Julian Rogers Home Rain gauge Rain gauge - more construction

Under Construction

Rain gauge circuit and software

Circuit

I am using a slot-opto sensor to detect when the bucket flips. Infrared light from an LED is interrupted by thin wire as the bucket flips. The output of a photo transistor momentarily goes from close to zero volts to 5 volts then back to zero (this part of the circuit runs at 5 volts and will need a level shifter when connected to the 3.3 volt Feather). The voltage change is a relatively slow process and if this signal were applied directly to the input of a microprocessor it could result in inaccurate counting. Some slot-optos have a built in Schmitt trigger which produces an output pulse with very rapid changes from zero to 5 volts and back again. However, my slot-opto which I found in my bits box from rather a long time ago was not one of these!


The answer is add a Schmitt. I often resort to the trusty 555 timer which can be wired up to do this job without any extra components. To complete the signal processing, I will use the output of the Schmitt to trigger a monostable circuit (also using a 555 timer) to produce a fixed length pulse of about 11 mS which, hopefully, will avoid any confusion on the part of the microprocessor which will be counting the pulses. To simplify things slightly I will use a 556 chip which is two 555 circuits in one package.


The eighteen pin chip under the 555s (556 actually) on the left is a PICAXE 18A which is a Microchip PIC16F819 microprocessor pre-programmed so that a a serial connection from a PC or similar running a simple version of BASIC can get it do all sorts of (fairlysimple) things with inputs and outputs.



In this particular case a program of a few lines can get the PICAXE to count the pulses emanating from the 556. The count is registered as an 8-bit number on the PICAXE’s eight output pins. These are fed into eight of the MCP23017’s which will be configured as inputs. I2C then transmits this number to the remote computer. The remote computer can reset the PICAXE through a pin on the MCP23017 connected to the PICAXE’s reset pin thus resetting the count. At this stage, I anticipate the remote computer might read the count every hour and a max count of 255 should be sufficient (I hope!)

Alternative schemes


My PICAXE 18A is now redundant and newer versions have I2C facilities. In an I2C network there can only be one master, other devices connected have to be configured as slaves. I haven’t investigated whether these PICAXEs can be configured as slaves.

Arduino Unos, however, can be configured as slaves. The slave can be configured to do stuff like count events and transmit the current value when requested by the master. I have bread boarded this and it works fine. (The 556 would still be needed if I were using the slot-opto - unless it had a built in Schmitt. A reed switch would not need a Scmitt, of course!)

Using a whole Uno for this would seem to be a waste! I tried an Adafruit Trinket which would seem to be ideal but couldn’t get it to work. I’ve had trouble with the Trinket and I2C before. I could get it to read one TMP102 temperature but not two (they were set to separate addresses). It’s probably just me - you can stare at a simple program for hours and not see a blindly obvious mistake!

Back to the original concept! The picture on the left shows the circuit on a breadboard. The breadboard is one specially designed for PICAXEs

555s (I used individual ones here!)

MCP23017

PICAXE

Slot-opto

Arduino Uno

Circuit on a pcb connected to an Uno for testing purposes

The pcb didn’t need the usual modifications for once!

//test port expander inputs on breadboard BASIS for RAIN GAUGE

//picaxe 18A is counting (8-bit) the number of times a slot opto beam is broken

//the signal is conditioned by a Schmitt trigger and a monostable (556 IC)

//Feather reads the value periodically

//20.1.18


// pin 6 (a6 on chip numbering) is connected to picaxe reset pin (needs to be high to run)

// pins 8 to 15 (b0 to b7 using chip numberings) are used to log the count from the picaxe

// The input to the Picaxe from the 556 is on input 2

// input 1 is connected to pin 7 on the MCP23017 (a7) for possible future use


#include <Adafruit_MCP23017.h>

#include <Wire.h>

Adafruit_MCP23017 mcp;

byte b0, b1, b2, b3, b4, b5, b6, b7;

int val;


#define TMP102_I2C_ADDRESS 0x48 // I2C address TMP102 A0 to GND (0x48 = 72 = 1001000 for GND, 73 for vcc)

#define TMP102_I2C_ADDRESS_2 0x49  //second tmp102


void setup() {

mcp.begin(1);      // use address 0x21 (0,0,1) rather than default


Serial.begin(9600);

mcp.pinMode(8, INPUT);

mcp.pinMode(9, INPUT);

mcp.pinMode(10, INPUT);

mcp.pinMode(11, INPUT);

mcp.pinMode(12, INPUT);

mcp.pinMode(13, INPUT);

mcp.pinMode(14, INPUT);

mcp.pinMode(15, INPUT);


mcp.pinMode(6, OUTPUT);

mcp.digitalWrite(6, HIGH);


}



int getTemp102(byte ADD_TMP102){

  byte firstbyte, secondbyte; //these are the bytes we read from the TMP102 temperature registers

  int val; //an int is capable of storing two bytes, this is where we "chuck" the two bytes together.


  float convertedtemp; //We then need to multiply our two bytes by a scaling factor, mentioned in the datasheet.


