DHT Sensors

Try opening up the sensor in your read function.  And at the end, closing it.

Posted by: @mikeburton_2000

Hi folks, I have built a greenhouse monitoring system using multiple soil moisture probes and a DHT11 sensor. I have it all up and running on my NodeMCU sending the info to Blynk and on to my Smartphone. i am quite happy with the way it all working except for the permanently ON DHT. I have the soil sensors powered off and they are switched via a 22N2222 transistor, left on for 20s before the reading is taken, then switched off again. This is done to reduce power consumption from the battery and also reduce corrosion on the probes.

Mike

Hi Mike...  I am interested in using several Moisture Probes, etc. also...  

Can you describe your Moisture Probes, where to buy, etc. & how they work?

It would be nice if you could write-up your complete application...  Might save me a little time from reinventing the wheel.  🙂

Thank you very much!

@mikeburton_2000

If you want to save power, try this library - Has some examples:

Low-Power

Cheers!

@triform I have tried moving dht.begin inside TransON and SensorRead functions but it has no effect.

@mikeburton_2000

when using a DH11 sensor a 10K resistor is usually used between VCC and the data pin as a pull up as the in built arduino pullups are too weak.  As your Fritz diagram does not show this I wonder if this is the cause of your problem.

@joelyddon

The moisture sensor modules were bought off Amazon from A-ZDelivery ( https://www.amazon.co.uk/gp/product/B07CNRJN8W/ref=ppx_yo_dt_b_asin_title_o06_s00?ie=UTF8&psc=1 ) and are the fairly generic resistive type. I made a longer probe for one of my sensor modules to measure a deeper bed where I grow root veg. This was 2 stainless welding rods kept separated at an even distance by 3 insulators I made from some Polymorph using a highlighter pen cap as a mould. (Picture to follow)

I also use an ADC from A-Z (ADS1115 https://www.amazon.co.uk/gp/product/B07QHWLTTS/ref=ppx_yo_dt_b_asin_title_o04_s00?ie=UTF8&psc=1 ) which allowed the additional Analogue inputs for the NodeMCU

To help reduce power and also corrosion of the moisture probes, I used a transistor to switch the power to the moisture modules on and off. These are running from 5v supplied through one of those cheap breadboard power modules ( https://www.amazon.co.uk/gp/product/B06XQ1DSCP/ref=ppx_yo_dt_b_asin_title_o07_s00?ie=UTF8&psc=1 ) which I am running off a 6v lead acid battery.

Here is my fritzing drawing - forgive me if it's not 100% as I am new to that tool too.

My Code looks like this:

//INCLUDES
#include <ESP8266WiFi.h>
#include <BlynkSimpleEsp8266.h>
#include <Blynk.h>
#include <DHT.h>
#include <Wire.h>
#include <Adafruit_ADS1015.h>
//------------------------------


// DEFINITIONS
int16_t adc0, adc1, adc2; //create 16bit integer to read adc channel A0
//int16_t adc1;
//int16_t adc2;

#define BLYNK_PRINT Serial
#define TRANS 15 // Digital pin for Transistor
#define DHTPIN D4 // Digital pin for T and H
#define LEDG D7 // Digital pin for LED Green
#define LEDR D0 // Digital pin for LED Red
#define DHTTYPE DHT11 // DHT 11
//------------------------------

//WiFi & Blynk Info
char auth[] = "cz0G22Y*******Ieoa6Nsdoe";
//char ssid[] = "TP-LINK_B12408";
char ssid[] = "Riverside 4G 2.4Ghz";
//char pass[] = "*******";
char pass[] = "*******";
int status = WL_IDLE_STATUS;

//Class
DHT dht(DHTPIN, DHTTYPE);
Adafruit_ADS1115 ads;
BlynkTimer timer;
WiFiClient client;
//------------------------------

//FUNCTIONS

void SendSensor() //Sending DHT Sensor Info
{

// read the value from Moisture sensors:
digitalWrite(TRANS, HIGH); //Turn on transistor
delay(2000);

float h = dht.readHumidity();
float t = dht.readTemperature(); // or dht.readTemperature(true) for Fahrenheit

if (isnan(h) || isnan(t))
{
Serial.println("Failed to read from DHT sensor!");
return;
}

delay(1500);
adc0 = map(ads.readADC_SingleEnded(0), 10000, 30784, 98, 0);
delay(250);
adc1 = map(ads.readADC_SingleEnded(1), 12000, 30784, 98, 0);
delay(250);
adc2 = map(ads.readADC_SingleEnded(2), 14000, 30784, 98, 0); //Calibration Note: Compost is less resistive then soil/soil& compost mix. Use 8000-9000 for wet compost 6500 - 8000 for wet soil. dry sand = 30784
delay(250);



digitalWrite(TRANS, LOW);
delay(500);
Wificonnect();
delay(2000);

