Voltage & Current Measuring Circuit

So as discussed in a previous post, the state-of-charge can be approximated by measuring the voltage of the battery. We wanted rough statistics for our Interstate battery and so we emailed them. We got the chart below in return.

Now unfortunately, batteries will vary as they degrade so this chart won't be valid for long and further more the chances of us obtaining the same battery in Uganda are very slim. So what we incorporated into the design of the voltage measuring circuit was a variable resistor which is in series with the reference voltage resistors. By varying this resistor we will be able to calibrate the measuring circuit to the individual battery.

The Design:

The diagram below is the circuit which will accommodate both current and voltage measuring. Works much like the current measuring circuit from before. The major differences in this circuit is that the reference voltages to each op-amp is coming from a single voltage dividing branch instead of individual branches which is slightly more in accurate because more resistors with non-zero tolerances are involved in the one conductive path but this will save on resources, complexity and power consumption since the total branch resistance is much higher.

The only other major difference is that this circuit will utilize that same array of LEDs and op-amps for both voltage and current measuring. When measuring the current that the solar panel is outputting we take our test voltage (input signal) to be the voltage drop across the ballast resistor (MP915) through which all the current from the charging battery and load flows towards ground. Ground in this case is defined to be the return of the solar charge controller (SCC). To measure voltage from the battery our test voltage is the voltage after the battery voltage is dropped via a zener diode (NTE141A). We needed this zener diode to ensure a constant drop of voltage to a value comparable to the voltage we measure for current measuring. Our 11.5 V Zener diode will give is 13.5 - 11.5 = 2 V for a full battery and 11.5 - 11.5 = 0 V for a completely dead battery.

The switches are a bit of a mess but essentially there will be one push switch (S1) to power all the op-amps, one switch to choose weather you want to measure something with reference to the resistor branch calibrated for voltage or current (S2) and finally a large 10 A rated push switch (S3) which will divert current through the ballast resistor (MP9100) when measuring current. We have S1 to ensure that the meter isn't running all the time and wasting current by powering the resistor branch on op-amps. S2 is a double switch and will not only choose the resistor branch but also what the input signal is; voltage from battery or voltage across ballast resistor. S3 was necessary because we don't want current constantly flowing through the ballast resistor but just momentarily as a measurement is taken.

Calibration:

The voltage dividing branch of resistors which give the reference voltage to each op-amp will actually work perfectly if the Zener diode drops 11.5 V exactly. For the NTE141A there should be an 11.5 V drop in the reverse bias direction but in our application we found through testing that there was only ~ 11 V dropped. We might experiment with having more current flow through the diode to give a more accurate drop. (Just putting a resistor in series with the diode directly to ground rather than using the extremely high impedance of the op-amp inputs to get to ground.) We will also be testing the circuit with a 12 V Zener diode and perhaps that will drop 11.5 V with our small current load.

The resistor branch which gives each reference voltage was pick precisely such that it can accommodate increments of 0.1 V from 0.1 to 0.8 V for current measuring given that the variable resistor at the start of the branch is correctly tuned (in theory 42 kΩ for the current measuring branch and 12 kΩ for the voltage measuring branch). These static resistors must also be able to accommodate jumps of 0.25 V from 13.5 to 11.75 V when a different variable resistor begins the branch. In theory all the values calculate perfectly however in practice the inaccuracy of the components causes a skew of the LED array which is not too bad. The more we calibrate away from the intended values the worse the skew gets and the more inaccurate one end of the LED array is. If we can be sure of things like an 11.5 V on the Zener diode the accuracy will be much higher.

Demonstrations: