Read analog data, in Java, part III

Important: This document goes along with the code at https://github.com/OlivierLD/raspberry-coffee/.
In the sources, refer to ADC

Analog to Digital Converter: Monitor a 12 Volt system

We have seen in the previous documents (available here) how to read and monitor a varying power source from the Raspberry PI using a chip MCP3008.
The power source was in those cases provided by the Raspberry PI itself. Those voltages range from 1.2 Volt (at the CPU level), to 5 Volts (power supply, pin #2 on the GPIO Header), the GPIO pins deliver a 3.3 Volts output.
The MCP3008 itself would not take any current greater than 5 Volts.
If we want to monitor a 12 Volt system (on a boat, an RV, a solar panel...), we will need to use some voltage divider, in order not to overload - and fry - the components we will use.
This divider will sit between the power source to monitor, and the pins CH0 to CH7 of the MCP3008.

The divider

In the figure above:

V_in is the 12 Volt system to monitor. It can go up to 15 Volts.
V_out is what will be used from the MCP3008 standpoint. The + terminal will go to its CH[0-7] pins.

V₁ is the tension across the resistor R₁.
V_out (which can also be named V₂) is the tension across the resistor R₂.

We (should) know that

V₁ = V_in ×	R₁	and	V₂ = V_in ×	R₂

	R₁ + R₂			R₁ + R₂

If we go with:

V_in = 15 V
R₁ = 100 kΩ
R₂ = 33 kΩ

we find out that V_out = 3.72 V, which is suitable for the MCP3008. We will go with those values.

Do your own:

Vin	V
R1	kΩ
R2	kΩ
Vout	3.72 V

The setup

Here is the full setup. Notice the following elements:

The wiring on the Raspberry PI, very close to what it was before, but with some differences regarding the hot wires
The potentiometer and the voltmeter, which will be used for calibration
The ground, common to the Raspberry PI and the divider.

All set with the hardware.

Calibration

A Note: This calibration step is not mandatory. Its goal is to display voltage values as close as possible to the reality. If you are only interested in seeing the way the voltage varies, you can skip it.

The MCP3008 returns values ranging from 0 to 1023 (2¹⁰ = 1024).
We need to calibrate the hardware so we can directly read the voltage from the program running on the Raspberry PI.
To do so, with the setup above, on the Raspberry PI, start the following script:

 
 Prompt> ./battery.monitor -calibration

Turn the potentiometer all the way counter-clockwise


  Prompt> ./battery.monitor -calibration
  Read an ADC
	Parameters are:
	  -calibration or -cal
	  -debug=y|n|yes|no|true|false - example -debug=y        (default is n)
	  -ch=[0-7]                    - example -ch=0           (default is 0)
	  -min=minADC:minVolt          - example -min=280:3.75   (default is    0:0.0)
	  -max=maxADC:maxVolt          - example -min=879:11.25  (default is 1023:15.0)
	  -tune=ADC:volt               - example -tune=973:12.6  (default is 1023:15.0)
	  -scale=y|n                   - example -scale=y        (default is n)
	  -log=[log-file-name]         - example -log=[batt.csv] (default is battery.log)
  Prms: ADC Channel:0 MinADC:0, MinVolt:0.0, MaxADC:1023, maxVolt:15.0
  readAdc:286 (0x11E, 0&100011110) Volume:27% (286) Volt:4.193548
  readAdc:293 (0x125, 0&100100101) Volume:28% (293) Volt:4.296188
  readAdc:283 (0x11B, 0&100011011) Volume:27% (283) Volt:4.14956
  readAdc:289 (0x121, 0&100100001) Volume:28% (289) Volt:4.2375364

Read the voltmeter: In this example, let's say that it displays 3.75 Volts, when the ADC value is at 289.
Now, turn the potentiometer all the way clockwise, and read the voltmeter:


  ...
  readAdc:606 (0x25E, 0&1001011110) Volume:59% (606) Volt:8.885631
  readAdc:654 (0x28E, 0&1010001110) Volume:63% (654) Volt:9.589443
  readAdc:717 (0x2CD, 0&1011001101) Volume:70% (717) Volt:10.513197
  readAdc:808 (0x328, 0&1100101000) Volume:78% (808) Volt:11.847507
  readAdc:872 (0x368, 0&1101101000) Volume:85% (872) Volt:12.785924

In this example, the voltmeter reads 11.75 Volts when the ADC says 872.
To summarize:

	ADC	Voltmeter
Mini	289	3.75
Maxi	872	11.75

Now, re-run the script as below. Notice the -min and -max parameters, and how they relate to the table above. Move the potentiometer back and forth.


