v2.0.0-dev10 - AM2320 issues

[bluejedi]

In v2.0.0-dev10 the AM2320 is now supported but I observed the following issues:

• The are 3 different firmware version in the package: normal test and dev.
  In normal AM2320 is not supported (but expected because AM2320 is mentioned in the release notes), in test it is supported and dev I have not checked.

• AM2320 is not displayed in I2C Scan. The AM2320 works but in I2C Scan only 3 of 4 I2C devices (in my config) are listed and AM2320 is not one of them.

• AM2320 is listed as 'Environment - AM2320 [Test]' but because it only supports I2C interface it should be named 'Environment - AM2320 (I2C)' (consistent with ('DHT12 (I2C)').

• AM2320 can be used with both I2C and 1-wire bus interfaces. It is currently supported only for I2C.
  In comparison with DHT12. The DHT12 has a separate device entry 'DHT12 (I2C)' for I2C and it is also supported for 1-wire under device 'Environment - DHT 11/12/22'
  The AM2320 should be supported for 1-wire similarly.
  For consistency device 'Environment - DHT 11/12/22' should probably support AM2320 1-wire too and then be renamed to 'Environment - DHT 11/12/22/AM2320'.
  I have created a separate thread about naming issues.

[krikk]

• The are 3 different firmware version in the package: normal test and dev.
  In normal AM2320 is not supported (but expected because AM2320 is mentioned in the release notes), in test it is supported and dev I have not checked.

the am2320 plugin is new, so its only in the testing firmware, because it has only undergone basic testing...

• AM2320 is not displayed in I2C Scan.

this is a know issue of the sensor, no workaround known...

and yes it would be nice if we would also support the 1-wire mode... but someone has to code it ...i wrote the i2c am2320 plugin as a "quick hack", but generally i think its a crappy sensor, so i will not invest more time to support it via 1-wire

[bluejedi]

Crappy sensor? pity.

I'm still looking for a good accurate (not too expensive) temperature and humidity sensor preferably with I2C interface.

BME280 is not designed as precision thermometer therefore not a real alternative.

[Shardan]

Crappy sensor? pity.

I'm still looking for a good accurate (not too expensive) temperature and humidity sensor preferably with I2C interface.

BME280 is not designed as precision thermometer therefore not a real alternative.

Definitely the DHT11/12/22 sensors imho are not made for detailed measuerement.
I think it's more for devices like (de-)humidifiers where you turn a knob until you reach
a sufficient room climate.
I've tested a bunch of DHT22, the result was disappointing. (Anyone need some? Used only once... )

Well good, accurate and not to expensive is an oxymoron.
Cheap choices are BME280 or SI7021. The SI7021 is quite well but i think some
improvement in the plugin is needed as they tend to show a too high temperature.
This might be caused by internal heating.

A highly precise sensor is the SHT1x.
Tolerances on humidity are
SHT11: 4,5%
SHT12: 3%
SHT15: 2%
(Datasheet: http://downloads.cdn.re-in.de/500000-52 ... _SHT15.pdf
These use a connection similiar to I2C but it is not I2C, don't mix!
And swiss made precision has a price. Just the sensor chip is around 25€ here.
On top you need a pcb for it and a covering case.
They are a bit cheaper on AliExpress, but i doubt if that are really
precise sensors or just waste. Anyways the pcb's and covering cases
on AliExpress are nice.

Regards
Shardan

[bluejedi]

Thanks,
€25+ for a temp/humidity sensor is 'over budget'. For €25 you can get a multi-core MCU board with memory and more...
It's remarkable that with all today's technology it appears so difficult to create rather accurate temp/humidity sensors for a low price.

[Shardan]

Thanks,
€25+ for a temp/humidity sensor is 'over budget'. For €25 you can get a multi-core MCU board with memory and more...
It's remarkable that with all today's technology it appears so difficult to create rather accurate temp/humidity sensors for a low price.

Hello,

yes you can get a multidevice for 25€.
But don't be surprised if you find one of these cheap and inaccurate sensors inside that device.
You won't get a Lamborghini at the price of a Fiat.

I've searched a lot for really accurate sensors, be it temperature, humidity, pressure, current, voltage...
I can assure you: Really good sensors are expensive. No workaround, no excuses.

Ruskin's rule applies here too:

"There is absolutely nothing in the world that some man cannot make a little worse and sell a little cheaper;
and the people who consider price only are this man's lawful prey."

What may be a way around:
As ESPEasy supports "formula" at the setup some limited calibration is possible.
I've done this for some DS18B20 temperature sensors using a calibrated lab thermometer near the sensor.

[bluejedi]

Thanks.
I was aware of the formula in ESPEasy but that will still require a good reference device (which I don't have).

SHT31-D(IS) looks promising and has I2C interface.

Article Comparing Temperature/Humidity Sensors (including SHT31-D):
https://www.facebook.com/notes/phil-moy ... 8075605557

SHT-3x data sheet:
https://www.sensirion.com/fileadmin/use ... igital.pdf

https://www.aliexpress.com/item/Free-sh ... 06758.html

[Update]
After carefully re-reading Phil Moyer's article I came to the conclusion that it contains a big mistake: results of 2 sensors seem to have been exchanged, based on hard to read colors in the legenda of the analysis graphs.
The author writes:
"It's a bit hard to read the legend. The red points are the BME280 data, the olive points are the DHT22 data, the blue points are the HTU21D-F data, green points are the SHT31-D data, and, finally, the purple points are the Si7021 data."

However, the colors in the agenda appear so hard to read that the author has problems reading them himself. He writes "blue points are HTU21D-F data, green points are the SHT31-D data". But that appears incorrect!
After analyzing colors in the legenda I come to the conclusion that HTU21D-F data is green (not blue) and SHT31-D data is blue (not green). This is also consistent with the colors that have been used in the observation graphs. The exchange has major impact on the analysis conclusion: it is the HTU21D-F which comes out best from the analysis, not the SHT31-D.

I will take a further look at the HTU21D-F (which is 50% cheaper on AliExpress than SHT31-D).

[Shardan]

The SHT sensors made by Sensirion give quite a good reference as they have a low tolerance equaled to the usual DHT and DS18 sensors.

For temperature I prefer a simple old fashioned thermometer... a glass thermometer used in the laboratories is below 10€.
It's far more difficult to find a good reference for humidity, pressure and luminosity.

It's a bit funny to compare SHT21, HTU21 and SI7021. These breakout boards usually use the same sensor.
Comparing shows more the tolerances of this single sensor type then quality differences.

[krikk]

this is also a nice comparison of sensors: https://www.kandrsmith.org/RJS/Misc/hygrometers.html

[bluejedi]

For temperature I prefer a simple old fashioned thermometer... a glass thermometer used in the laboratories is below 10€.

You mean glass thermometer as reference? They have rather low resolution.

[Shardan]

Far good enough for room temperature... as a standard DS18B20 has
a tolerance of +/- 0,5 degrees a glass themometer is sufficient.
My appartment isn't a lab