Introduction

This circuit is a battery characterizer that applies a fixed load to a charged NiCad or NiMH cell and measures milliamp-hour capacity as it discharges.

There is no single load that can be applied to a battery to measure milliamp-hours. It depends on the battery type, it's intended application, etc. This device is simply one methodology that gives me a relative idea of the performance of NiCad and NiMH batteries in my drawer. Discharge is approximately 200mA which is approximately 0.1C for AA size cells or packs constructed with AA size cells.

When NuMH¹ batteries come into being, I'll update this device to handle those too.

Click on photo to display full size image.

Features

• Up to five cells can be characterized at one time
• To avoid total discharge the load is disconnected when test stops
• Determines internal resistance and accumulated milliamp-hours.

Specifications

• Power: 8-16VDC at ~100mA
• Sampling: once per minute
• Load 0.1C (~200mA)

Operation

Connect an 8-16VDC power supply to the battery characterizer.

Insert the cell(s) or connect the battery pack to be tested. After a charged cell is inserted in a socket, the unit waits for the voltage to stabilize (to avoid false measurements during first contact,) then the internal resistance is measured for later display when the cell test is complete. Thereafter, the cell is placed in TEST mode, the load is turned on, and milliamp-hours are accumulated. When the cell voltage falls below the cutoff voltage (1.0V per cell,) TEST mode is terminated and the load is turned off. DONE mode thereafter displays the total milliamp-hours and starting internal resistance until the cell is removed from the socket.

Socket #5 is designed for multiple-cell packs. The number of cells is calculated from the initial measured charged voltage and is in turn used to calculate the proper cutoff voltage by dividing by the nominal 1.2V per-cell voltage.

TEST screen

The TEST screen displays "T" and the cell number and it's accumulated milliamp-hours.

T#2 569maH     

DONE screen

The DONE screen displays "D" and the cell number, the total accumulated milliamp-hours (rounded to the nearest 100,) and the internal resistance (rounded to the nearest tenth) of the charged cell at the beginning of the test.

D#2 2400maH 0.1Ω

DETAIL screens

The DETAIL screens display various measured and calculated parameters during TEST and DONE cycles. This set of screens is invoked by pressing the button for longer than a half second. The various screens sequentially appear for two-seconds each. DETAIL mode is ended when the complete set of screens has been displayed, or if the button is pressed momentarily.

During TEST mode

In DONE mode

Cell#2: in TEST  

Cell#2: is DONE  

Num cells: 1     

Num cells: 1     

Init Vopen: 1.34 

Init Vopen: 1.34 

Vloaded: 1.21    

Int res: 0.14Ω   

Avg load: 246mA  

Avg load: 246mA  

Accum: 2176mAh   

Final: 2423mAh   

Time 3.2 hrs     

Time 3.5 hrs     

If all sockets are empty, the display shows a message to that effect.

Sockets empty  

Operational notes

• It's assumed that the battery pack to be tested is fully charged.
• The pushbutton cycles the display through the status screen of each cell. If the pushbutton is not pressed for ten seconds or more, the display will cycle every two seconds to the next cell in succession. Empty sockets are skipped. A cell that measures less than 0.4V is interpreted as an empty socket.
• The internal resistive load is disconnected when the battery discharges to the cut-off voltage (1.0V.) This prevents discharging the cell down to zero volts. For a multi-cell pack this precludes the possibility of reverse biasing and damaging a weak cell in the pack.
• Repeated tests of the same battery can yield maH results that vary 20% or more...it's highly dependent on actual charge.
• If the button is held down for more than four seconds, the system will go into reset mode, just as if the power were first turned on.

Technical Reference

HARDWARE

• The PIC is configured for power-on delay and brownout reset to guarantee proper resetting of the PIC.
• R1 and R2 form a voltage divider for testing battery packs of multiple cells. Whatever the divider resistors are, they should be several tens of kilohms, divide the voltage down to 5.0V (max,) and the ratio should be inserted into the C code before compiling.
• If using my preprogrammed PIC without specifying specific values for R1, R2, and Vdd, since I have very specific values for R1 and R2 in my own unit and you undoubtedly won't have my exact measured values, do the following...
  1. Do this instead...use a 10K fixed resistor for R2. This will not only divide, but also limit current into PIC pin 7 if the voltage on that socket is from a high voltage cell of some sort. Use a 10K multi-turn pot for R1.
  2. Once built, you can calibrate the socket five voltage to read accurately on the screen.
  3. Since the 5V power supply is used as the Vreference for the PIC's ADC, if your power supply is something other than that, the voltage readings will be off a bit and won't match what you'd read on your DVM. Try to find a voltage regulator that gets it quite close to 5.00V.
• Channels 1-4 have 100K resistors across the battery socket. That prevents the ADC inputs from floating and forces empty sockets to zero volts.
• Display contrast is controlled by R3 and R4, or insert a 10K pot instead.
• Loading is done by the 5-ohm 1/2W and 20-ohm 1W metal film resistors. These values were selected to discharge at approximately 250mA.
  
