For one of my projects I want to use an arduino to measure voltage for from 20 sensors.
To do that I saw many shield that allow to multiplex voltage source connected to a few arduino. ADCs ... But It's for fun, so I decide to build my onwn including standalone AT328 (see this Arduino link).
Requirements
To begin this is my requirements list :- max 20 sensors inputs. Max voltage Vref /AA.
- work on battery so as low consumption as I could
- wireless communication (as cheap as I could)
- one raspberry pi as server (receive/ persist datas comming from my 328)
- proof of concept/validation (is all is working, communication, sensors and so on)
- optimization / validation (reduce consumption, resolve dysfunction)
- optimization / validation
- end of projet : build circuit board, solder, put in box , validation
Proof of concept
At this point I have to :- choose components
- prototype this card
- and when all will work I will thinl about circuit board.
Component list
- atmel 328 with bootloader (It more expensive as ATmel without bootloader, but one thing at a time)
- 16 Mhz quartz
- few resistor and capacitor
- L7805CV regulator. (I will try LM317 latter)
- emitter / receiver 433MHz (Warning you should verify if the 433 Mhz use is allowed in your country...)
- 2 x CD4051 8 channels multiplexer (for test it's enough)
- raspberry pi (I have a B model with wifi)
- battery connector
For prototyping
- breadboard (I have 2 : 840 points)
- 1 usb / rs232 converter (to upload sketch in arduino)
- I have one arduino (will allow me to make comparison)
Schematics
TODO/Work in progresslet's start building
first build an arduino like (see "Building an Arduino on a Breadboard"), at this point I could not test anything except Vcc.Then add an USB to Serial arduino module to be able to push sketch in arduino.
To connect this module you have to connect
- +Vcc,
- Gnd,
- Tx (to atmel328 Rx)
- Rx (to atmel328 Tx)
- Reset (to atmel 328 pin 1 reset using a 100nF capacitor)
At this point I could test my arduino and usb connection.
I use the blink example and connect a LED to pin arduino pin 13. Then push this sketch to my arduino bread board. And... all is ok.
Now it's time to add wireless comunication. I choose a 433Mhz emitter end receiver (very cheap)
In my plan, in want an arduino send periodically information to my raspberry pi. In my mind, at the end I'd like to build a "server ask --> client response" and see if it decrease consumption.
So I had now to connect an emmiter to arduino and the receiver to my raspberry pi, and see if all work toghether.
First I'll start with receiver on raspberry pi.
As I found a lot of informations about internet I saw there difference between RPI revision and GPIO.What is my RPI model ?
So first I have to identify which RPI model I have.So I connect the RPI through SSH using moba xterm. And in a terminal I use the command line :
cat /proc/cpuinfo
at the and of the command output you should see something like "Revision : Number"
Go through http://www.magdiblog.fr/divers/connaitre-le-modele-exacte-dun-pi/
and look at your model.
mine is : 000e Model B Revision 2.0 512MB, (Sony)
Ok , let's connect RPI and arduino
see http://www.homautomation.org/2013/09/21/433mhtz-rf-communication-between-arduino-and-raspberry-pi/Mainly :
Receiver on RPI- connect pin GPIO 21 (wiring pi 2; chip 13) on receiver data out
- GND to GND
- and 5v to vcc
- connect pin 10-PWM (physical 16) on emitter data in
- GND to GND
- and 5v to vcc
Software
For software on both I use RCswith (VirtualWire does not work for me...) see (https://code.google.com/p/rc-switch/)For RPI I use RFSniffer
on the arduino side the sketch is as in http://www.homautomation.org/2013/09/21/433mhtz-rf-communication-between-arduino-and-raspberry-pi/
And all is working...
Ok now,
- due to the nature itself of this protocol, message are received many times... Anyway, just modify the RFSniffer.cpp (as follow) relying on this source (see last program : http://faitmain.org/volume-1/dispositifs.html)
- I want to send more than activity form my arduino... I want to send many data, in fact one for each sensor connected... To do this I have to find or write some kind of protocol wich is allowing to :
- to identify my arduino card (with an adress)
- to identify on sensor (relying on multiplexer adress ??? 3 or 6 bit ie 8 or 16 sensor)
- to add for each sensor my digital measured value.
first, what is the max size of a "data packet" ? To determine this I firstly search but could not find. So I use rule of thumb : send bit since reaching the limit... I found that I could send 31 bits (31 + 1 bit for sync I think)
mySwitch.send(2147483647, 31);
so I choose
- 14 bit for base address (0 to 16383) identifier
- 5 bit for sensor identifier (32 sensors max for each arduino)
- 12 bit data (0 to 4095)
At this time on receiver side I could have :
while sending on arduino side :
mySwitch.send("1111111111000000000000000000100"); // base adress 16368 and sensor value 0
// and value 4
mySwitch.send("1111111110010000000000000001000"); // base adress 16356 sensor module 0
// and value 8
How to send data ?
