High-side current sensor has period output

You use high-side current monitoring in many battery-powered products that require accurate monitoring of load current, charger current, or both. In applications for nonportable designs, high-side-current monitoring serves as a power-supply watchdog that can flag a failure in downstream devices. The monitoring can also eliminate hazardous conditions by preventing power-supply overloads. Further, high-side-current monitoring of motor/servo circuits can produce useful feedback in control applications. These applications require a device that converts high-side current directly to a digital signal (Figure 1). IC1 is a low-cost, high-side-current-sense amplifier that converts high-side current to a proportional, ground-referenced voltage. Its two internal comparators (latching and nonlatching) implement a voltage-to-pulse converter that produces an output pulse width proportional to the measured current.

Figure 1 The duration of a negative-going pulse at COUT1 is proportional to the current flowing through RSENSE.

IC1’s Out pin charges C1 via R1. When C1’s voltage reaches 0.6V, Comparator1 latches in the high-impedance state. The time required to charge C1 to 0.6V is proportional to the measured current. Comparator2, in conjunction with the Reset pin, initiates the conversion and removes the previously existing charge on C1. The Reset and C IN2 pins, tied together and connected to a TTL-compatible microcontroller output, CTRL, control the conversion process. Normally, CTRL is high. The microcontroller starts a conversion by pulsing CTRL low, discharging C1 and clearing the latch in comparator1 (C OUT1 goes low.) The microcontroller now measures the time from the CTRL transition to the low-to-high transition at C OUT1 (Figure 2). The period begins at the low-to-high transition of CTRL and ends at the low-to-high transition of C OUT1. As a function of the current levels of interest, you select R1 and C1 values to create pulse durations in the tens of milliseconds. As a result, the Out settling time of 20 µs and the comparator propagation delays of 4 µs have negligible effects on the measurement accuracy.

Figure 2 These waveforms illustrate the operation of the circuit in Figure 1.

To derive an expression for the output pulse width, start with the relationship for an RC-charging circuit: VTHRESH = Out(1 – e – TPULSE/R1C1). For the Out-pin voltage, substitute the expression ILOAD ×RSENSE × A V and solve for ILOAD: ILOAD = VTHRESH/ [RSENSE ×AV (1 – e – TPULSE/R1C1)], where ILOAD = measured current in amperes, VTHRESH = comparator threshold = 0.6V, RSENSE = current-sense resistor in ohms, AV = gain of IC1, and TPULSE = time to charge C1 to VTHRESH in seconds.

For example, selecting R1 = 1 MW, C1 = 0.1 µF, RSENSE = 0.075W, and AV = 20 produces a TPULSE measurement of 0.022 seconds in response to a 2A current. Thus, given a microcontroller timer port, an external interrupt, or simply an available microcontroller input, IC1, and two external passive components implement high-side-current-to-digital conversion without the need for a discrete A/D converter.