Reply To: Battery Measurement accuracy

So just checking, the Vbat voltage range. I’ve an application that doesn’t have a solar panel.     I’m trying to figure out     1) can a LiSOCl2  3.6V 19Ahr be used  2) 3*D Mn 18Ahr    3) generically characterize  the LiPo measured Voltage,

https://www.batteryspace.com/primary-lithium-thionyl-chloride-battery-d-size-3-6v-19-ah-er34615-saft-lsh20-non-rechargeable-68-4wh-200ma-rate-un-38-3-passed-5-7.aspx

So I’m finding for some reason there is a need for 0.55V overhead from the processors 3.3V. Seems this the LDO and the MOSFET.

I’ve generated some data, graph below. This is a miniumium load, of the Mayfly taking a measurement every two minutes, and sleeping inbetween. Its measuring  the Vbatt through uP A6 (blue line), then  ADS1115 through a divider 1M/1M measuring the Vbat (orange line) and also the processors 3.3V (grey line)

Graphs or attachment below

I also added a low ESR 680uF cap to V_batt (Digikey 732-9079-1-ND ) to buffer any current pulls. Though under V_batt=3.45V  the processor starts to reset.

My method is for an FTDI 3.3 cable to J2. I’ve disconnected J2-4/Vusb and J2-1/DTR. No USB cable.  Variable power supply connected to JP1. Then vary the power supply Vbatt, measure externally with a Meterman 37XR volt meter to set the voltage, wait for a measurement to show on the TTY. Then drop the voltage and repeat.

The orangle line is accurate, the blue starts to droop under a real 4.1V, and flattens underneath a real 3.8V, as the 3.3V then starts to droop.

So I’m wonder if you have seen the same and any thoughts on the 0.55V overhead. That kinda eliminates the compact LiSOCl2 battery for me.