Differential signals are standard for many different communications protocols including LVDS, CAN, USB, RS485/422, HDMI, and so on and incorporates two complementary signals (differential pair) that are used to transmit data.

While single ended communication protocols such as I2C or SPI are often more ubiquitous due to their simplicity and are referenced to a single, stationary potential, differential signaling offers numerous advantages over single ended communication schemes and several are as follows:

• EMI (electromagnetic interference) and Common Mode Noise Resistance

• EMI and Common Mode Noise can be introduced from a variety of sources that can be coupled to the data transmission signals and can distort what the receiver is reading. In protocols with differential signaling, this effect is mitigated significantly as the noise is emitted onto both signals meaning that each inverted and non-inverted signal will constructively cancel each other out once they reach the receiver:

• Ground Shift Immunity

• In addition, many differential signal protocols do not share any common ground or return current path, so ground reference shifts and ground noise acquired along the path are of no concern when utilizing such differential signaling protocols. However, for protocols that do call for common termination (such as the CAN interface as an example), there needs to be more care taken to reduce ground offsets as much as possible.

• Reduced Power Requirements and Higher Frequency Operation

• Differential signaling often calls for voltage signals to be in the hundreds of millivolts or just up to 1 V or 2 V and can greatly simplify power consumption needs of an application. In turn, this also allows these differential signals to operate at higher frequencies due to the much lower radiated emissions these signals have from the smaller voltages used (in addition to the resistance from external EMI imposed by using a differential pair).

Common mode signals can be simply defined as the voltage that is common between the input terminals of a device and are often an unwanted element in most systems. For instance, take the visual representation as shown in Figure 1-3.

It is evident that when a common mode voltage is introduced to a system, this will shift the reference point of the affected components to be outside of the anticipated operating range by a given VCM (as shown above from both a graphical and equivalent circuit perspective). While common mode signals can be attributed to EMI, ground shifts, miswirings, coupling, or even lightning strikes, they can be influential enough to the system at large that it can significantly affect measurement accuracy or even permanently damage devices by causing them to exceed absolute maximum operating ratings. Hence, it is essential that the design is defined by the environment the system will be used in and components selected that can tolerate such conditions.

For multiplexer applications that may be exposed to high common mode voltages, there are several multiplexer parameters that need to be considered for these situations.

The most evident parameter that one must consider is the operating supply range of the multiplexer. During nominal operation, the expected signals propagating through the multiplexer can usually be fairly low. However, if there exists common mode voltages that can potentially be upwards of ±36 V in some cases, then a higher voltage capable multiplexer would be needed to be able to survive and operate in these instances. While it may seem that this would necessitate a higher supply voltage, which certainly can be the remedy, there is a way to still utilize a lower supply rail while being able to survive an event where a high common mode voltage is present. Below illustrates how external diodes and current limiting resistors (typically in the 1k to 10k range) can be implemented to suppress damage to the multiplexer while allowing time for the fault to subside and resume nominal operation:

Lastly, if there still needs to be operation during a high common mode voltage event and there is no access to high supplies in the system, beyond the supply multiplexers can be used instead which offer low voltage supply rails, but can pass and tolerate voltages on the multiplexer inputs beyond those supply rails.

In many applications that are susceptible to common mode interference, higher speed communications are typically implemented and the multiplexer must be able to pass these signals unattenuated. The bandwidth specification found in all TI multiplexer data sheets can give an indication of the frequency of signals that can be handled by the multiplexer so they can be passed without issue. A good rule of thumb is to select a multiplexer that can handle a bandwidth of at least 1.5x-3x the application requirement to allow for margin in the design for layout and loading constraints.

At higher frequencies, there is also concern about channel to channel crosstalk. Crosstalk is defined as the magnitude of the signal of an ON-channel that appears on an adjacent channel. This can cause errors in any measurements or can introduce noise on adjacent lines that can negatively impact communications if the multiplexer does not have very good crosstalk performance.

All multiplexers will have some form of leakage current that can be injected into the application, some more and some less. What can be problematic during an event that imparts a high common mode voltage onto the inputs of a device is that the leakage current will certainly increase and if not accounted for in the system, can potentially affect the accuracy of the application.