The DS250DF230 output driver has a three-tap finite impulse response (FIR) filter which allows for pre- and post-cursor equalization to compensate for a wide variety of output channel media. The filter consists of a weighted sum of three consecutive retimed bits as shown in Figure 8-1. C[0] can take on values in the range [-31, +31]. C[-1] and C[+1] can take on values in the range [-15, 15].

Figure 8-1 FIR Filter Functional model

When using the FIR filter, it is important to abide by these general rules:

• |C[-1]| + |C[0]| + |C[+1]| ≤ 31; the FIR tap coefficients absolute sum must be less or equal to 31
• sgn(C[-1]) = sgn(C[+1]) ≠ sgn(C[0]), for high-pass filter effect; the sign for the pre-cursor and/or post-cursor tap must be different from main-cursor tap to realize boost effect
• sgn(C[-1]) = sgn(C[+1]) = sgn(C[0]), for low-pass filter effect; the sign for the pre-cursor and/or post-cursor tap must be equal to the main-cursor tap to realize attenuation effect

The FIR filter is used to pre-distort the transmitted waveform to compensate for frequency-dependant loss in the output channel. The most common way of pre-distorting the signal is to accentuate the transitions and de-emphasize the non-transitions. The bit before a transition is accentuated through the pre-cursor tap, and the bit after the transition is accentuated through the post-cursor tap. The waveforms in Figure 8-2 through Figure 8-4 give a conceptual illustration of how the FIR filter affects the output waveform. These characteristics can be derived from the example waveforms:

• VOD_pk-pk= v₇ – v₈
• VOD_{low-frequency} = v₂ – v₅
• Rpre_dB = 20 × log₁₀ (v₃ ⁄v₂ )
• Rpst_dB = 20 × log₁₀ (v₁ ⁄v₂ )

Figure 8-2 Conceptual FIR Waveform With Post-Cursor Only

Figure 8-3 Conceptual FIR Waveform With Pre-Cursor Only

Figure 8-4 Conceptual FIR Waveform With Both Pre- and Post-Cursor