  Prompt> ./battery.monitor -min=289:3.75 -max=872:11.75 -debug=y
  Read an ADC
	Parameters are:
	  -calibration or -cal
	  -debug=y|n|yes|no|true|false - example -debug=y        (default is n)
	  -ch=[0-7]                    - example -ch=0           (default is 0)
	  -min=minADC:minVolt          - example -min=280:3.75   (default is    0:0.0)
	  -max=maxADC:maxVolt          - example -min=879:11.25  (default is 1023:15.0)
	  -tune=ADC:volt               - example -tune=973:12.6  (default is 1023:15.0)
	  -scale=y|n                   - example -scale=y        (default is n)
	  -log=[log-file-name]         - example -log=[batt.csv] (default is battery.log)
  Prms: ADC Channel:0 MinADC:289, MinVolt:3.75, MaxADC:872, maxVolt:11.75
  readAdc:759 (0x2F7, 0&1011110111) Volume:74% (759) Volt:10.1994
  readAdc:746 (0x2EA, 0&1011101010) Volume:72% (746) Volt:10.021011
  readAdc:711 (0x2C7, 0&1011000111) Volume:69% (711) Volt:9.540737
  readAdc:667 (0x29B, 0&1010011011) Volume:65% (667) Volt:8.936964
  readAdc:642 (0x282, 0&1010000010) Volume:62% (642) Volt:8.59391
  readAdc:617 (0x269, 0&1001101001) Volume:60% (617) Volt:8.250858
  readAdc:601 (0x259, 0&1001011001) Volume:58% (601) Volt:8.031303
  readAdc:593 (0x251, 0&1001010001) Volume:57% (593) Volt:7.9215264
  readAdc:609 (0x261, 0&1001100001) Volume:59% (609) Volt:8.141081
  readAdc:625 (0x271, 0&1001110001) Volume:61% (625) Volt:8.360635
  readAdc:647 (0x287, 0&1010000111) Volume:63% (647) Volt:8.662521
  readAdc:681 (0x2A9, 0&1010101001) Volume:66% (681) Volt:9.129074
  readAdc:716 (0x2CC, 0&1011001100) Volume:69% (716) Volt:9.609348
  readAdc:754 (0x2F2, 0&1011110010) Volume:73% (754) Volt:10.130789
  readAdc:794 (0x31A, 0&1100011010) Volume:77% (794) Volt:10.679674
  readAdc:842 (0x34A, 0&1101001010) Volume:82% (842) Volt:11.338336
  readAdc:862 (0x35E, 0&1101011110) Volume:84% (862) Volt:11.612779
  readAdc:808 (0x328, 0&1100101000) Volume:78% (808) Volt:10.871784
  readAdc:756 (0x2F4, 0&1011110100) Volume:73% (756) Volt:10.158234
  readAdc:716 (0x2CC, 0&1011001100) Volume:69% (716) Volt:9.609348
  readAdc:680 (0x2A8, 0&1010101000) Volume:66% (680) Volt:9.115352
  readAdc:648 (0x288, 0&1010001000) Volume:63% (648) Volt:8.676244
  readAdc:629 (0x275, 0&1001110101) Volume:61% (629) Volt:8.415524
  readAdc:615 (0x267, 0&1001100111) Volume:60% (615) Volt:8.223413
  readAdc:602 (0x25A, 0&1001011010) Volume:58% (602) Volt:8.045026
  readAdc:586 (0x24A, 0&1001001010) Volume:57% (586) Volt:7.825472
  readAdc:580 (0x244, 0&1001000100) Volume:56% (580) Volt:7.7431393
  readAdc:587 (0x24B, 0&1001001011) Volume:57% (587) Volt:7.839194
  readAdc:605 (0x25D, 0&1001011101) Volume:59% (605) Volt:8.086192
  readAdc:630 (0x276, 0&1001110110) Volume:61% (630) Volt:8.429245
  readAdc:669 (0x29D, 0&1010011101) Volume:65% (669) Volt:8.964409

The voltage displayed on the screen should match what's read on the voltmeter.

Two calibration options

You can use 2 points as above, or one only.
In the case you use only one, we will assume that for 0 ADC, we have 0 Volts.
Let us see.
If we use the following calibration parameters:


	 -min=695:9.1 -max=973:12.6 -scale=y

The voltage ranges from -0.3500028 V for ADC=0, to 13.265468 V for ADC=1023. This produces the following figure.

If we use:


    -tune=973:12.6 -scale=y

The voltage ranges from 0.0 V to 13.247482 V for ADC=1023, and produces the following figure.

This seems good enough.

Operation

Run the script without the -debug parameter.


  Prompt>./battery.monitor -tune=973:12.6
  Read an ADC
	Parameters are:
	  -calibration or -cal
	  -debug=y|n|yes|no|true|false - example -debug=y        (default is n)
	  -ch=[0-7]                    - example -ch=0           (default is 0)
	  -min=minADC:minVolt          - example -min=280:3.75   (default is    0:0.0)
	  -max=maxADC:maxVolt          - example -min=879:11.25  (default is 1023:15.0)
	  -tune=ADC:volt               - example -tune=973:12.6  (default is 1023:15.0)
	  -scale=y|n                   - example -scale=y        (default is n)
	  -log=[log-file-name]         - example -log=[batt.csv] (default is battery.log)
  Prms: ADC Channel:0, tuningADC:973, tuningVolt:12.6

    ....

