Category: electronics

I work on 900MHz point to multi-point data radio systems.  Humidity can greatly affect the propagation of rf waves at this frequency.  I wanted to do some data logging and analysis of how moisture in the air affected the received signal strength indication over time.

The GE MDS 9710 P70 radio system has an analog voltage test point that corresponds to the rssi on it’s interface board.  I used a DHT22 to measure the temperature and humidity,  an MCP3004  10-bit analog to digital converter to read the rssi voltage, and a pi zero W for the brains.

I originally implemented this project with an arduino nano every.   It worked great.  I was going to log the data by connecting the nano every to a pi zero via UART while a screen session was running on the pi with the screen output (STDOUT) saved to a file .  This method would require a logic level converter which I did not have on hand at the time, so I decided to implement the whole project on my pi zero.

It is much easier to use a DHT22 sensor with arduino using adafruit’s unified sensor library.

Here is the source:

 
// REQUIRES the following Arduino libraries:
// - DHT Sensor Library: https://github.com/adafruit/DHT-sensor-library
// - Adafruit Unified Sensor Lib: https://github.com/adafruit/Adafruit_Sensor

#include "DHT.h"
#include 
#define DHTPIN 2     // Digital pin connected to the DHT sensor

// Uncomment whatever type you're using!
//#define DHTTYPE DHT11   // DHT 11
#define DHTTYPE DHT22   // DHT 22  (AM2302), AM2321
//#define DHTTYPE DHT21   // DHT 21 (AM2301)

DHT dht(DHTPIN, DHTTYPE);

const int rs = 12, en = 11, d4 = 10, d5 = 9, d6 = 8, d7 = 7;
LiquidCrystal lcd(rs, en, d4, d5, d6, d7);

//number of samples
int sn = 0;

void setup() {
  Serial.begin(9600);
  Serial1.begin(9600);
  // set up the LCD's number of columns and rows:
  lcd.begin(16, 2);
  // Print a message to the LCD.
  lcd.print("MDS RSSI Data   Logger");
  delay(2000);
  dht.begin();
}

void loop() {
  // Wait a few seconds between measurements.
  delay(2000);

  //increment sample number
  sn++;

  // Reading temperature or humidity takes about 250 milliseconds!
  // Sensor readings may also be up to 2 seconds 'old' (its a very slow sensor)
  float h = dht.readHumidity();
  // Read temperature as Fahrenheit (isFahrenheit = true)
  float f = dht.readTemperature(true);

  // Check if any reads failed and exit early (to try again).
  if (isnan(h) || isnan(f)) {
    Serial.println(F("Failed to read from DHT sensor!"));
    Serial1.println(F("Failed to read from DHT sensor!"));
    return;
  }

  //read rssi voltage
  int bv = analogRead(A0);
  float v = (5 * bv) / 1023.0;
  float rssi = v - 1.5;
  rssi = (rssi/0.8) * 20 - 120; 
  
  // Compute heat index in Fahrenheit (the default)
  float hif = dht.computeHeatIndex(f, h);

  //print to console
  Serial.print(F("Humidity: "));
  Serial.print(h);
  Serial.print(F("%  Temperature: "));
  Serial.print(f);
  Serial.print(F("°F  Heat index: "));
  Serial.print(hif);
  Serial.println(F("°F"));

  //print to pi
  Serial1.print("Humidity: ");
  Serial1.print(h);
  Serial1.print(F(" Temp: "));
  Serial1.print(f);
  Serial1.print("F RSSI: ");
  Serial1.print(rssi);
  Serial1.print("dBm");
  Serial1.print("\n");

  lcd.clear();
  lcd.print("Humidity:");
  lcd.print(h);
  lcd.print("%");
  lcd.setCursor(0,1);
  lcd.print(f);
  lcd.print("F ");
  //lcd.print(v);
  //lcd.print("V");
  lcd.print(rssi);
  lcd.print("dBm");
}

To use perl to interface with the DHT22 on my pi required the RPi::PIGPIO::Device::DHT22 module on cpan.  Before I could use this,  I needed to install PIGPIO and run it as sudo before executing my script.  PIGPIO is a VERY interesting library for controlling a pi’s gpio’s locally or remotely via socket, and I will hopefully find the time to explore more possibilities in the near future.  It is, however, somewhat of a resource hog.

Once all the modules are loaded on the pi and the PIGPIO daemon is fired up,  I just had to translate the ardino C code into perl.

 
#!/usr/bin/perl 
use strict;
require "/home/pi/gpio/MCP3004.pl";
use RPi::PIGPIO;
use RPi::PIGPIO::Device::DHT22;

=pod
loggs the temperature, humidity, and rssi of a
control point radio using a DHT22.  requires
PIGPIO daemon to be started first.  from home dir, 
cd/PIGPIO.  then, sudo pigpiod.  cd back into
gpio dir, and sudo perl CTC_datalogger.pl.  
=cut

my $pi = RPi::PIGPIO->connect('127.0.0.1');
my $dht22 = RPi::PIGPIO::Device::DHT22->new($pi,23);

#ADC channel
my @ch0 = (1,1,0,0,0);

#init ADC
init3004(3,4,5,6);

#open log file
open RD, ">", "ctc_data.log" or die $!;
#print table description
print RD "index,epoch,date,humidity,temperature,voltage,rssi\n";

my $i = 0;
for(;;){
	#formatted date string 
	my $time_str = `date`;
	print "$i\n";
	print $time_str;
	chomp($time_str);

	$dht22->trigger();
	my $deg_f = ($dht22->temperature * (9/5)) + 32;
	#$deg_f = sprintf("%.1f", $deg_f);
	$deg_f = int($deg_f);

	print "Temp: $deg_f \n";
	my $humidity = $dht22->humidity;
	$humidity = sprintf("%.1f", $humidity);
	print "Humidity: ".$dht22->humidity."%\n";
	int($deg_f);

	#take 5 samples from ADC and avg
	my $v_avg = 0;
	my $v_accum = 0;
	for(my $s=0;$s<5;$s++){
		my ($reading, $binval, $voltage ) = read3004(\@ch0, 50, 4.966);
		#print "binval: $binval\tvoltage: $voltage vdc\n";
		#print RD "$s\n$voltage\n";
		usleep(1000);

		$v_accum += $voltage; 
	}#end for
	$v_avg = $v_accum / 5;
	$v_avg = sprintf("%.2f", $v_avg);
	#print "avg: $v_avg\n";
	
	#calculate RSSI
	my $rssi = 0;
	$rssi = $v_avg - 1.5;
	$rssi = ($rssi / 0.8) * 20;
	$rssi -= 120;
	
	#format rssi
	$rssi = sprintf("%.2f", $rssi);
	print "rssi: $rssi dBm \n";
	print "-----------------------------------\n";

	#print to LCD
	`./lcd "Humidity: $humidity%     $deg_f F $rssi dBm"`;

	#print to log file
	my $log_str = '';
	my $uts = time;
	my $hr_time = localtime();
	$log_str = "$i,$uts,$hr_time,$humidity,$deg_f,$v_avg,$rssi\n";	
	select((select(RD), $|=1)[0]);
	print RD $log_str;

	sleep(2);
	
	$i++;
}#end for

close RD;

So far, it’s working great on the test bench.

Forgot to mention.  I used a DS3231 real time clock to keep time on the pi.  It works great and was easy to set up using this tutorial.

I will be deploying the data logger in the field this week and will post the results.  My bench test results are as follows.  Here is my humidity graph after nearly 9K samples.

And here is the rssi over the same dataset.  You can definitely see a higher rssi when the humidity is lower as expected.