Electrical – Advantages and disadvantages of reading a current-sense amplifier with different gains on I2C

amplifieranalogcurrenti2c

I am writing an I2C wrapper for the MAX9611 chip. For the current-sense amplifier (CSA), according to the datasheet there are 3 selectable gains (1x, 4x and 8x):

MAX9611 internal diagram

When I read the 12bit ADC registers for all 3 gains, I get the same raw values. But to convert it to real current I have to multiply them with some constants which are different for each gain, e.g. for 1x it is 1.075 and for 4x/8x constants are 0.269 and 0.134. I have found these constants on some sample codes and I have no idea how they are calculated but I have found this table in datasheet but I can't make sure why:

enter image description here

So my guess is, when I select the gain=8x, the resolution (or accuracy?) of my reading is much higher compared with gain=1x but what will I lose in doing so?

So my question is, what are the advantages and disadvantages for reading from lower or higher gains? Also, what should be the correct multipliers if I want to convert raw ADC to real current? (the ADC is 12bit) I power the chip from a 3.3V source.

Best Answer

The different gains allow you to sense a wider range of currents. The higher the current, then the higher the gain you will need (although that sounds backwards).

The current-sense resistor is limited by the heat it can dissipate, usually just a couple of watts. So to sense higher current, you need the sense voltage to be lower (a smaller value sense resistor, like .005 ohm) to keep the wattage in range. A common maximum sense voltage would be around 250 millivolts for sensing around 10 amps or less, and more like 50 millivolts if you're sensing up to 50 amps.

For best resolution, you want to amplify the sensed voltage to maximize the range of the ADC, without saturating the ADC. So you will select your gain for your desired maximum sense voltage. There is no advantage or disadvantage to a particular gain - there is simply one gain that will work best for your maximum sense voltage.

Your second question: What are those silly multipliers? You have a 12-bit ADC (4096 values) that convert a voltage range of only 440 millivolts. So the "Full-Scale Sense Voltage" is simply the .44 volt ADC input range divided by the gain. The "LSB Step Size" is simply the voltage necessary to change the output by one.

Related Topic