Other Parts Discussed in Thread: LMV342, INA168
I'm not sure this is the right forum for this, but I'll start here and see where it goes.
I'm attempting to measure the current on a 48-VDC bus. I have a 50-mΩ high-side shunt resistor that is being monitored by an INA168 with a 100-kΩ output resistor. When the bus current is 14 mA, the output of the INA168 is 14 mV. So far, so good. I then buffer this through an LMV342 op-amp connected as a unity-gain buffer. The output of the op-amp is 14 mV. Still good. The output of the op-amp is connected to a series 470-Ω resistor (thick film, 0603) and a 1-µF capacitor to ground (X7R, 0603) – a passive low-pass filter. This is then connected to an analog input (P6.1, pin 2) of the MSP430F5310 – and nothing else. The voltage here is 82 mV. Huh? Where did that come from? The ADC in the MSP430 dutifully reads the voltage as 82 mV, which corresponds to a current almost six times higher than the actual current!
I cannot figure out what is causing this. The only things between the output of the buffer amp (where the voltage is 14 mV) and the analog input of the MSP430 (where the voltage is 82 mV) is a resistor and capacitor.
In total, I am reading 4 analog channels (P6.0–P6.3) in repeated autoscan mode, and the other three channels are all fine. P6.2, which is scanned right before this one, has an identical op-amp buffer + RC low-pass filter input configuration. I've checked the GPIO configuration bits, and they all look fine.
Is current somehow being sourced from the MSP430's input pin? Is this bleed from the previously scanned channel? (But it seems like if this was bleed, only the conversion result would be affected. In this case, the voltage is actually physically there on the board.)
