Current Sensor ADC scaling.

Rob_975 · Joined: 14 Aug 2012 Posts: 16

Hi all.

I'm using an ACS713 as current sensor with a 18F67k22 @ 12bit res.
I don't have any clue about how to calculate the scaling factor to get the right Ampere reading.
As per specs. this sensor has a 0.5V output @ 0Amps.
I tryed some kind of formulas that looked "barely" correct to but the outcome was completely wrong.

Cheers.

alan · Joined: 12 Nov 2012 Posts: 350 Location: South Africa

Not that difficult. First work out the offset, 0.5v *4096/5v = 410. Now this get deducted from any ADC value you measure. If your current are 10A then the ADC value will be 0.5V + 10A * 185mV/A (for the 20A version) = 2.35V or 1925 digital. Less your offset of 410 gives 1515. Convert back to amps 1515 * 5V/4096 / 0.185= 9.997 Amps. Use this with scaled integers discussed on this forum a couple of times and you don't have to use floats.

Regards

asmboy · Joined: 20 Nov 2007 Posts: 2128 Location: albany ny

A bit of analog processing to zero the offset AND expand the input
ADC range of the PIC-- to your expected current range --
will extract the maximum usability from your design too.

Subtracting a presumed zero-I offset and lack of analog span control
actually makes the LEAST of the 12 bits you have.

Being able to apply 4095 steps to the expected range, and fully utilize
the ADC will be the best outcome, neh ??
SO....
what is the expected RANGE of current you wish to measure?

Rob_975 · Joined: 14 Aug 2012 Posts: 16

OK, thank you very much guys.
Problem solved by means of Alan's formula and a few of small adjustments because of the components tolerances.
ASMboy, cant change the design, I already have the PCB with all the components on it.
It's a speriemental board that I designed to make experiements.
The ACS goes to a opamp buffer and then to the Pic's ADC.
The voltage reference is a 4.096V ref. generator.
The maximum current I would like to measure is 10A.
With the code I'm gonna share with you, the error of the system is 0.01A along the range 0-2A. My psu cant supply more than that on a load of 8ohm.

asmboy · Joined: 20 Nov 2007 Posts: 2128 Location: albany ny

i sure hope the VAR 'current' is an int32

gpsmikey · Joined: 16 Nov 2010 Posts: 588 Location: Kirkland, WA

If you are only looking at one channel, you can set the adc channel outside the loop which eliminates that and the following delay from the inner loop.

mikey
_________________
mikey
-- you can't have too many gadgets or too much disk space !
old engineering saying: 1+1 = 3 for sufficiently large values of 1 or small values of 3

Rob_975 · Joined: 14 Aug 2012 Posts: 16

Rob_975 · Joined: 14 Aug 2012 Posts: 16

asmboy · Joined: 20 Nov 2007 Posts: 2128 Location: albany ny

"No, it's an float. "

since you are always adding INTEGER ADC readings
in that variable -
make it an INT32 as FLOAT takes MANY more machine cycles
to to an "add" than an integer32 .....

only convert to a FLOAT if you need to do decimation fractional output

ruth · Joined: 14 Apr 2014 Posts: 11

Alan, i had read your reply for the post of Current Sensor ADC scaling. Can u explain aboutt the "Less your offset of 410 gives 1515." Why 410 will give 1515? I am using LTS 15NP current sensor and i am trying to get back the ampere from the voltage output of the sensor. Thank for your concern Smile

alan · Joined: 12 Nov 2012 Posts: 350 Location: South Africa

The LTS 15NP outputs 2.5V @ 0Amp, that will be your offset.
If current positive it will add 41.6mV for each ampere that flow through the sensor, similarly will subtract 41.6mV for each ampere.

So work out what your ADC reading will be @ 2.5V and that is what I call the offset, it is fixed throughout your calculations.

Meaning if you read 2.5416V (Nice ADC this Smile

) work out what your ADC reading will be. If you subtract the offset from this it will give you a result 41.6mV in digital format (the value depends on the resolution of your ADC) which equals 1A.

Regards