  ^C
  Shutting down
  Closing log file
  Shutting down...
  Prompt>

Let it run as long as you want, and stop it with a SIGTERM (Ctrl+C).
The data are logged in the file named battery.log, as CSV data. Columns are date, ADC, Volts.


  Prompt> tail battery.log
  2014-04-19 11:45:39.804;523;6.9609776
  2014-04-19 11:45:39.907;558;7.441252
  2014-04-19 11:45:40.009;589;7.866638
  2014-04-19 11:45:40.112;636;8.511578
  2014-04-19 11:45:40.229;681;9.129074
  2014-04-19 11:45:40.331;736;9.883791
  2014-04-19 11:45:40.439;788;10.597342
  2014-04-19 11:45:40.542;850;11.448113
  2014-04-19 11:45:40.645;875;11.791166
  2014-04-19 11:45:41.455;869;11.708834
  Prompt>

Slice of Pi

Once the prototyping phase is validated, all the components are set on a Slice of PI.

The components on the board

Wired. At work.

One terminal is used to hook up the 12v system to monitor, the other is used for calibration, with the voltmeter.

Logging and rendering

The logging is performed as above.

At work.

The data in red are raw data, the blue curve is the result of a smoothing process.
The big spike downward is the starter of the engine...

Version 2

After a couple of tests, there was some room for improvement, on the software side mostly.
On the hardware side though, a smaller potentiometer considerably reduced the height of the device, and it fits in the chart table.
Also, it could be a good idea to leave some possibility to use some I²C board. As the Slice of PI uses the full GPIO header, four wires have been soldered (GND, 3V3, SCL, SDA), they could be plugged onto a breadboard to use breakout boards like a BMP180, for example.

In the chart table. Smaller potentiometer, an I²C extension. NMEA Logging.

And NMEA multiplexing can be used as well. There is - as far as I know - no NMEA String for the battery tension..., I just came up with my own. From there, regular NMEA logging can be used:


	$IIMTW,+14.5,C*38
	$IIMWV,151,R,04.4,N,A*16
	$IIMWV,150,T,04.5,N,A*10
	$IIRMC,030143,A,3730.078,N,12228.856,W,00.0,148,230514,15,E,A*1B
	$IIVHW,,,359,M,00.0,N,,*6B
	$XXBAT,14.16,V,966,94*19
	$IIVLW,08189,N,000.0,N*5B
	$IIVWR,151,R,04.4,N,,,,*64
	$IIGLL,3730.078,N,12228.856,W,030143,A,A*4F
	$IIHDG,359,,,15,E*19
	$IIMTW,+14.5,C*38
	$IIMWV,152,R,05.1,N,A*11
	$IIMWV,151,T,04.4,N,A*10
	$IIRMB,A,0.23,R,,HMB-3   ,,,,,001.20,184,,V,A*1F
	$IIRMC,030145,A,3730.078,N,12228.857,W,00.0,148,230514,15,E,A*1C
	$IIVHW,,,359,M,00.0,N,,*6B
	$IIVLW,08189,N,000.0,N*5B
	$IIVWR,152,R,05.1,N,,,,*63
	$IIGLL,3730.078,N,12228.857,W,030145,A,A*48
	$IIHDG,000,,,15,E*16
	$IIMTW,+14.5,C*38
	$IIMWV,150,R,05.0,N,A*12
	$IIMWV,152,T,05.1,N,A*17
	$IIRMC,030145,A,3730.078,N,12228.857,W,00.0,148,230514,15,E,A*1C

From there, it's all downhill, data can be analyzed...

For real...

The downward notches correspond to the periods when the fridge is on. The global tension goes up with the sun (solar panel).
The white and gray backgrounds show night and daylight, enhancing the influence of the solar panel.
And as we log all the NMEA data, we can also tell what the wind looks like, and as a result, what would be the contribution of the Wind Generator.

The wind...

All set, Raspberry PI B, USB Hub, and a breadboard.
The connectors plugged on the breadboard are I²C ones.
I used it with a BMP180 (Air temperature & Pressure).

Note: Since this project begun, the Raspberry PI B+ has been released. As it has 4 USB ports, you would not need the USB Hub featured above. This would allow a smaller setting, or a bigger breadboard.

The budget

As of April 2014, here is what it takes:

A Slice of PI	$ 9.05	from Amazon
A 50K potentiometer	$ 0.95	from Adafruit
An MCP3008	$ 3.75	from Adafruit
2 connector terminals	5 for $ 2.95	from Adafruit
Resistors, wires, etc
Total	$ 16.70

Schema

Fritzing schema available here.

Oliv did it