High Speed, Digital-to-Analog Converters Basics

Application Report
SLAA523A – March 2012 – Revised October 2012
High Speed, Digital to Analog Converters Basics
Chris Pearson ..................................................................................................................................
ABSTRACT
The goal of this document is to introduce a wide range of theories and topics that are relevant to highspeed, digital-to-analog converters (DAC). This document provides details on sampling theory, data-sheet
specifications, common system-level concerns, and the common functions of interpolation DACs such as
FIR filters, Digital Mixing and Quadrature Modulator Correction.
1
2
3
4
5
6
Contents
Introduction .................................................................................................................. 2
Spectral Performance Terminology ....................................................................................... 3
DAC Nyquist Zones, Zero Order Hold, and Images .................................................................... 5
Interpolating DAC Topic 1: FIR filters .................................................................................... 7
Interpolating DAC Topic 2: Digital Mixing .............................................................................. 12
Interpolating DAC Topic 3: Quadrature Modulator Correction (QMC) .............................................. 19
List of Figures
................................................................................... 2
.................................................................................... 3
Single tone DAC Performance Terminology ............................................................................ 4
WCDMA ACLR .............................................................................................................. 4
IMD3 .......................................................................................................................... 5
Ideal Sine Wave ............................................................................................................. 5
Ideal DAC Sine Wave ...................................................................................................... 5
Hypothetical Spectrum (sampled at Fs) .................................................................................. 6
Zero Order Hold Time and Frequency Domain plots................................................................... 7
Convolution in the Time Domain .......................................................................................... 7
Convolution in the Frequency Domain ................................................................................... 7
DAC5688 Block Diagram – FIR filters .................................................................................... 9
Interpolation Example ..................................................................................................... 10
Single FIR Filter and a Composite FIR Filter Spectrum .............................................................. 11
DAC5688 Block Diagram – Clock Section ............................................................................. 11
DAC5688 Block Diagram – Digital Mixing Section .................................................................... 12
Digital Mixer, Complex IQ In and 1 Real Out .......................................................................... 13
Digital Mixer, Complex IQ In and 1 Real Out (2 examples).......................................................... 14
Digital Mixer, Complex IQ In and 1 Real out ........................................................................... 15
Digital Mixer, Complex IQ In and 1 Real out – Wideband example ................................................ 16
Real Output System Architecture ....................................................................................... 17
Digital Mixer, Complex IQ In and Complex IQ Out.................................................................... 17
Digital Complex IQ In, Digital Complex IQ Out, With an Analog Mixer ............................................. 18
Digital Mixer, 2 Real In and 2 Real Out ................................................................................ 18
Quadrature Modulator, Common Source of Errors ................................................................... 19
DAC5688 Block Diagram – QMC Functions ........................................................................... 20
1
Basic DAC Diagram and Terminology
2
Frequency Domain vs. Time Domain
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
SLAA523A – March 2012 – Revised October 2012
Submit Documentation Feedback
High Speed, Digital to Analog Converters Basics
Copyright © 2012, Texas Instruments Incorporated
1
Introduction
www.ti.com
List of Tables
1
1
DAC Case Study ............................................................................................................ 8
Introduction
A Digital to Analog Converter (DAC or D-to-A) is a device that converts digital codes to an analog signal.
There are many categories of DACs. Some of these categories include sigma-delta DACs, pulse width
modulators, interpolating and high speed DACs. In this paper the focus will be on the interpolating and
high speed DACs.
Figure 1 provides the basic block diagram, functionality and common terminology for DACs. This figure
shows a digital word applied to the inputs of the DAC, which is then converted to an analog signal at the
sampling frequency (Fs) applied to the DAC clock. Figure 1 is a time domain representation of the DACs
input and output signals.
Analog Output:
Frequency (Fout)
Digital Input
Time D5 D4
1/Fs
0 1
2/Fs
1 0
3/Fs
1 1
.
. .
.
. .
.
. .
N-2/Fs 0 1
N-1/Fs 0 0
N/Fs
1 1
D3
1
1
1
.
.
.
0
0
0
D2 D1 D0
1 1 1
0 1 0
0 0 0
. .
.
. .
.
. .
.
1 0 1
1 1 1
1 0 1
Volts
(V)
D5
D4
D3
D2
D1
D0
6 bit
DAC
Time (s)
Clock:
t1
t2
Sampling
Frequency (Fs)
Fout: Analog Output Frequency = 1/(t2 – t1)
Fs: Clock Frequency
N: Number of digital samples
n: Number of output bits; in this 6 bit DAC example n = 6
Figure 1. Basic DAC Diagram and Terminology
Time domain representations are often described as real world signals. In Figure 1, notice the analog
output’s amplitude is shown in volts (linear) and seconds (linear). Also notice the digital input codes are
listed with time stamps (1/Fs, 2/Fs, 3/Fs…). Time domain representations are often easy to visualize and
help with understanding gross concepts. However the time domain is poor when it comes to measuring
performance of DACs and other signal processing devices. Measuring performance is best done in the
frequency domain. Therefore, it is important to understand how the time domain and frequency domain
relate.
