Low voltage precision current sensing (INA219?)

Hi everybody,

I am developing a custom Voltage/Current/total Power meter for a small 5V, 3W solar panel system. Goal is a ultra low power solution.
Curently I have barebone Arduino on 3,3V (powered from REG710 charge pump) and running at 8MHz from internal RC oscillator. Applied various power reductions and consumption is only 4mA in active state and 1mA in sleep state (including Nokia 5110 LCD).

As I found there are only 3 types of current sensors generally availible - ACS712(Hall sensor based), MAX471(analog) and INA219(analog with II2C).

Because Hall based sensors are really inaccurate at measuring curents below 500mA ( only 185mV per 1A ), I purchased a “simple” sensor based on MAX471. Results are not bad, but still very inaccurate. After manual calibration, and adding filters at input, output is accurate only at current bigger than 350mA. Currents below 150mA are not sensible at all.

Of course I can develop my own circuit with amplifier and custom shunt, but I have a feeling that it is kind of re-developing a wheel…

Now I am looking at II2C solution INA219, looks good and even have built in power counter. There is even Arduino library so I dont need to write custom code.

Anybody have some experience with this sensor? Or have another idea how to effectively achieve this king of current measuring?

INA219 - very interesting chip. It's new for me.

Available on a breakout board here.

Clones on ebay.

Has been used many times on this site.
Do a search in the searchbox on top of this page.
Leo..

Indeed I know that it is availiable and it was used many people before.

But I am looking for specifications of measurement accuracy in low currents(or comparison with other sensors), I havent found any info about it.

Everything I found was just about that it works (its like with MAX471, nowhere is mentioned that you cant sense less than 150mA and that solid precision is over 300mA minimum).

Galvanic (shunt) measurement should go down to zero (and negative) lineair.
One small problem could be zero Amp offset, but that can be fixed in software.
The INA219 has the advantage of an inbuild 12-bit A/D, PGA, and more.
Leo..

Wawa:
Galvanic (shunt) measurement should go down to zero (and negative) lineair.
One small problem could be zero Amp offset, but that can be fixed in software.

Yes, they should be linear.
But MAX471 is also shunt based (internal 35 mOhm shunt). But measuring offset is pretty high - as mentioned it can measure only above 150mA.

Wawa:
The INA219 has the advantage of an inbuild 12-bit A/D, PGA, and more.

This is exactly what I really like on this chip. I guess that custom configuring preset of PGA & Bus voltage range should highly increase accuracy in low currents measuring.

It looks that I have to buy (yet) another samples to test :smiley: