[Edited] When should one adopt DC coupling versus AC coupling?
How about when high-speed signals are being cascaded between serial processing stages? DC coupling provides simpler connection styles, but is its performance equivalent to AC coupling?
Thanks for your interest in High-Speed Digital Design.
AC coupling works successfully only with binary signals that have equal numbers of ones and zeroes. Examples of such signals include a 50 percent duty-cycle clocks, Manchester-coded data signals, and ANSI Fiber-Channel 8B10B signals. Many other data codes also enforce equal numbers of ones and zeroes. The property of having equal numbers of ones and zeros is called the property of DC balance.
The property of DC balance is often specified like this: "The number of ones and zeroes, averaged over any sequence of N bits, differs by no more than M."
For a signal to have good DC balance means there is a low-frequency cutoff to the spectrum of the data signal below which no useful information is carried. The spectrum of the signal may include a spike at DC (i.e., a DC offset), but no useful, changing, data-dependant information lies below the cutoff.
If you are working with a code that guarantees no more than one bit of offset averaged over every pattern of N bits, you can assume the low-frequency cutoff is located at a frequency somewhere around 1/(10*B*N), where B is the bit interval, N is the specified averaging period, and 10 is a fudge factor intended to guarantee your cutoff lies sufficiently below the actual spectral content of the signal.
If the signal has good DC balance, it is possible to pass it through a suitable high-pass filter without altering the informational content. This works if the passband cutoff frequency of the high- pass filter lies below the low-frequency cutoff of the data signal.
A series-connected DC-blocking capacitor is a good example of a commonly used high-pass filter. The DC- blocking capacitor strips away all low-frequency content in the signal.
Whenever you pass a signal through any AC-coupled circuit there is always SOME tiny part of the signal that gets filtered out. This filtering effect induces a tiny difference between the informational part of the signal before and after filtering. In severe cases this difference manifests itself as the phenomenon of "DC-wander".
To estimate the degree of DC wander possible when passing a particular code through a certain high pass filter HPF(w), first set up a complimentary filter LPF(w), defined thus:
Then, pass the data code through the filter LPF(w) and look for the worst-case output. Whatever the magnitude of the output of LPF(w), that's the magnitude of the worst-case DC-wander error you will experience when passing the signal through HPF(w).
(An expanded discussion of DC wander appears in my newsletter vol. 7, #09, DC Blocking Capacitor Value )
Here are three good reasons why you might want to AC-couple a signal.
- To change the DC bias level when interconnecting logic families with different switching thresholds.
- To provide a removable interface that may be shorted to ground without damaging the output drivers.
- When combined with differential signaling and transformer coupling, to connect boxes without requiring any DC connection between the two product chassis.
If these advantages are important to you, and if you have a DC-balanced signal, use AC coupling. Otherwise, for general logic interconnections within the same printed circuit board, stick with a simple, direct-DC-coupled interface.
Given circuits of sufficient complexity, you can restore the lost low-frequency content of a high-pass-filtered digital signal even if that signal lacks the DC-balance property. Such "DC- level-restoration circuits" are commonly used in DSP-based transceivers.
Dr. Howard Johnson