Battery Life Test Project
2012-01-25 21:45 By Jason Birch
Using a PIC circuit board project to test how much power is provided in a variaty of PP3 batteries.
The background of this project is to experiment with the PIC16F630 single chip microcontroller. This microcontroller provides everything on a single intergrated circuit required to write programs which are capable of controlling external cuircuitry and devices. The PIC series of microcontrollers provide various packages which can be selected as most appropreate to the task required.
The PIC controller was selected for the following reasons for this project:
- It is one of the cheapest in the PIC range.
- It is compatible with the Vellaman PIC programmer.
- It provides an adequate number of I/O pins.
This PIC device provides the following features:
- 1KB EPROM program memory.
- 64 byte RAM.
- 128 byte EEPROM non-volatile memory.
- 12 I/O pins available as two 6 bit ports.
Program Features:
- Boot Display
- LCD Test
- Date / Time
- Set Date / Time
- Zero Date / Time
- Save / Restore Date / Time
- Low Power Mode
- Sound
- Play Tune
- Keypad
- Random Number Generator
- Save / Restore Random Seed
Circuit Features
- 20 x 4 LCD Display
- LCD Backlight low power mode
- Pizo electric sound
- 4 x 4 Keypad
- 5V Power regulation
- 8bit Output port
- 8bit Input port
- Low power circuit mode
Fitting a simple program into 1KB of memory is easy. However when controlling external devices and writing a proper structured set of routines which can be reused in future projects becomes more tricky. For this project the maximum number of features where implemented as were possible. This involved reviewing code and refining processes and structure several times in order to free up a few extra bytes of program memory.
The microcontroller drives the LCD display directly from it's I/O pins, the LCD has an enable signal. In order for the keypad to opperate an 8bit three state buffer is used, this provides the facility for the keypad to have an enable signal so it may use the same I/O pins as the LCD display. There are two more 8bit three state buffers in the circuit, one for input and one for output. In order to provide enough select signals for the devices, a 2-4 decoder has been used.
The keypad is a 4x4 matrix of buttons assigned the following values in a software lookup table:
|
7
|
8
|
9
|
D
|
|
4
|
5
|
6
|
C
|
|
1
|
2
|
3
|
B
|
|
#
|
0
|
*
|
A
|
One bit of the output buffer drives a piezo electric sounder via the transistor TR2. Variable resistor VR2 provides a volume adjustment for the sound, however this is not ideal for a piezo sounder and would be more appropriate if a speaker were used.
The program starts with a boot up display showing the author, application name and version information. This stays displayed for a few seconds, along with a boot up sound similar to a BBC micro or Apple computer.
The application has a power saving feature which will automatically dim the LCD back light and place the circuit into low power after a period of no user activity for approximately one minute. Below is a picture of the LCD in high power mode.
For this project only the display is placed in low power mode and the circuit is not powered down, consuming 56mA. This is so a battery is drained after a few hours, which provides enough definition in the results to distinguish power consumption from battery to battery. Below is a picture of the LCD display in low power mode.
The application clears low power mode when user activity occurs on the keypad or a sound is played. The LCD back light is switched off using transistor TR1, there is a resister bypassing TR1 to allow a dim back light when in low power mode. The variable resistor VR1 provides a contrast control for the LCD display.
The application provides a selectable mode the user can select using the '#', '*' and 'A' keys of the keypad. The modes available are LCD Test, Sound Test and Zero Date/Time. The LCD test displays the LCD's character set on the LCD as a series of pages which automatically cycle though. The sound test plays a tune and zero date/time resets the date/time and writes the zeroed value to EEPROM. The date/time value automatically saves itself to EEPROM every minute so that when the system boots, the last date/time value is restored from EEPROM and continues. The date/time value is displayed on the fourth line of the display.
When a key on the keypad is pressed, a set of values are displayed. The key code looked up in the key table. The random number seed is displayed, this is generated from user interaction and is saved to EEPROM and then restored at system boot time so that each system boot provides unique random numbers. The random number high byte and low byte is displayed. And also the microcontroller's calibration value, which is a value for calibrating the microcontroller's timers when using an internally generated clock source.
The date/time can be set by the user. Pressing the '5' key places the application in date/time set mode. Then pressing '4' and '6' to move a cursor left and right to select the digit to set. And then '8' and '2' to increase and decrease the digit selected. Finally pressing '5' again to bring the application out of date/time set mode.
The Project
As the microcontroller is using an internal clock a correction value was measured over the period of an hour to find the difference between a true clock and the applications clock. This was found to be x1.154 and checked over several measurements to verify the value. An external crystal to provide true time was not used for this project as all I/O pins has already been allocated to other tasks.
Towards the end of each batteries charge the current consumption will fall from 56mA until the current is on longer capable of powering the microcontroller. This means the discharge values measured here will be slightly higher than the actual value for the battery. But the cutoff point will be the same for all batteries tested so the comparisons here are very accurate.
Zinc composites perform differently to Alkaline composites. Though some of the duration values between the composites are similar, as they will power the microcontroller for a similar period of time. Zinc composites are not capable of providing a rate of power, to power the display for a long period of time.
Each battery has been purchased brand new. The test is only carried out on a sample of one battery per result published. So it is possible that a source of a battery supplied, provided a battery which is not representative of a typical battery capacity for the battery when new.
The purpose of this project was to discover from all of the many manufactures, types and claims by the manufacturers; which batteries are the best value for money and which last the longest in use.
The results measured are tabulated on the right side of this article and arranged in order of value of the power provided by each battery. The results are also graphed below comparing duration and value for money.
Conclusions
The results measured where not only surprising, but shocking. The brand associated as being a premium brand of battery not only performed the worst, also being the worst value for money in the batteries selected for testing. It also demonstrated the premium products of the premium brand performed no better than the standard product provided by the premium brand. This was so surprising that the first results found for this brand where considered inaccurate for some reason, and retested with batteries supplied from a leading high street store. These also performed very poorly in comparison to other battery brands tested.
Zinc based batteries provide a similar lifespan to Alkaline based batteries, but are only suitable for low current applications. With the price of high performing Alkaline batteries being very low, there is little point in purchasing Zinc based batteries. For example the Kodak XtraLife battery was one of the highest performers and when purchased from the high street store PoundLand and provided outstanding value for money.
NOTE: Battery manufactures may alter their materials, manufacturing processes or technologies over time. These results are only accurate for the batteries actually tested at the date they were tested.
|
