Battery State of Charge - Prototype

[digitaldanny]

Describe the solution you'd like

• Add ADC connection from microcontroller to battery (and other HW if necessary).
• Basic SW to detect voltage of battery.
• Test that battery charge can be detected with this SW. Make sure there are enough ADC steps to read battery's discharge voltage range (may be from 3.3V to 4.2V).

Additional context Research for this issue was completed in #26.

[digitaldanny]

I will estimate the battery’s remaining charge or “State of Charge (SOC)” using a method called “Coulomb Counting.” This method involves estimating the state of charge by subtracting the current consumption from the initial capacity. The battery percentage can be estimated with the following formula:

(Q_previous - i_current deltaTime) / Q_total 100%

However, the error in this estimation will increase each time the battery is recharged. My goal is just to provide a rough estimation so the user knows to plug it in when it’s in the 10-25% range, so this will work well for my application.

An alternative that I considered is Open Circuit Voltage (OCV) estimation. With this approach, the battery charge can be estimated by comparing the battery’s open circuit voltage to the linear portion of the battery’s discharge curve. This would be simpler and more accurate; however, the battery would have to be an open circuit for each measurement. This would power off the microcontroller, disabling the necessary ADC pin along with the rest of the glove’s functionality.

[digitaldanny]

I switched to using the AD8226 instrumental amplifier instead of the INA114AP because it is a single supply rail-to-rail amp that can operate at 3.3V, which will work for my glove's 3.3V power source. I have the new in-amp hooked up on a breadboard to make sure it will sense voltage drops across the shunt resistor. My test is described below.

• Powered with 3.3V and 5V in two separate test runs.
• IN+ input = 3.3V offset. The first pin on the shunt resistor should see a consistent 3.3V power supply input.
• IN- input = 10 mV amplitude sine wave with 3.299V offset. At an estimated 60-100 mA current draw, the shunt resistor should drop no more than 10 mV across the resistor.

Powered with 5V (Working correctly)

Powered with 3.3V (Not functional)

Results:

• When I power the amp with 5V, the amplification works correctly. When I power witth 3.3V, the output stays at ~0V. When I drop the offset down to ~1V instead of 3.3V, I get the expected sine wave even with 3.3V power supply. I believe this means there is some headroom required for the input difference. Planning to dig into the datasheet a little more to find the headroom required.

[digitaldanny]

Current sensing circuit changed from original design with INA114AP in the first figure below to the following design with the AD8226ARZ. Adding some voltage dividers to halve the voltage before going into the in-amp fixed the problem I was seeing in the previous comment.

Original (INA114AP)

New (AD8226ARZ)

[digitaldanny]

Current circuits being used for Coulomb Counting included below.

• Coulumb Counting Current Sensing circuit will be used to read current draw while all hardware is powered on. This circuit halves the voltage drop across the shunt resistor, so the software will need to double the voltage conversion.
• Open Circuit Voltage Sensing circuit will be used to read voltage of the battery when external hardware is powered off and the uC is in low power mode.
• The uC-Controlled power switch is used to enable/disable external hardware power.