In telecom power systems, the positive rail is usually grounded. This forces designers to place a shunt resistor on the system's negative rail to measure current. Yet there are other ways to sense the system's current.
Some methods even use a transducer, like Hall-effect or magnetic inductance current transducers, instead of a shunt resistor. These transducers are made for only certain current values (50, 100, or even 200 A). Although effective, such solutions are more expensive than the shunt-resistor transducer, which also is manufactured for many more current values.
Using a shunt resistor, some approaches implement analog optocouplers or even isolation amplifiers for positive-referencing applications. The actual ICs available for high-side current measurements are made for using shunts on the positive rail of the system, referencing on the negative. The technique presented here is based on this last solution, but adapted for telecom power systems.
Figure 1 shows the design's basic topology. A 50-mV shunt resistor senses the system's current. Op-amp IC1 amplifies the voltage on the shunt via an n-channel enhancement-mode MOSFET, whose gate is connected to the output. This MOSFET produces a current proportional to the voltage drop measured on the shunt. A part of the signal that appears on R1 is fed back to the noninverting input by means of a voltage divider formed by R2 and R3:
50 mV = (VfMAXR3)/(R3 + R2)
R4 will drop a voltage VO' proportional to the current produced to feedback a signal to IC1. Basically, the maximum voltage on R1 will appear when RSHUNT drops 50 mV, because:
VfMAX = \[50 mV(R3 + R2)\]/R3
This solution requires a pair of sources designed specially for IC1 and IC2. IC1 must be referenced to the negative rail, and it has to be a single-supply op amp biased by VCC. This supply must be greater than VfMAX plus the gate-to-source voltage needed to turn on the MOSFET. An offset nulling circuit has to be used on IC1 as well, because the maximum voltage drop on the shunt is just 50 mV. Virtually any offset would cause a relatively large reading error.
IC2 must be biased with both negative and positive supplies. The maximum voltage read on R4 should be within the negative rail. This is VEE, and it shall be greater than VO'. In this case, the positive supply can be the same 5-V supply used for the analog-to-digital converter (ADC). This particular issue determines whether IC2 should be chosen as a rail-to-rail op amp or as a standard split-supply op amp. IC2, configured as an inverting amplifier, has the task of converting the negative signal obtained in R4 to a positive voltage, VO, adequate for the ADC's input. In this case, if R5 = R6, then |VO| = |VO'|.
Figure 2 shows the complete design for a 38- to 60-V dc typical power system. This design can read positive and negative currents by adding IC3, IC4, and IC5. Texas Instruments' TLC271C op amp is used as IC1 and IC3. This particular op amp has a very low offset voltage drift and can be configured to consume a small amount of current. The n-channel MOSFETs are BS107s. They allow a maximum drain-to-source voltage of 200 V. If the dc power system has higher voltages, the power supplies for IC1 and IC2 then will have to be redesigned.
IC4, a unity-gain inverting amplifier, is a rail-to-rail input-output op amp. It's assumed that the ADC's reference is the 5-V rail. So when the output is maximum, the digital conversion is the highest.
The supplies used to bias IC1, IC2, IC3, and IC4 are designed using MPSA92 and MPSA42 transistors. If the maximum power to be dissipated by them is 0.625 W, then a regulator constructed like those shown in Figure 2 can deliver up to approximately 11 mA and 12 mA for Q3 and Q4, respectively.
The ICs used here won't consume that much current, even in the worst cases. A reference formed by a zener diode and a resistance sets the voltage of these supplies. These elements have to be calculated for an accurate and safe nonheating performance when used on higher-voltage systems.
IC3 does pretty much the same as IC1, but only for negative currents flowing through RSHUNT. This means that a battery bank's discharge current can be sensed using the same ADC input. IC4 is a simple single-supply op amp used as a comparator. It just reports the sign of the current measured on RSHUNT, via an optocoupler (IC5) like Sharp's PC817, which can be turned on with less than 2 mA.
Finally, the output voltage, VO, is expressed as follows:
VO = \[VSHUNTR4(R3 + R2 + R1)/R3R1)\]