Blynk.begin(auth, ssid, pass);
delay(1500);

Blynk.virtualWrite(V5, h);
Serial.print(h);
Serial.println("% ");
Blynk.virtualWrite(V6, t);
Serial.print(t);
Serial.println("°C");

Blynk.virtualWrite(V7, adc0);
Serial.print("Moisture Sensor 1 ");
Serial.println(adc0);
Serial.println(ads.readADC_SingleEnded(0));

Blynk.virtualWrite(V8, adc1);
Serial.print("Moisture Sensor 2 ");
Serial.println(adc1);
Serial.println(ads.readADC_SingleEnded(1));

Blynk.virtualWrite(V9, adc2);
Serial.print("Moisture Sensor 3 ");
Serial.println(adc2);
Serial.println(ads.readADC_SingleEnded(2));

Serial.println("----------------------------------------");

//adc0 = 1;
//adc1 = 2;
//adc2 = 3;

WiFi.disconnect();
digitalWrite(LEDG, LOW);
digitalWrite(LEDR, HIGH);
delay(2000);
}
//------------------------------

void Wificonnect() {
WiFiClient client;

WiFi.begin(ssid, pass);
while (WiFi.status() != WL_CONNECTED)
{
digitalWrite(LEDG, HIGH);
digitalWrite(LEDR, LOW);
delay(150);
Serial.print(".");
digitalWrite(LEDG, LOW);
digitalWrite(LEDR, HIGH);
delay(150);
}
digitalWrite(LEDR, LOW);
{
Serial.println("");
if (WiFi.status() == WL_CONNECTED)
{
digitalWrite(LEDG, HIGH);
Serial.println("Wifi OK");
}
if (WiFi.status() == WL_DISCONNECTED)
{
digitalWrite(LEDR, HIGH);
Serial.print("WiFi Not Connected");
}
}delay(2000);
}
//------------------------------


void setup() {
Serial.begin(9600);
//Blynk.begin(auth, ssid, pass);
ads.begin();
dht.begin();
ads.setGain(GAIN_ONE);

timer.setInterval(30000L, SendSensor);

pinMode(LEDG, OUTPUT);
pinMode(TRANS, OUTPUT);
pinMode(LEDR, OUTPUT); // set the digital pin as output

digitalWrite(LEDG, LOW);
digitalWrite(LEDR, LOW);

Serial.println("Connecting to ");
Serial.println(ssid);
SendSensor();
}

//------------------------------
void loop() {

Blynk.run(); // T&H
timer.run(); // T&H
//SendSensor();
//SensorSetUp();

}

Again probably not the neatest but I am leaning and it suits me.

I have added a simple Voltmeter to my build (it's in a plastic biscuit tin) which I can use to check the voltage from the battery. The state of the battery and also the temperature of the soil/compost has an effect on the reading the script pumps out to Blynk. However, the whole project gives a good idea of what is happening in the greenhouse when I am not there.

Screenshot from the Blynk app with the basic data I am currently collecting:

I'd like to add logging so I can see what the max-min temps are especially overnight.

Moving it on, I will be adding auto watering solutions for the 3 key areas (Deep bed, Staging and Tomato Bags) via a 12v high-pressure water pump again controlled by the NodeMCU.

My biggest bug-bear at the moment is the inability to turn the DHT on and off as per my post above. But having read a few bits of materials including the datasheet (Chinglish Is not my forte) I may have to resign myself to leaving it powered on all the time - certainly until such time as my skill level increases to be able to write a fancy function to power it back on with the right sequences of highs and lows across the relevant pins.

I would like to get to the bottom of that issue though.

Hope the above helps and if I can do anything more to assist with your project, let me know.

Hi Mike,

I may be reading your fritzing incorrectly as it is very small on my screen. But it looks like you are powering 2 LEDs and the DHT from one of the GIO output. The LEDs also look like they are connect across the GPIO output to ground. If this is how it is set up then you are only getting about 1 volt to the DHT. I don't think the DHT alone can be powered from a GPIO reliably. If you need it I would be happy to provide a simple circuit using a PNP transistor to provide power to you DHT.

Hi @ruplicator and thanks for your input.

In my OP, removing the 2 red LED's does not change the result. I put them there to verify that I was actually getting power where and when I thought I should.

In my full-scale project (my second and larger post above) I have a small red and green LED in the circuit each one powered directly from D0 & D7 respectively. The DHT11 gets its VCC & GND from the 3v bus bar on the right of the breadboard which is in turn powered from an external source (breadboard power supply running from 6v lead-acid battery). The DHT uses pin D7 for its data IO.

