DETAILED DESCRIPTION

Dual-Channel Mode

FIR1

FIR2

FIR3

FIR4

IOUTB1

IOUTB2

IOUTA1

IOUTA2

16-Bit

DAC

FIR5

x

sin(x)

DA[15:0]

DB[15:0]

2fDATA

2y–16y fDATA

8fDATA

4fDATA

16fDATA

fDATA

y2

A Gain

A

Offset

FIR1

FIR2

FIR3

FIR4

16-Bit

DAC

FIR5

x

sin(x)

2fDATA

8fDATA

4fDATA

16fDATA

y2

B

Offset

B Gain

DEMUX

B0020-01

Single-Sideband Mode

DAC5686

www.ti.com ............................................................................................................................................................ SLWS147F – APRIL 2003 – REVISED JUNE 2009

In dual-channel mode, interpolation filtering increases the DAC update rate, thereby reducing sinx/x rolloff and

enabling relaxed analog post-filtering, and is useful for baseband I and Q modulation or two-channel low-IF

signals. The dual-channel mode is set by mode[1:0] = 00 in the config_lsb register. Figure 18 shows the data

path architecture in dual-channel mode. The A- and B-data paths, which are independent, consist of four

cascaded half-band interpolation filters, followed by an optional inverse sinc filter. Interpolation filtering is

selected as 2×, 4×, 8×, or 16× by sel[1:0] in the config_lsb register. Magnitude spectral responses of each filter

are presented following in the section on digital filtering. Full bypass of all the interpolation filters is selected by

fbypass in the config_lsb register. The inverse sinc filter is intended for use in single-sideband and quadrature

modulation modes and is of limited benefit in dual-channel mode.

Figure 18. Data Path in Dual-Channel Mode

Single-sideband (SSB) mode provides optimum interfacing to analog quadrature modulators. The SSB mode is

selected by mode[1:0] = 01 in the config_lsb register. Figure 19 shows the data path architecture in

single-sideband mode. Complex baseband I and Q are input to the DAC5686, which in turn performs a complex

mix, resulting in Hilbert transform pairs at the outputs of the DAC5686's two DACs. NCO mixing frequencies are

programmed through 32-bit freq (4 registers); 16-bit phase adjustments are programmed through phase (2

registers). The NCO operates at the DAC update rate; thus, increased amounts of interpolation allow for higher

IFs. More details for the NCO are provided as follows. For mixing to fDAC/4, DAC5686 provides a specific

architecture that exploits the {… –1 0 1 0 …} resultant streams from sin and cos; the NCO is shut off in this mode

to conserve power. fDAC/4 mix mode is implemented by deasserting nco in register config_msb while in

single-sideband or quadrature modulation mode.

Copyright © 2003–2009, Texas Instruments Incorporated

Submit Documentation Feedback

17

Product Folder Link(s): DAC5686