DHT22 generates off scale values

[Pavel Hrudka]

Hi.
Iam building battery node for temperature.
Before I move to battery related stuff I wanted to have two nodes up and running. In that way I can have reference node which is powered from USB DC 5V and battery powered node.

This way I could evaluate battery nodes measurements before moving in to production.

BTW: All nodes are hooked to NRF radio, I did not mentioned it before. That is why there is 4.7 uF capacitor.

1. I left my second node powered from programmer, but its measurements were total different. So I connected it to the same 5V DC as the first one. Its values are now almost identical.
2. However, from time to time I see reading offscale values (spikes).

R = 4K7
V = 4.7uF

Here are some graphs

and

Iam using
Arduino 1.6.7
DHT sensor library by Adafruit v 1.2.3 (also Adafruit DHT unified 1.0.0)
MySensors library 1.5.4
OpenHab 1.8.3 with openHab binding.

Sketch is here: 0_1465243821164_HumiditySensor_101.ino

Only reference to similar problem I found here.
Which says that when connected to 5V problem disappeared. But DHT22 should be running fine from 3V?

Does anyone have idea what could cause these occasional spikes?
Please help.

Thank you.
PH

UPDATE 2016-06-07 DHT22 connected to 5V suply. Spikes are still there. Ill try to move capacitor closer to radio.

[drock1985]

The DHT22 temperature sensor requires a minimum of 3.3v to operate; any less and if fluctuates. The only luck I have ever had with the DHT22 is powered from a stable 5v supply; forget about 3.3v.

The SI7021 and BMP 180 /BME 280 do operatore @ 3.3v though, and quite well on battery power.

[Pavel Hrudka]

Hi thank you for responding.
Interesting thing for me is that first Arduino does not show the spikes.

Anyway I tried to connect DHT22 sensor on second one to 5V DC from USB and will monitor it.

Maybe its in the way that DHT22 likes 3.3V and any fluctuations will lead to instabilities?

Cheers PH

[alexsh1]

@Pavel-Hrudka I have DHT22 on Particle Photon, which is 3.3V. Your power supply must be rock solid and this is not the best sensor for batteries.

[Pavel Hrudka]

Connected DHT22 to 5V suply and spikes are still present.
Ill try to move capacitor closer to radio to decouple it from the rest.

Cheers.

[drock1985]

You could try this. Solder your 4.7uf cap right to the GND+VCC pins on the NRF radio. That would at least isolate that.

IF not, you may have a bad radio too.

[Pavel Hrudka]

Just to let you know, that best result I got was when I changed the DHT library (sparkfun) for the one with this sketch. with this sketch.

Thank toy for your time.
PH