INA828 monitoring current shunt charging and discharging question

Question

I am designing a battery current shunt monitor that measures current while discharging and also charging. I am planning on using the INA828 amplifier feeding an ADC.
My questions revolve around the Ref pin connection of the INA and whether I need a +/- supply? On attached diagram, in lower right corner are my two questions;

1. Do I need +/- supply to the INA to measure the +/- input differential voltage in both directions?
2. Do I need to provide an offset voltage to the Ref pin of the INA to convert the +/-50mV input to 0 - 2.048V output? (The delta-sigma ADC has internal 2.048V Ref that I intend to use during ADC conversion). The ADC sampling rate will be 20hz (10k+ oversampling ratio) with Sinc4 digital filter.

The shunt is not pushed hard. It's a 500A shunt. +150A is the max charging current available. -300A is the max load current that can reasonably be realized. The INA gain is chosen for shunt full range. This is a low side shunt.

I know there is a bunch to consider beyond my two questions. For now, the two questions are where I am currently stuck.

Answer 1

You don't need a dual supply to operate this chip. However, the 5V single-ended supply is quite low for your requirements - this amplifier is not "rail to rail" inputs. In other words you are out of headroom.

You could use the 12V from the battery to power the amplifier - this solves your input problem. But, you still have an issue with the output. You are going to need a buffered reference voltage for the low-impendence reference pin. This is normally created from a precision bandgap reference (6V, 5.1V, or something that's in the middle of the 0V-12V range) fed into a unity-gain amplifier. This band gap also serves as the reference for the ADC. Then you have to scale (resistor divider) the output of the instrumentation amplifier to be in the middle of the ADC range. This is so you can measure both charging and discharging current.

Personally, I would use something like this: https://www.ti.com/lit/ds/symlink/ina185.pdf?ts=1704224694708&ref_url=https%253A%252F%252Fwww.ti.com%252Fproduct%252FINA185

It's a high-side bi-directional shunt amplifier. It operates on 5V, covers your input common-mode range, designed for exactly what you are trying to do, and comes in a tiny SOT-563 package. You will still need to offset the reference using this approach if you want bidirectional readings. Or, use 2 devices with a common shunt but with flipped inputs, and use 2 ADC inputs: 1 for charging and 1 for discharging.

Answer 2

You need to look at the common-mode range requirement in the data sheet.

This shows that correct operation is only guaranteed when the inputs are no closer than two volts of the supply rails - so yes a negative supply of at least two volts is required for your application.

Datasheet

Answer 3

The common mode question was answered. Other feedback:

-You don’t need a negative supply if you choose a different amplifier

-Generally an Inst. Amp is not the best choice for current sensing because you don’t need high impedance inputs. Many dedicated shunt amplifiers exist or you can make a difference amplifier with resistors (see matched packs ACAS or LT5400) and any regular opamp (common mode includes ground).

-Yes you want to bias the reference to the ADC midpoint (assuming you want to measure symmetric +- currents).

-If you do use an inst amp I’d probably use smaller resistors for the input filter. No need to take a nice low impedance shunt signal and add 10kohm+ IMO.