Dynamic and static CMR: impacts to signal integrity

Optocouplers can protect systems and users from high-voltage surges. They can also reject high common-mode transient noise that would otherwise result in abnormal voltage transitions or excessive noise on the output signal.

By Yeo Siok Been • Avagago Technologies -- EDN Europe, 01 Aug 2010

Noise can have many sources. It can, for example, come from nearby electric fields’ capacitive coupling, magnetic fields’ inductive coupling, or conductive coupling due to differences in ground potentials. To combat this noise, devices often use CMR (common-mode rejection) to reject signals that are common to the input. This parameter is important especially in noisy environments because noise appears as offset in the inputs. The ability of a device to attenuate the noise and to transmit signals across it is important for signal integrity.

The key factors that determine the CMR performance of a device are common-mode voltage and common-mode transient noise. When the common-mode noise voltage is larger than the interface’s input range or the input common-mode voltage’s rejection range, you must use galvanic isolation. The switching current that causes voltages to spike often generates common-mode transient noise, which you define using the voltage and the rate of change. You must filter both common-mode noise and transient noise to ensure that the input signal does not become corrupt.

The Isolator’s Role

An isolator connects two electrical modules within a system to ensure that the system can transmit signals without a physical connection to the destination. Light is the most common mode through which you can transmit a signal within an isolator. In this case, you use an optocoupler as an isolator not only to protect systems and users from high-voltage surges but also to reject high common-mode transient noise to either set of floating points, which otherwise may result in abnormal voltage transitions or excessive noise on the output signal.

The key parameter that measures the success of rejecting common-mode signals in an isolator is the common-mode current during a voltage transient. The following equation defines the common-mode current: ICM=C(dV/dt), where C is the parasitic capacitances due to packaging—pin- to-pin leadframe or wirebond leadframe, for example—and capacitance at the signal-coupling interface between the LED and the detector; and dV/dt is the rate of the transient-voltage signal. This equation assumes that external factors, such as PCB (printed-circuit-board) layout or component placement, are optimized and contribute little parasitic capacitance.

To reduce the common-mode current, it is important to reduce these parasitic capacitances. The large separation between the LED and the photodetector—0.08 to 1 mm in a representative part, such as Avago’s (www.avago.com) ACPL-J313—minimizes the parasitic capacitance and thus results in small leakage current during common-mode transients. The part also has a built-in proprietary Faraday shield between the input LED and the photodiode to provide increased common-mode noise rejection.

This internal transparent, conductive shield allows optical coupling to the photodiode but diverts electrically coupled current to the ground pin, improving the parasitic capacitance, which in turn improves the CMR of the optocoupler. In addition, this shield helps to discharge the charges that accumulate on the detector chip due to any large common-mode voltage that is applied across the device for a substantial period of time. The ACPL-J313 can improve CMR performance by as much as 40 kV/μsec at 1.5-kV common-mode voltage.

On the other hand, a design in which the microcontroller is far away from the isolation interface and which thus has long wire traces, can easily pick up inductive noise and corrupt the signal. In this case, a direct-drive optocoupler, such as the ACPL-M61L, acts as a “natural” filter. A resistor is required to limit the current that drives the LED in an optocoupler (Figure 1). Together with the intrinsic input capacitance of the LED, the resistor/capacitor pair acts as a lowpass RC filter to shunt away the high-frequency noise.

Static and Dynamic CMR

An isolator has both static and dynamic CMR. Static CMR is the signal-rejection capability when the input is static at either a logic high or a logic low. This situation usually occurs when the system is in the idle, or standby, state. During this state, some parts of the system are turned off to save power, leaving certain modules on to detect the input signal. The system must maintain and hold at the same logic regardless of the static common-mode noise in the environment. This requirement ensures that the noise does not falsely trigger the system.

Typically, thedynamic CMRof a system is worsethan the stat icCMR, and the CMRcapability is worseat higher commonmodevoltagesfor DC voltage.

In a dynamic environment, the system transmits a signal that toggles between logic high and logic low. To prevent the common-mode noise from corrupting the input signal, the system must filter the noise using dynamic CMR. Typically, the dynamic CMR, or ac voltage, of a system is worse than the static CMR, or dc voltage, and the CMR capability is worse at higher common-mode voltages for dc voltage.

Dynamic CMR ensures that the system does not lose signals during operating mode. During this mode, most parts of the system are operating, and any erroneous signal the system transmits will cause the connecting devices to erroneously turn on or off. This situation may result in a short circuit; overheating; or the destruction of expensive devices, such as motors and machines.

Rejecting common-mode transients

Some optocouplers adopt a level-triggered coding scheme. In this scheme, the LED detects the level of forward current the input signal sets and pulses the light output to the detector. In contrast, an edge-triggered coding scheme adopted in other isolators generates small voltage pulses during the transition edge of the input signal (Figure 2).

  1. Go to the Talkback section at www.edn.com/ ms4368 to post a comment on this article.
  2. For more technical articles, go to www.edn. com/features.

The duration of the current or voltage pulses is 100 nsec, for example, in a level- triggered coding scheme, such as that in an Avago ACPL-072L. As such, dynamic transient noise is less likely to corrupt the signal in these systems than in edge-triggered coding schemes, in which these pulses last 2 to 3 nsec. Due to the short rise and fall times of the LED current, the dynamic transient noise will probably result in a slight increase or decrease of pulse-width distortion in the level-triggered coding scheme. The edgetriggered scheme, in contrast, can possibly momentarily miss pulses. In short, optocouplers with level-triggered coding have inherently better dynamic-CMR performance.

Figure 3 shows the dynamic-CMR performance of a level-triggered optocoupler—the ACPL-072L—that can withstand a 10-kV/μsec dynamic-CMR surge at a 25-Mbps signal rate, rejecting the dynamic CMR to preserve signal integrity

AUTHOR’S BIOGRAPHY
Yeo Siok Been is digital-optocoupler-product manager at Avago Technologies, where she has worked for five years. She is responsible for product development and marketing. Siok Been has a master’s degree in electrical engineering from the University of Singapore. Her personal interests include travel. You can reach Siok Been at siok-been.yeo@avagotech.com.


 

Our Sponsors



Ads by Google