serial communication at 9600 bauds and sending the same data structure with the same checksum calculation, just with less data (only measured PM2.5 and "calculated" PM10 based on PM2.5 measurement).

[alexsh1]

@vladimir The answer is it depends. If you are happy with +/- 40% error, you probably be better of with a generic dust sensor. I think the best would be if you do a bit of research online. This is would be a good start - http://aqicn.org/sensor/

Personally, I moved from a cheap sensor to Plantower PMS7003 (about $23). The amount of time one spends in calibrating and understanding cheap sensors, is just not worth it in my view.

[vladimir]

@alexsh1 Plantower PMS7003 this sensor need not be calibrated? Is it accurate enough?
Unfortunately I have some problems with understanding English.:confused:
My GP2Y1014AU0F sensor now shows a value of about 150 units. As I understand, this is absolutely not the right value for the apartment.

[alexsh1]

@vladimir можно на русском и в личку ;))

[vladimir]

@alexsh1 :smile: :+1:
Wrote in the chat.

[NeverDie]

I just now hooked up the PMS5003. Incredibly easy. Regarding the OP, it says the particles in my living rooms are:
Concentration Units (standard)
PM 1.0: 5 PM 2.5: 8 PM 10: 8

Concentration Units (environmental)
PM 1.0: 5 PM 2.5: 8 PM 10: 8

Particles > 0.3um / 0.1L air:894
Particles > 0.5um / 0.1L air:286
Particles > 1.0um / 0.1L air:61
Particles > 2.5um / 0.1L air:2
Particles > 5.0um / 0.1L air:0
Particles > 10.0 um / 0.1L air:0

The above is from using the Adafruit library.

I don't see how the PM numbers can be right though. I mean the particles >5.um is 0, yet the PM10 is 8. How does that make sense?

[NeverDie]

Hmm.. Looks like the Honeywell may not provide the actual particle counts, but just the PM number? If so, that makes for a much coarser measurement. Also, interfacing to the Honeywell is a bit more involved. You have to request measurements. They don't just arrive on their own.

[Nca78]

@neverdie said in Which dust sensor do you use and why?:

Hmm.. Looks like the Honeywell may not provide the actual particle counts, but just the PM number? If so, that makes for a much coarser measurement. Also, interfacing to the Honeywell is a bit more involved. You have to request measurements. They don't just arrive on their own.

They should, by default it's supposed to send measurements at regular intervals.

[korttoma]

@neverdie

Just received an Plantower PMS5003 allso, do you have a "MySensors" sketch that you could share?

[NeverDie]

@korttoma I haven't connected it to mysensors. At present, I'm just using it to explore emissions from my 3d printer and how best to filter them.

[gohan]

@korttoma if you have a library that can read the values from the sensor, you can easily add the mysensors code

[korttoma]

@gohan said in Which dust sensor do you use and why?:

@korttoma if you have a library that can read the values from the sensor, you can easily add the mysensors code

yeah I know, I just thought that if @NeverDie already made a MySensors sketch I would not have to invent the wheel again.

@NeverDie what lib are you using and how? I2C?

[NeverDie]

@korttoma Adafruit library. It communicates over UART.