  Wire.beginTransmission(ADD_TMP102); // start talking to sensor

  Wire.write(0x00);

  Wire.endTransmission();

  Wire.requestFrom(ADD_TMP102, 2);

  Wire.endTransmission();



  firstbyte      = (Wire.read());

  //read the TMP102 datasheet - here we read one byte from

  //each of the temperature registers on the TMP102

  secondbyte     = (Wire.read());

  //The first byte contains the most significant bits, and

  //the second the less significant

  val = firstbyte;

  if ((firstbyte & 0x80) > 0) {

    val |= 0x0F00;

  }

  val <<= 4;

  //MSB

  val |= (secondbyte >> 4);    

  // LSB is ORed into the second 4 bits of our byte.

   

  convertedtemp = val*0.625; // temp x 10

  

  int temp = (int)convertedtemp;

  return temp;

}



/////////////////////////////////////////////////////////////////////////////


void loop() {

b0 = mcp.digitalRead(15);

b1 = mcp.digitalRead(14);

b2 = mcp.digitalRead(13);

b3 = mcp.digitalRead(12);

b4 = mcp.digitalRead(11);

b5 = mcp.digitalRead(10);

b6 = mcp.digitalRead(9);

b7 = mcp.digitalRead(8);


val = b7 * 128 + b6 * 64 + b5 * 32 + b4 * 16 + b3 * 8 + b2 * 4 + b1 * 2 + b0;

Serial.println(val);


int tempNow2 = getTemp102(TMP102_I2C_ADDRESS_2);

Serial.println(tempNow2);


delay(1000);

 

}

/////////////////////////////////////////////////////////////////////////////////


The software below runs on a Feather connected up for testing purposes. The program reads the count on the PICAXE every second together with the temperature registered on a TEMP102 sensor. Eventually, similar software will be incorporated in the overall monitoring and control system.

// test master slave I2C Arduino

// this is the master



#include <Wire.h>

#define DEV_ADD (0x10)


void setup() {

Wire.begin();

Serial.begin(9600);

}


void loop() {

 

  Wire.beginTransmission(DEV_ADD);

  Wire.requestFrom(DEV_ADD, 1);     //request 1 byte

  Wire.endTransmission();


  byte count    = (Wire.read());

  

  Serial.println(count);

 

  delay(3000);

}



// test master slave I2C Arduino

// this is the slave


#include <Wire.h>

#define DEV_ADD (0x10)


int count;


void setup() {


Serial.begin(9600);

Wire.begin(DEV_ADD);

Wire.onRequest(reqHandler);

}


void reqHandler() {

  Wire.write(count);

}


void loop() {

  count++;

  delay(500);

  Serial.println(count);

  if( count == 255) {

    count = 0;

  }


}


For the sake of completeness, the programs I used to link two Unos over I2C are shown below.

These are the waveforms I expect to get from the slop-opto and 555/556 sections of the circuit. Waveform 1 is what I expect from the slot-opto. With nothing obscuring the optical path between the infra-red diode and the photo transistor, the transistor should be saturated (high current) and the output close to zero volts. As the wire swings across the beam, the beam is reduced, then increased again. Current flowing in the transistor is reduced then increased again so the voltage across the 4.7k transistor load resistor swings from nearly 5 volts to nearly zero volts back to nearly 5 volts again (Ohm’s law) which, when measured from the ground, or zero volts, gives an output of zero volts, then 5 volts, then back to zero.

Waveform 2 shows the output of the first 555. Pins 2 and 5 are connected together as the input. When the input voltage is rising, once it reaches 2/3 of the supply voltage, pin 6 causes the output to go to zero. When the input is falling, once it reaches 1/3, pin 2 causes the output to go high to the supply voltage.

1

2

3

4

Thus a relatively slowly varying voltage is converted into square wave which is either on or off but nothing in between. These are the sorts of waveforms digital computers like!

Next the output is passed on to the second 555 via the 0.01 micro Farad capacitor and the 4.7k resistor network. Capacitors can only pass voltage changes not steady voltages hence the negative-going spike. (There would be a corresponding positive-going spike but the diode shorts it out (clamps it to +5 volts). The negative spike connects to pin 2 (“trigger”) and this starts a pulse the length of which is determined by the 100k and 0.1 micro Farad capacitor (11 milli seconds). This pulse is fed to an input of the PICAXE to be counted and should provide an unambiguous signal avoiding double counting or missing pulses.

I love the 555 which I use at every opportunity! It’s probably because it’s a bit complicated but not so complicated that I can’t understand all its little foibles, with a bit of effort! (A bit like the BBC ‘B’ computer - modern PCs and even microcomputers like the Arduino are too complicated for me to feel like I’ve got anything like a complete understanding. This also includes programming languages like C, Python etc. I suppose you have to content yourself with just knowing what works. This is a sort of epistemological problem - how much can we ever know? To digress for a moment, one of my hobby horses involves whether the physical world (assuming it actually exists!) can be described by laws expressed in mathematical terms. Most people including Brian Cox and Stephen Hawking say yes. I say the Universe just “is” and we invent a language called mathematics to correspond as closely as possible to how we experience the Universe through our senses. So there! (However, I’ve been wrong before!)



Software

'counter for rain gauge

pins = 0

b2= 0


main:

if pin2 = 1 then add

goto main


add:

b2 = b2 + 1

pins = b2

pause 200

goto main


Left is the very simple program for the PICAXE written on the editor provided by Revolution Education, the suppliers of the PICAXE. “pins” is a byte variable whose bits are mapped to the eight output pins of the PICAXE. “b2” is a built-in 8-bit byte variable.

PICAXE programming revives the good old “goto” that we used to know and love before structured programming (thanks to Mr Dijkstra).