Before adding multiplexer, I have decide how to send Sensor datas.What i already know :
- ADC conversion take 13ADC clock cycles
- ADC clock is Atmel frequency (16MHz in my case)/ 128 (prescale set toobtain between 50kHz and 200Khz ADC clock for 10bit resolution) , that gives 125KHz.
- So the conversion time is 13 cycles / 125KHz = 0.104 * 10^-3 that gives 104 uS
As I could read the ADC frequency could be increase up to 1Mhz without much degradation by changing the prescale.
In my case 1 sensor value send each minute is more than enough. it means
1 multiplexer adressing + 1 conversion time + 1 send (multiple send beacause of 433 Mhz protocole)
so it means
AAA us + 104 us + BBB us .....
I am missing data, let grasp them.
Multiplexing time
I could try to measure it when i will have it...Sending time duration
To have an idea I does :void loop() {
unsigned int i;
for(i=0;i<10;i++) {
start_times[i] = micros();
mySwitch.send("1111111111111100000000000000111");
delay(1000);
stop_times[i] = micros();
}
Serial.println("\n\n--- Results ---");
for(i=0;i<10;i++) {
Serial.print(" elapse = ");
Serial.print(stop_times[i] - start_times[i]);
Serial.print(" us\n");
}
That is giving me ~500ms for BBB.
Connecting multiplexer (CD 4051) and Mock sensor
Firstly, I don't want to already use sensor, so I will fake it :- Mock sensor : potentiometer gnd, +5V , middle pin to in multiplexer
connect multiplexer pin z (4051) to pin A0
connect the voltage sensor to pin Y0 (4051). As I fake only one entry I connect the Y1-Y7 to gnd.
And after adding some lines to my sketch I obtain on my Raspberry :
- an adress (in the future based on DIP switch)
- a sensor id
- the sensor value.
What if wifi card is cheaper...
A the very beginning I'd like to use Wifi shield to does exchange with my RPI, but the cost was unaffordable.Recently I found an ESP8266 device from 5$ to 8$ ... So I decide to test it. I'm waiting for delivery. After a few test I will probably change the 433Mhz emitter. This evolution will allow to exchange more data in a request-response way with no receiver to install.
I finally received my ESP8266... But I was a little bit optimistic and I missed a couple of important things :
- ESP8266 run at 3.3V
- ESP could required 300mA peak
For the connection, I found many link for example :
- http://yaab-arduino.blogspot.fr/2015/03/esp8266-wiring-schemas.html
- http://www.instructables.com/id/noobs-guide-to-ESP8266-with-Arduino-Mega-2560-or-U/step2/Software-needed/
So let's start and see if a zener regulation with BZXC55C3V3 500mW (with a serial resistor) is possible (see Zener Diode introduction from M. H. Miller or http://electronics.stackexchange.com/questions/28944/selecting-correct-zener-diode).
With this kind of regulation, the serial resistor is calculated with the max current in the "load" (here 300mA for ESP8266) and the minimum for zener work (I took 10mA) so 310mA. When no more load is required, the execess should be absorb be zener... when nothing is required from the load the ESP 8266 require low current. The wort case is 0 A , the datasheet gives 0,9 mA in standby mode.
So the zener should absorb 309,1mA .... far more than the maximum current (I= P/U) 151mA (Izrm 115mA)
This solution is not suitable.
So as I have LM317, let's use one....
To build a regulation with a LM317 you could use this schematic
and this formula
Vout=Vref(1+R1/R2)+IadjR2
$E=mc^2$
Vref = 1.25
Iadj = 50A
to be continued...
Next step :
- serve information collected : exposing data through REST API on raspberry PI
- Using arduino interrupt
- what if other 433Mhz peripherals exists ?.... One : how to check that.... Two :change ardess with dip switch WIP