2
High Speed, Digital to Analog Converters Basics
SLAA523A – March 2012 – Revised October 2012
Submit Documentation Feedback
Copyright © 2012, Texas Instruments Incorporated
Spectral Performance Terminology
www.ti.com
Analog Output:
Frequency (Fout)
Digital Input
Time
D5 D4
1/Fs
0 1
2/Fs
1 0
3/Fs
1 1
.
.
.
.
.
.
.
.
.
N-2/Fs 0 1
N-1/Fs 0 0
N/Fs
1 1
D3
1
1
1
.
.
.
0
0
0
D2
1
0
0
.
.
.
1
1
1
D1
1
1
0
.
.
.
0
1
0
Volts
(V)
D0
1
0
0
.
.
.
1
1
1
D5
D4
D3
D2
D1
D0
6 bit
DAC
Time (s)
Clock:
t1
t2
Sampling
Frequency (Fs)
Digital Input - Frequency Domain
Frequency Fin
Analog Output - Frequency Domain
Frequency Fout
Fout
Fin
Harmonics = Fout ´ Integer
Power
(dBm)
Power
(dBm)
2 3
Signal Images from Nyquist Zone 1
(amplitude determined by sinc function)
2 3
…
.
Noise Floor
Zone 1
Zone 2
Zone 3
…
Zone 4
Frequency (Hz)
Nyquist Zones
Frequency (Hz)
Fs/2 - Nyquist
Fs
2 ´ Fs
Fout: Analog Output Frequency = 1/(t2 – t1)
Harmonics: Integer Multiples of Fin (2 × Fout, 3 × Fout, …)
Fs: Clock Frequency
Nyquist: Fs/2
Nyquist Zone 1 = 0 to Fs/2; Zone 2 = Fs/2 to Fs; Zone 3 = Fs to 3 × Fs/2, ...
Images: Frequencies > Fs/2 (Nyquist) that land in the analog output frequency domain, explained later in more detail
Sinc function: Sinx/x
Figure 2. Frequency Domain vs. Time Domain
Figure 2 demonstrates a high level overview of the differences between the time domain and the
frequency domain. Frequency domain plots are measured in Signal Power (log scale) and frequency
(linear). Figure 2’s frequency domain plots the signal imperfections are labeled as noise and harmonics.
Notice the ease with which one can identify and quantify these imperfections. Frequency domain plots are
also commonly termed spectrums, spectral plots, or Fast Fourier Transforms (FFT). In Figure 2 the terms
Nyquist, harmonics, imaging and the sinc function are introduced. These important signal processing
terms will be discussed in more detail later.
2
Spectral Performance Terminology
SNR: Stands for Signal to Noise Ratio. SNR is the ratio of the RMS value of the fundamental (PS) output
signal to the RMS sum of all other spectral components below the Nyquist frequency, including noise (PN),
but excluding the first six harmonics and dc. Refer to Figure 3 for the SNR equation and illustration.
SLAA523A – March 2012 – Revised October 2012
Submit Documentation Feedback
High Speed, Digital to Analog Converters Basics
Copyright © 2012, Texas Instruments Incorporated
3
Spectral Performance Terminology
www.ti.com
SFDR: Stands for Spurious Free Dynamic Range: SFDR is the ratio of the power of the fundamental (PS)
to the next highest spur (PH). Refer to Figure 3 for the SFDR equation and illustration. Unless noted in the
datasheet SFDR includes the harmonics, but excludes dc.
Analog Output - Frequency Domain
Frequency Fout
SNR = 10log10 (PS /PN)
SFDR = 10log10 (P S/PH )
Fout
PS : Signal Power (Red)
PN : Noise Floor Power (Blue)
PH : Power of Next Highest Spur (Purple)
Power
(dBm)
Signal Images from Nyquist Zone 1
(amplitude determined by sinc function)
2 3
….
In this plot harmonic
#3 would be PH in the
SFDR calculation, since
it is the largest nonfundamental spur.
Zone 1
Zone 2
…
Zone 4
Zone 3
Nyquist Zones
Frequency (Hz)
Fs
2 ´ Fs
Figure 3. Single tone DAC Performance Terminology
ACLR: Stands for Adjacent Channel Leakage Ratio. ACLR is used to measure a variety of standards with
different signal bandwidths and offsets. Figure 4 is provided as an example of the 3.84Mcps 3GPP WCDMA input signal measured in 3.84MHz bandwidth at a 5MHz offset from the carrier with a 12dB peakto-average ratio. Refer to Figure 4 for ACLR equation and illustration.
Analog Output - Frequency Domain
Frequency Fout
ACLRLow = CHPCarrier – CHPLow
ACLRHigh = CHPCarrier – CHPHigh
CHP: Channel Power
Fout
Power
(dBm)
3.84MHz
3.84MHz
3.84MHz
ACP
Low
Carrier
ACP
High
….
5MHz
5MHz
Frequency
Figure 4. WCDMA ACLR
IMD3: refers to the InterModulation Distortion from the 3rd harmonic of a circuit. IMD3 is of particular
importance in narrow or wide band signals because its distortion products can land in the signal band or
the adjacent channel. This can be seen by looking at a signal that is made up of two sine waves, one at f1
and one at f2. The 3rd order products of these sine waves create frequencies at 2 x f2 – f1 and 2 x f1 – f2,
which are very close to the signals f1 and f2. Refer to Figure 5 for the IMD3 equation and illustration.
4
High Speed, Digital to Analog Converters Basics
SLAA523A – March 2012 – Revised October 2012
Submit Documentation Feedback
Copyright © 2012, Texas Instruments Incorporated
DAC Nyquist Zones, Zero Order Hold, and Images
www.ti.com
Analog Output - Frequency Domain
Frequency Fout
IMD3 = P 1 – P 3
P1 : Signal Power (Red)
P3 : IM3 Power (Blue)
f1
f2
Power
(dBm)
IM3
products
Assumes power of f1 = f 2
Assumes power of IM3 1 = IM3 2
2 ´ f1 – f2
2 ´ f2 – f1
Frequency (Hz)
Figure 5. IMD3
3
DAC Nyquist Zones, Zero Order Hold, and Images
The terms Nyquist zones, Zero Order Hold and Images are some basic DAC terms. Once understood,
many of the advanced DAC concepts discussed later will follow naturally. The following paragraph asks
some questions whose answers will lead to an understanding of these terms.
Why do the frequency domain plots differ in Figure 6 and Figure 7? The answer lies within the red circle in
Figure 7’s time domain plot. For ease of explanation, there are 2 differences in Figure 7’s Frequency
domain plot that will be discussed. 1) Why do the Fout images exist in every Nyquist zone? (What is a
Nyquist Zone?) 2) Why do the Fout images decrease in amplitude?
Ideal Sine Wave Output Spectrum
Ideal Sine wave - Time domain
Ideal Sine wave - Frequency domain
Fout
Time (s)
Frequency (Hz)
Figure 6. Ideal Sine Wave
Ideal DAC Sine Wave Output Spectrum
Ideal DAC Output - Time domain
Ideal DAC Output - Frequency domain
Fout
Fout Images from Nyquist Zone 1
(amplitude determined by sinc function)
Time (s)
Zone 1
Zone 3
Zone 2
Zone 4
…
Nyquist Zones
Stair stepped
waveform sampled at
DAC clock rate (Fs)
Frequency (Hz)
Fs
2 ´ Fs
Figure 7. Ideal DAC Sine Wave
SLAA523A – March 2012 – Revised October 2012
Submit Documentation Feedback
High Speed, Digital to Analog Converters Basics
Copyright © 2012, Texas Instruments Incorporated
5
DAC Nyquist Zones, Zero Order Hold, and Images
www.ti.com
1) In Figure 7, why do the Fout images exist in every Nyquist zone? What is a Nyquist Zone?
The Poisson summation formula indicates that the samples of function x(t) are sufficient to create a
periodic extension of function X(f). The result is:
¥
X ( f )=
å
X( f - kFs)
(1)
k = -¥
The variable k in Equation 1 causes the Fourier transform output to output a copy of the signal X(f) at
every multiple of Fs. Equation 1’s output is depicted visually in Figure 8 as copies of X(ƒ) are shifted by
multiples of Fs. Figure 8 also provides a visual example of the Nyquist Frequency, Fs/2. Since these
signals are repeated at multiples of Fs, then for bandwidths greater than Fs/2 information loss (red area)
will occur. As shown in Figure 8 multiples of the Nyquist frequency determines the Nyquist Zones. In most
literature, Nyquist Zones are only defined for the positive frequency spectrum.
-Fs/2
-Fs-fBW
-Fs -Fs+fBW
Fs/2
X(f)
-fBW
0
Nyquist Zones
fBW
Fs-fBW
Zone 1
Fs
Zone 2
Fs+fBW
Zone 3
X(f)
-Fs
-fBW
0
fBW
Fs
Figure 8. Hypothetical Spectrum (sampled at Fs)
Using Equation 1 in conjunction with Figure 7, X(f) would be sin(f) and the sample rate is Fs. Using
Equation 1 this expands to
X(f) = sin(f+∞× Fs)…+ sin(f+2×Fs) + sin(f+Fs) + sin(f) + sin(f-Fs) + sin(f-2×Fs) ...+ sin(f-∞× Fs).
In Figure 7, what is not shown is the –Fout plot of the spectrum. Adding Fs to Fout and –Fout results in the
tones shown in Nyquist Zones 2 and 3. Likewise adding 2×Fs to Fout and –Fout results in the tones shown in
Nyquist Zones 4 and 5. This explains why the Fout exist in every Nyquist zone in Figure 7.
2) In Figure 8, why do the Fout images decrease in amplitude?
The answer is due to frequency response of the zero-order hold function. In Figure 9, a zero order hold
function is shown in the time domain and frequency domain. The frequency domain plot shown in Figure 9
is also the sinc function (sin(x)/x). The sinc function has a decreasing amplitude with frequency due to the
denominator x.
6
High Speed, Digital to Analog Converters Basics
SLAA523A – March 2012 – Revised October 2012
Submit Documentation Feedback
Copyright © 2012, Texas Instruments Incorporated
Interpolating DAC Topic 1: FIR filters
www.ti.com
0
Response - dB
-10
1
Time
-Tsymbol/2
0
-20
-30
-40
Tsymbol /2
Time Domain
-50
-60
0
1
2
3
4
5
Frequency Domain
Figure 9. Zero Order Hold Time and Frequency Domain plots
The DACs sampling nature creates the analog output to have a zero order hold function at each sample.
This is shown pictorially in Figure 10 and Figure 11. The math behind these concepts is described in many
textbooks and is a very lengthy topic. The decreasing amplitude of the Fout images in Figure 7, and
Figure 11 are determined by the sinc function of the zero order hold as a result.
Sampled at Fs
1
and
Time
-Tsymbol/2
0
Tsymbol /2
Figure 10. Convolution in the Time Domain
0
Sampled at Fs
-10
and
-20
-30
-40
-50
-60
0
1
2
3
4
5
Figure 11. Convolution in the Frequency Domain
4
Interpolating DAC Topic 1: FIR filters
To understand the need for FIR filters in interpolating DACs, the design constraints of high speed DACs
should be understood initially. Consider the following three cases in Table 1. Spend some time to
understand the concerns and comments in each of the 3 cases.
SLAA523A – March 2012 – Revised October 2012
Submit Documentation Feedback
High Speed, Digital to Analog Converters Basics
Copyright © 2012, Texas Instruments Incorporated
7
Interpolating DAC Topic 1: FIR filters
www.ti.com
Table 1. DAC Case Study
Case 1:
Customer Requirements:
Fs = 250Msps
Fout = 20MHz
SFDR = 80dBc
text
Concerns/Comments:
None
By using a Low-Pass Filter (LPF),
the amplitude of the images will
be below 80 dBc SFDR
LPF
….
Zone 1
Zone 3
Zone 2
…
Zone 4
Frequency (Hz)
Fs
2 ´ Fs
Case 2:
Customer Requirements:
Fs = 250Msps
Fout = 100MHz ← change from case 1
SFDR = 80 dBc
text
Concerns/Comments:
Tough LPF requirements due to image amplitude and
frequency spacing.
Sinc Function:
– Amplitude loss in Fout (SNR degraded)
– Larger Image at Fs-Fout
If this LPF exist it will be more expensive than the one
used in Case 1
By using a Low-Pass Filter (LPF),
the amplitude of the images will
be below 80 dBc SFDR
LPF
….
Zone 1
Zone 3
Zone 2
…
Zone 4
Frequency (Hz)
Fs
2 ´ Fs
Case 3:
Customer Requirements:
Fs = 500Msps ← change from case 2
Fout = 100MHz SFDR = 80dBc
text
Concerns/Comments:
A higher sample rate solves LPF and amplitude (SNR)
concerns in Case 2.
Higher DAC sample rates forces the DAC to have an
LVDS interface
– complexity, 2× of signals/pins
– cost: to achieve higher clock and data rates a higher
cost ASIC/FPGA is required
Fout
LPF
By using a Low-Pass Filter (LPF),
the amplitude of the images will
be below 80 dBc SFDR
….
Zone 1
…
Zone 2
Frequency (Hz)
Fs/2
Fs
In Table 1, cases 2 and 3 illustrate that as Fout frequency increases there are technical trade-offs of SNR,
low pass filtering requirements, and data rate requirements. These trade-offs can increase the cost and
complexity of the system design.
The FIR filters in interpolating DACs were designed to address the trade-offs shown in Table 1. This basic
idea is to provide a slower data rate to the DAC inputs and oversample the data inside the interpolating
DAC. Interpolating DACs have the slower data rate benefit of Table 1’s case 2 and the LPF and SNR
benefits of Table 1’s case 3. When compared to a standard DAC the trade-offs are increased cost, a
slightly larger package footprint, and possibly a more complicated clock scheme.
8
High Speed, Digital to Analog Converters Basics
SLAA523A – March 2012 – Revised October 2012
Submit Documentation Feedback
Copyright © 2012, Texas Instruments Incorporated
Interpolating DAC Topic 1: FIR filters
www.ti.com
8x interpolation block diagram
Figure 12. DAC5688 Block Diagram – FIR filters
Interpolating DACs often have capabilities to perform 8x or 16x interpolation by daisy chaining multiple
FIR filters as seen in Figure 12. Figure 13 is a pictorial example of how one FIR filter in an interpolating
DAC functions. Figure 13 is a simple example of one FIR performing 2x interpolation of the input signal.
SLAA523A – March 2012 – Revised October 2012
Submit Documentation Feedback
High Speed, Digital to Analog Converters Basics
Copyright © 2012, Texas Instruments Incorporated
9
Interpolating DAC Topic 1: FIR filters
www.ti.com
Stet 1: Samples sent to DAC inputs
Samples: 8192, 13984, 16383, 13784, 8192, 2399, 8192
16384
14336
12288
10240
8192
6144
….
4096
2048
0
0
2
4
6
8
10
12
14
16
Zone 1
Zone 2
…
Zone 4
Zone 3
Fs
2 ´ Fs
Stet 2: Internally the DAC inserts 0’s between each sample (2x oversampling - zero stuffing)
Samples: 8192, 0, 13984, 0, 16383, 0, 13784, 0, 8192, 0, 2399, 0, 0, 0, 2399, 0, 8192
Zero stuffing causes
- a 6dB loss in signal amplitude
16384
14336
12288
10240
8192
….
6144
4096
2048
0
0
2
4
6
8
10
12
14
16
Zone 1
…
Zone 2
Fs
Fs/2
Stet 3: Using samples from Step 2. The DAC FIR filter creates following samples
Samples: 8192, 11327, 13984, 15760, 16383, 15760, 13784, 11326,
8192, 1557, 2399, 623, 2399, 5057, 8192
Interpolation improves signal
amplitude to near original
16384
14336
12288
10240
8192
6144
….
4096
2048
0
0
2
4
6
8
10
12
14
16
Zone 1
…
Zone 2
Fs/2
Fs
Figure 13. Interpolation Example
A DACs datasheet will provide the FIR (Finite Impulse Response) filter shapes and specifications.
Figure 14 provides an example of some plots that can be found in datasheets.
10
High Speed, Digital to Analog Converters Basics
SLAA523A – March 2012 – Revised October 2012
Submit Documentation Feedback
Copyright © 2012, Texas Instruments Incorporated
Interpolating DAC Topic 1: FIR filters
www.ti.com
Figure 14. Single FIR Filter and a Composite FIR Filter Spectrum
Interpolation does create a new design consideration. After the FIR filters the DAC will need to be clocked
at a multiple of the data rate. This implies there would be two clocks, one for the data and one for the
DAC. For instance in a 2x interpolation mode as shown in Figure 13, the DAC will require a clock 2x the
speed of the input data rate. Similar statements can be made about 4x, 8x, and 16x interpolation modes.
Most interpolating DACs provide 3 options to solve this clocking concern. Each option has a
cost/complexity vs. performance trade-off.
Figure 15. DAC5688 Block Diagram – Clock Section
Option 1: PLL Clock Mode
The PLL clock mode allows for the user to input a reference clock at the data rate. The internal VCO/PLL
will than generate the higher frequency DAC clock from the reference clock. This mode reduces system
cost and complexity by allowing the designer to use the DAC without the need for a higher speed clock.
However, often the PLL/VCO option generates more phase noise than an external clock. This added
phase noise will affect the DACs SNR and SFDR performance.
SLAA523A – March 2012 – Revised October 2012
Submit Documentation Feedback
High Speed, Digital to Analog Converters Basics
Copyright © 2012, Texas Instruments Incorporated
11
Interpolating DAC Topic 2: Digital Mixing
www.ti.com
Option 2: External Clock Mode
A high speed clock is applied to the DAC at the DAC sample rate. The data sample rate is divided down
from this high speed clock. Compared to option 3 this reduces system cost and complexity as only one
clock is needed. This method also has good clock phase noise. If multiple DACs are used in a system,
then dividing the high speed DAC clock to the lower speed data clock rate can create some phase
uncertainty between multiple DACs, as it is not certain that all the DAC clock dividers are triggering on the
same clock edge. This may be a problem in some system designs.
Option 3: Dual Clock Mode
One clock is supplied at the DAC speed and another clock is supplied at the data rate. When used with
multiple DACs this method solves the phase uncertainty problem with Option 2. This method has good
clock phase noise. The method adds cost to the design when compared with the other methods.
5
Interpolating DAC Topic 2: Digital Mixing
In many system designs the digital mixing function is used to shift a signal to another frequency for
channel selection or to improve image filtering requirements. Often this function of digital mixing is
performed with an external digital chip, either a DSP or FPGA. Most interpolating DACs include this
function in their digital logic section. This helps reduce the complexity of the digital baseband design by
moving this functionality inside the DAC.
Figure 16. DAC5688 Block Diagram – Digital Mixing Section
Figure 16 provides the block diagram of the digital mixing section. In this digital section signals from the
FIR filters are mixed with a digital sine wave from the Numerically Controlled Oscillator (NCO). The output
of the mixer will shift the incoming signal to another frequency based on the NCO’s frequency.
NCO
The NCO creates a cosine and a sine wave at identical frequencies. The frequency of the NCO is
calculated by the DAC clock frequency and the values for the NCO programmed in the serial interface.
Equation 2 provides an example of an equation used to set the NCO output frequency. For a given DAC it
is best to refer to the datasheet for the exact equation for that DAC. In the Equation 2 example the
denominator is 232. The denominator sets the frequency resolution of the NCO:
fNCO = freq x fNCO_CLK /232 (
)
(2)
Where:
fNCO: output frequency of NCO
freq: 32bit frequency word programmed by serial interface
12
High Speed, Digital to Analog Converters Basics
SLAA523A – March 2012 – Revised October 2012
Submit Documentation Feedback
Copyright © 2012, Texas Instruments Incorporated
Interpolating DAC Topic 2: Digital Mixing
www.ti.com
fNCO_CLK: clock frequency supplied to NCO
Digital Mixer (Digital Up Converter – DUC)
The signals from the NCO and FIR filter or Data inputs are then fed into the digital mixer.
The Digital mixer can be used in one of three modes. The three modes are:
1. Complex IQ In and 1 Real Out
2. Complex IQ In and Complex IQ Out
3. 2 Real In and 2 Real Out.
The following will describe how the digital mixer works. Initially mode 1, Complex IQ In and 1 Real Out
will be analyzed to understand the mixer behavior.
IIN (t)
IIN(t)*cos(2*pi*f NCO*t)
The mixer
multiplies the
inputs
cos(2*pi*f NCO*t)
IIN(t)*cos(2*pi*fNCO*t) Q IN(t)*sin(2*pi*fNCO*t)
+
-
QIN(t)
Q IN(t)*sin(2*pi*fNCO*t)
sin(2*pi*f NCO*t)
Input:
From FIR filters
Input:
From NCO
Output:
To DAC A or
Figure 17. Digital Mixer, Complex IQ In and 1 Real Out
At this point the math is a little cumbersome at the digital mixer output in Figure 17. Using trigonometry
identities, and by making an assumption, the output equation can be simplified. The assumption is that
input signals IIN and QIN are the same signal, but are 90 degrees out of phase. The reason this assumption
is made is that this is how most customers will use this DAC function in their application. Also, for ease of
illustration IIN and QIN will be single frequency signals (sine wave and cosine waves). The result of this
assumption is shown in Figure 18.
SLAA523A – March 2012 – Revised October 2012
Submit Documentation Feedback
High Speed, Digital to Analog Converters Basics
Copyright © 2012, Texas Instruments Incorporated
13
Interpolating DAC Topic 2: Digital Mixing
cos(2*pi*f 1t)
www.ti.com
Use Trigonometry identities
to simplify Output equation –
see figure 21
cos(2*pi*f1 *t)*cos(2*pi*fNCO *t)
cos(2*pi*f NCO*t)
+
-Ex1) sin(2*pi*f 1t)
Ex1)
cos(2*pi*f 1t)*cos(2*pi*f NCO *t) sin(2*pi*f1t)*sin(2*pi*f NCO *t)
Ex2)
Ex1)sin(2*pi*f 1*t)*sin(2*pi*fNCO *t)
Ex2) -sin(2*pi*f1t)
sin(2*pi*f NCO*t)
Input:
From FIR filters
Input:
From NCO
Output:
To DAC A or
Figure 18. Digital Mixer, Complex IQ In and 1 Real Out (2 examples)
At this point the math is still cumbersome, but it allows for some simplifications using trigonometry
identities shown below. Figure 19 shows the result of this simplification.
Product to Sum Identities
cos(Ф) × cos(ψ) = 1/2 × [cos(Ф-ψ) + cos(Ф+ψ)]
sin(Ф) × sin(ψ) = 1/2 × [cos(Ф-ψ) - cos(Ф+ψ)]
sin(Ф) × cos(ψ) = 1/2 × [sin(Ф+ψ) + sin(Ф-ψ)]
cos(Ф) × sin(ψ) = 1/2 × [sin(Ф+ψ) - sin(Ф-ψ)]
14
High Speed, Digital to Analog Converters Basics
SLAA523A – March 2012 – Revised October 2012
Submit Documentation Feedback
Copyright © 2012, Texas Instruments Incorporated
Interpolating DAC Topic 2: Digital Mixing
www.ti.com
cos(2*pi*f 1t)
½*[cos(2*pi*(f NCO -f1)*t) + cos(2*pi*(fNCO +f1)*t)]
Real Out
cos(2*pi*fNCO*t)
Complex In
+
--
Ex1) cos(2*pi*(fNCO+ f1)t)
Ex2) cos(2*pi*(fNCO- f1)t)
Ex1) sin(2*pi*f1t)
Ex1) ½*[cos(2*pi*(fNCO -f1)*t) - cos(2*pi*(fNCO +f1 )*t)]
Ex2) -sin(2*pi*f1t)
Ex2) -½*[cos(2*pi*(fNCO -f1)*t) + cos(2*pi*(fNCO +f1)*t)]
sin(2*pi*fNCO*t)
Input:
From FIR filters
or data inputs
Input:
From NCO
Output:
To DAC A or
correction functions
Mixing Results in the Frequency Domain
Complex
Spectrums
Real
Spectrums
X
f1
Example Ex1
When IIN (t) = Q IN (t-90degrees)
fNCO
-(fNCO+f1)
fNCO+f1
X
f1
Example Ex2
When IIN (t) = Q IN (t+90degrees)
fNCO
-(fNCO-f1) fNCO-f1
Figure 19. Digital Mixer, Complex IQ In and 1 Real out
The output in Figure 19 produces a very simple formula for this digital mixer operation. Notice the real
output is simply a frequency shift of the complex input signal. Whether IIN and QIN are –90 or +90 degrees
out of phase, would determine whether one would add or subtract the NCO and Input frequencies (fNCO ±
f1). This operation is shown in the frequency domain plots in Figure 19.
For wideband signals the same concept of adding or subtracting the wideband signal from the NCO
frequency will result in a frequency shift. The wideband signal example is shown in Figure 20.
SLAA523A – March 2012 – Revised October 2012
Submit Documentation Feedback
High Speed, Digital to Analog Converters Basics
Copyright © 2012, Texas Instruments Incorporated
15
Interpolating DAC Topic 2: Digital Mixing
www.ti.com
Wideband signal
centered at @ f1
- 90 degrees
out of phase
cos(2*pi*fNCO*t)
Wideband signal
centered @ (fNCO + f1)
+
--
Wideband signal
centered at @ f1
sin(2*pi*fNCO*t)
Input:
From FIR filters
or data inputs
Input:
From NCO
Output:
To DAC A or
correction functions
Mixing Results in the Frequency Domain
X
f1
fNCO
-(fNCO+f1)
fNCO + f1
Figure 20. Digital Mixer, Complex IQ In and 1 Real out – Wideband example
Figure 19 and Figure 20 also begin to hint at one of the weaknesses of the complex in to real out
architecture. All real signals have a positive and a negative frequency that are equal in amplitude. This
weakness is seen in with a radio architecture where the DAC output needs to be up converted to an RF
signal by an analog mixer. Notice the analog mixer up converts both the positive and negative IF
frequencies. This results in duplicate information since two different RF±IF frequencies (sidebands) are
created. Duplicate information has the results of increasing the noise floor by 2x and decreasing the
mixer’s bandwidth by a half. The end result of this is that BPF #2 is used to remove one of the two signals.
In Figure 21, there is a term labeled LO feed through. LO feed through is a common issue with analog
mixers. With a real signal a BPF is used to remove the LO feed through. Real signals do not allow for the
option of digital correction techniques to remove LO feed through.
16
High Speed, Digital to Analog Converters Basics
SLAA523A – March 2012 – Revised October 2012
Submit Documentation Feedback
Copyright © 2012, Texas Instruments Incorporated
Interpolating DAC Topic 2: Digital Mixing
www.ti.com
Digital
LPF
Analog
I*cos(2 pf IFt)
I
BPF
1
Encoder:
Input
Data
Symbol
Mapping
DAC
S
cos(2 pf IFt)
BPF
2
LPF
Q
sin(2 pf RFt)
Q*sin(2 pf IFt)
sin(2 pf IF t)
LO feedthrough
Upconverted
Image
Image
-f IF
0
f IF
fRF -f IF
BPF
fRF +f IF
Figure 21. Real Output System Architecture
Mode 2, Complex IQ In and Complex IQ Out is a fairly similar concept in the digital section, with the one
significant exception being that 2 digital mixers are used. The outputs of the two digital mixers create a
complex IQ signal. In theory this complex IQ out signal solves the weakness of the Complex IQ In to Real
Out architecture in Figure 21. In practice these theoretical improvements were not possible until recent
advances in DSP correction algorithms that are more readily available today. A block diagram of mode 2
is shown in Figure 22.
Ain [15 :0]
X
Interp
Aout
COS
+
Bin [15:0]
DAC
SIN
-
X
Interp
2 nd Digital
Mixer allows
for complex
IQ out
X
Bout
-SIN
-COS
+
DAC
X
Figure 22. Digital Mixer, Complex IQ In and Complex IQ Out
Figure 22 shows a typical Complex IQ In to Complex IQ Out application schematic. In Figure 23 the
Complex IQ Out signal is up converted to a real signal by an analog mixer. Also notice this architecture
has the ability to program a –NCO frequency. This is another advantage of this mode over the Complex
IQ in to Real Out mode.
SLAA523A – March 2012 – Revised October 2012
Submit Documentation Feedback
High Speed, Digital to Analog Converters Basics
Copyright © 2012, Texas Instruments Incorporated
17
Interpolating DAC Topic 2: Digital Mixing
www.ti.com
Digital
cos(wct)
Analog
LO
I
sin(wc t)
This diagram uses the same
Trigonometry functions as
Figure 18-20. However, for
this example only the results
are shown in the table
below.
DAC A
Q
-sin(wct)
90
-cos(wct)
DAC B
Figure 23. Digital Complex IQ In, Digital Complex IQ Out, With an Analog Mixer
Analog mixers inherently have some errors that cause LO feed through or sideband issues. As mentioned
previously, in theory the Complex IQ In, Complex IQ Out mode can correct for these issues by adjusting
the DACs phase and offset values. This will be discussed in more detail in the Quadrature Modulator
Correction (QMC) section.
Mode 3, 2 Real In and 2 Real Out, takes advantage of some unique properties of the mixer and
interpolation filters. This mode can only be used when the NCO is set to half of the DAC clock speed. By
running the NCO at this rate the cos function of the NCO will output { ..-1,1,-1,1,….}. Likewise the sin
function of the NCO will output {…0,0,0,0…}. The result of this the sin function will cancel out the signal
and the resulting block diagram is shown in Figure 24.