  
  Click on photo to display full size image.
  
  
• For best accuracy, measure the resistors and insert their real values into the software. Measure the voltage drop across the FETs and calculate their resistance and insert that into the software.
• Also for best accuracy, ensure that the VDD in the software matches the measured VDD in the hardware. Otherwise voltages will not be displayed accurately. This is because VDD is used as the reference voltage for the ADC in the PIC.
• If you don't want to deal with actual values for R1 and R2, simply make R1 a 10K pot and R2 a 10K fixed resistor. Adjust the pot to make socket #5 voltage read correctly.
• The 1-line, 16-character LCD display I used is made by Optrex obtained from MPJA. Similar LCDs are sold by DigiKey and Crystalfontz, listed in my main PIC page.
  
  
  Click on photo to display full size image.
  
  

SOFTWARE

• Basic sequence: The main loop code sets up the display, ADC, clock, and variables after which occurs an endless loop of checking switches, sampling, and displaying to the LCD screen. Timing and switches are handled by a 31mS interrupt. State is controlled by a simple state machine.
• During test, the cut-off voltage must be reached for a full minute before testing is automatically stopped. This avoids early stops due to electrical noise at the ADC input or due to poor connections to the battery pack.
• The per-cell cutoff voltage is set by a #define statement in the C code. For multi-cell battery packs, this cutoff voltage is further multiplied by the number of cells, which is determined by dividing the measured voltage by the nominal cell voltage (also a #define statement.)
• Actual load resistor values, FET resistance value, and other key parameters may be inserted at the top of the code.

// crystal frequency (Hz)
#define CRYSTAL_FREQ         4000000
// actual power supply voltage (volts)
#define VDD                  5.02
// cell nominal voltage (volts)
#define CELL_NOM_VOLTAGE     1.2
// voltage at which mAh accumlation stops (volts)
#define CUTOFF_VOLTAGE       1.0
// actual load resistances (ohms)
#define LOAD_0_OHMS          4.9
#define LOAD_1_OHMS          4.9
#define LOAD_2_OHMS          4.9
#define LOAD_3_OHMS          4.9
#define LOAD_4_OHMS          19.5
// actual FET RDS(on) resistance (ohms)
#define FET_RDS_OHMS         0.17
// actual scaling resistors R1 and R2 for channel 5 (ohms)
#define R1                   5350
#define R2                   10220
// display time for logo screens (normally 750 mS)
#define LOGO_TIME            750
// wait time for which voltage is displayed after a cell is inserted into a socket (seconds)
#define INITIAL_DELAY        2
// wait time after which display goes into rolling status mode (seconds)
#define STATUS_DELAY         10
// screen cycle time when in rolling status mode (seconds)
#define SCREEN_CYCLE_DELAY   2
// screen cycle time when in cell detail mode (mS)
#define DETAIL_CYCLE_DELAY   1500
// restart IRQ timer at adjusted value for 1-second accuracy (0-255, higher makes shorter IRQ tick)
#define IRQ_RESTART_TICK     3
// time at cutoff voltage required to end test (seconds, normally 60)
#define AUTOSTOP_TIME        60
// button time after which unit will display cell details (31mS interrupt counts, normally 16, ~= 0.5 second)
#define BUTTON_DETAIL_TIME   16
// button time after which unit will reset (31mS interrupt counts, normally 128, ~= 4 seconds)
#define BUTTON_RESET_TIME    128
// round to nearest 10's or 100's
#define ROUNDNUM             100
// socket definitions
#define SOCKET_1             0
#define SOCKET_5             4
// overvoltage definitions
#define CH_1_4_OVERVOLTAGE   2.0
#define CH_5_OVERVOLTAGE     6.0

Footnotes

1 - "Nuclear Metal Hydride"