Measuring sense-sesistor’s voltage drop, Part 4: isolation amplifiers

Measuring the voltage drop across a current-sense resistor can be a trivial or challenging issue depending on the rail voltage and other factors.

The final part of this four-part FAQ looks at specific isolation amplifiers and their techniques for solving CMV and related issues.

Q: What’s in an isolation amplifier compared to a non-isolation amplifier?
A: In Figure 1, you can see that an isolation amplifier functionally looks like a standard amplifier but is actually built in two parts with no ohmic (galvanic) path between them. The amplifier’s front end takes the voltage being monitored, uses it to modulate a carrier, transmits this modulated carrier across an isolation barrier, and then demodulates it. The result is an analog signal corresponding to the input (although it may also have some gain) and so can be used just as the non-isolated signal is used.

Figure 1: The basic isolated amplifier functionality is the same as a non-isolated amplifier except for the absence of a galvanic path from input to output. (Image: TT Electronics)

Q: That seems straightforward, but is it?
A: Yes and no. In many ways, the isolation amplifier (often referred to as isoamp) in just about as straightforward to use as its non-isolated sibling, but there are some caveats. As with non-isolated amplifiers, they have specifications beyond their maximum isolation voltage; these are related to static and dynamic performance, additional errors, temperature coefficients, input and output offsets, and more.

Second, the isolated front end needs power, which must also be isolated from the system power supply; if this is not done, the signal isolation function is negated. To provide this power, the user must add a small isolated supply (not a big deal) or select a more advanced iso amp, including an integral isolated power supply,y as shown in Figure 2.

Figure 2. The Analog Devices AD210 isoamp aso includes an isolated power supply for the sense-resistor front-end circuitry; other isoamps require an external supply (which makes sense if an isolated power rail is already available. (Image: Analog Devices)

Q: How is the isolation implemented?
A: In an all-analog approach, the barrier is implemented via magnetics (transformer), capacitive coupling, or even a modulated RF carrier with a tiny wireless link. Note that an optical barrier could also be used in principle. Still, it is difficult to maintain linearity of the optocoupler transfer function, so it is generally not used despite its other virtues.

Some vendors provide products with multiple approaches, while others focus on only one isolation technology. In the latter case, any comparative summary they offer of key pros and cons of isolation techniques may be biased, of course. However, the key attributes of each also vary with signal frequency, voltage range, and other factors, so it can get complicated.

Q: Thus far, you have only discussed all-analog isolation techniques. Is there a digital-focused alternative?
A: Yes, we can observe the hybrid analog/digital approach in Figure 3. There are several variations, but in each, the signal to be isolated is amplified and then digitized prior to passing through the isolation barrier. The digital signal is then used to modulate a carrier that crosses from input to output via one of the energy-transmission paths cited above (magnetic, capacitive, RF). In addition, it can be used to turn an LED on/off in conjunction with a phototransistor in an optocoupler since transfer-function linearity is not an issue for digital signals. At the output side, the digital signal is demodulated and restored as an analog signal or can be used directly as a digitized version of the isolated input, depending on the system design.

Figure 3. In the hybrid isoamp approach, the sense-resistor voltage is digitized and then is available as a serial bit stream, which opens up new options for the isolation barrier itself and also removes the linearity demands on the isolated signal’s source, path, and receiver. (Image: Texas Instruments)

Q: What’s the advantage of hybrid technique?
A: First, it allows for the use of optical coupling for isolation, which is often a desirable or even preferred option. Second, it can be used with magnetic, capacitive, and RF isolation. Still, it minimizes concerns about signal amplitude linearity of the modulation/demodulation process, as the key to the linear performance of input versus output amplitude is the linearity of the D/A and A/D conversions rather than the modulation scheme, and that is maybe easier to maintain over a wide dynamic range.

Conclusion
This FAQ has only dealt with some issues related to resistor voltage sensing. Among the issues that also could be explored further are:

The pros and cons of using a sense resistor versus other sensing transducers such as Hall-effect devices or various types of magnetic coils. Also, how do these pros/cons change over the range to be sensed, under what circumstances, and when comparing, even with the biases of vendors of each approach?
The different technologies for isolation and the key attributes of each and relative to each other; also, where each might be a “better fit,” which is, as expected, a function of the application, range, and, again, individual vendor perspectives.
Key top-tier and secondary-tier specifications associated with non-isolated and isolated amplifiers, including, but not limited to, initial offset, gain accuracy, temperature-induced drifts in offset and gain, power consumption, size, and cost.
There is a need in many cases for four-wire Kelvin connections when sensing a high-current voltage across a resistor, and how many of these higher-current resistors are designed to facilitate such contact arrangements.

Conclusion
Using a sense resistor for measuring current via voltage drop seems simple, and it is, and in many ways, it is. Of course, as with so many other simple-appearing techniques, subtleties and issues need to be understood and addressed, as the details make the difference between a successful solution and an inadequate or perhaps even dangerous one. Resistor-based current sensing over ranges from small currents in the single digits range to tens and hundreds of amps is widely and successfully used, so the needed components, applications expertise, and hands-on guidance are available.

Related EE World Content

External References (these are just a few of the many available ones; current sensing via a resistor or other means is a very important and widely discussed topic)