The breadboard power supply also powers the 5v bus bar on the left side of the breadboard and this also used to push 5v into the NodeMCU VIN pin and runs the soil moisture modules due to the cable length to the probes. Both bus bars are connected to ground pins on the NodeMCU also. There is a 3rd (yellow) LED which is drawing power from the D8 pin which is used for the transistor too. 

I would be very glad to see your diagram of a transistor switching the DHT as every attempt I have made to switch power to the DHT11/22 (including using a 2n2222 similar to that on the Moisture Sensor array) results in it failing to read the T&H. Though if leave everything the same and switch power to the DHT from the bus bar, it works fine.

I hope this helps and look forward to seeing your diagram.

@mikeburton_2000

You second fritz diagram also appears to lack a 10k resistor I mentioned above, have you tried this?

First I'm a little concerned that you are powering the DHT from a 3 volt supply. It's spec sheet says it will work on 3.3V - 5.5V. Even if your supply is 3.3V that leaves no room for us to add a transistor (with its voltage drop) in the power circuit. You indicated that you have a 5V supply available and if you can use it that would be much better. The circuit below uses a PNP transistor rather than your 2N2222 which is an NPN. Almost any PNP transistor will work as you will need very little current to run your DHT. The 4.7K resistor is also not critical critical in this application, just use something close. Also you will need to change your software to output a LOW rather than a HIGH when you want power to be sent to the DHT. One last note: The spec sheet did indicate that the DHT shouldn't be read for at least 1 second after power has been applied and that it might be a good idea to add a 100 nf capacitor to it's input power source.

Data Sheet: https://www.electronicoscaldas.com/datasheet/DHT11_Aosong.pdf

Please disregard my last post. I tried to edit it after I was able to look at your problem more closely but the system wouldn't let.

After looking at you layouts more closely and the spec sheet of the DHT11 the recommendation to switch the power as proposed doesn't seem to be your most probable problem.

1. If your sensor is/will be located away from the processor board the 3.3 V supply may not be adequate for a good signal back to the processor.

2. Some older/cheaper DHT11 sensors don't include a pull-up resistors on the sensor as Byron indicated. If you don't know if yours has this then you will need to include it on the data line.

3. The spec sheet indicates the DHT11 only pulls 1.5 mA maximum during measurement. If this is the case for your unit it may not be worth it to try to switch the power. If you decide you still need to switch the power I think the circuit you are using on on your other sensors should also work fine for the DHT11.

Sorry for the confusion. 

@byron Hi and thanks for your thoughts on this. I have added the pullup to the VCC and Data Pin since the fritzing was drawn up and even changed the bare sensor for a DHT module that includes the resistor. The results alas was the same. 

@ruplicator I have made a boob. The Fritzing for the larger project, as you rightly say, shows the DHT running from 3v. However, on the actual project, I have changed it to the 5v bus bar and not updated my Fritzing Image.

The DHT sensor wiring is about 350mm long so it is not mounted a long way from the project box and I did run the details through my calculator to check it would be OK at that length and that is probably what made me swap it to the 5v rail.

I have tried it with and without the pullup and it is currently running with a DHT11 module which includes the resistor, and again, this made no difference.

I will find a PNP and try it. In all the permutations I have messed with I only used that one transistor type (it's all I have in my box at the moment) so it is definitely worth another mix to see what happens. If I add the DHT11 Module to a similar setup as the soil moisture sensors (2N2222) it powers on fine, but does not read the sensor and instead returns "nan" (Not A Number) as the value for both D and T. I have verified that the sensor is actually on when I think it is, and I have inserted a delay to ensure the sensor is not read in the first 2 seconds to avoid the issue noted on the datasheet. The Chinglish datasheet also gave me the impression that some fancy data pin switching may be required to effectively reset the state of the sensor before it can be read. But it was a difficult read and for a simpleton like me, I may have got the wrong end of the stick.

As for the power saving, that would definitely be something I would like to achieve as the small 6v lead-acid lasts a couple of days and I would like to stretch that as best I can. But as much as anything, it's now also about me wanting to find a solution to the switching on/off issue because I feel I must and don't want to be outsmarted by a little pile of... whatever those things are made out of! If I simply can not find an answer I am considering jumping to a BME280 instead - but I don't like to give up!

Once again, thanks for your thoughts and keep em coming.

When you say connecting the DHT11 VCC to the 3V on the NodeMCU makes it work fine, do you mean you can disconnect the 3V and then reconnect it and it works with your code (i.e. after a 3.5 second delay you get correct readings) but does not work when VCC is supplied from the D5 pin (again after a 3.5 second delay)?  If so out with the multimeter 😀. But I may have got the wrong end of the probes.