Ain[15:0]
Aout
X
Interp
DAC
FDAC/2
FDAC/2
Bin[15:0]
X
Interp
Bout
DAC
Figure 24. Digital Mixer, 2 Real In and 2 Real Out
18
High Speed, Digital to Analog Converters Basics
SLAA523A – March 2012 – Revised October 2012
Submit Documentation Feedback
Copyright © 2012, Texas Instruments Incorporated
Interpolating DAC Topic 3: Quadrature Modulator Correction (QMC)
www.ti.com
6
Interpolating DAC Topic 3: Quadrature Modulator Correction (QMC)
To understand the need for the QMC section of the interpolating DACs it is best to understand that in
many applications the DAC outputs are connected to the input of a quadrature modulator. The path
between the DAC and the modulator itself have DC, gain and phase errors. As seen in Figure 25 these
DC, gain and phase errors create spectral errors. In Figure 25, it can be noticed that 2 unwanted spurs
are created in the RF spectrum due to these errors. The unwanted carrier spur (FLO) is a result of the DC
error mixing with the LO frequency. The unwanted side band spur (FSB) is a result of the gain and phase
errors.
Analog Quadrature Modulator
IIN(t)+
GAIN ERROR
DC ERROR
IIN(t)-
45
LOIN
-45
QIN(t)+
+
--
PHASE ERROR
RFOUT
GAIN ERROR
DC ERROR
QIN(t)-
RF Output of Analog Quadrature Modulator
FRF = FLO + FIN (desired RFOUT signal)
FSB = FLO – FIN
FLO = LO Input Frequency
FIN = IIN & QIN Input Frequency
Result of DC ERROR
Result of GAIN and
PHASE ERROR
FSB
FLO
FRF
Figure 25. Quadrature Modulator, Common Source of Errors
The goal of the QMC section of a DAC is to manipulate the DC, phase and gain of the DAC signal in such
a way that it offsets the errors seen internal to and at the inputs of the quadrature modulator. In an ideal
case QMC would remove the spurs at FLO and FSB in Figure 25. In practice, the QMC section has been
seen to improve the carrier suppression and sideband suppression to about –70dBc.
Figure 26 shows a typical block diagram of the QMC section of a DAC. As shown in Figure 26, the QMC
section provides the DAC an easy way to force offset, gain and phase errors that cancel out the errors at
the quadrature modulator. Once the DC, gain and phase errors are known, programming the DAC is
straightforward.
SLAA523A – March 2012 – Revised October 2012
Submit Documentation Feedback
High Speed, Digital to Analog Converters Basics
Copyright © 2012, Texas Instruments Incorporated
19
Interpolating DAC Topic 3: Quadrature Modulator Correction (QMC)
QMC Gain and Phase Correction
www.ti.com
QMC Offset - DC Correction
Figure 26. DAC5688 Block Diagram – QMC Functions
These DC, offset and phase errors can drift with phase, temperature and time. In an ideal case the
application would have a real time measurement system that could measure these errors and program the
DAC to correct these errors. However, a system like this adds complexity and cost to the system design.
At the time of this document, most end users manually calibrated their systems QMC values at time zero
and 25°C only. A few end users calibrated their systems QMC values at time zero and several
temperatures. The values for QMC correction values are then stored in an EEPROM and written to the
DAC via DSP or FPGA when the system is in the field.
20
High Speed, Digital to Analog Converters Basics
SLAA523A – March 2012 – Revised October 2012
Submit Documentation Feedback
Copyright © 2012, Texas Instruments Incorporated
IMPORTANT NOTICE
Texas Instruments Incorporated and its subsidiaries (TI) reserve the right to make corrections, enhancements, improvements and other
changes to its semiconductor products and services per JESD46, latest issue, and to discontinue any product or service per JESD48, latest
issue. Buyers should obtain the latest relevant information before placing orders and should verify that such information is current and
complete. All semiconductor products (also referred to herein as “components”) are sold subject to TI’s terms and conditions of sale
supplied at the time of order acknowledgment.
TI warrants performance of its components to the specifications applicable at the time of sale, in accordance with the warranty in TI’s terms
and conditions of sale of semiconductor products. Testing and other quality control techniques are used to the extent TI deems necessary
to support this warranty. Except where mandated by applicable law, testing of all parameters of each component is not necessarily
performed.
TI assumes no liability for applications assistance or the design of Buyers’ products. Buyers are responsible for their products and
applications using TI components. To minimize the risks associated with Buyers’ products and applications, Buyers should provide
adequate design and operating safeguards.
TI does not warrant or represent that any license, either express or implied, is granted under any patent right, copyright, mask work right, or
other intellectual property right relating to any combination, machine, or process in which TI components or services are used. Information
published by TI regarding third-party products or services does not constitute a license to use such products or services or a warranty or
endorsement thereof. Use of such information may require a license from a third party under the patents or other intellectual property of the
third party, or a license from TI under the patents or other intellectual property of TI.
Reproduction of significant portions of TI information in TI data books or data sheets is permissible only if reproduction is without alteration
and is accompanied by all associated warranties, conditions, limitations, and notices. TI is not responsible or liable for such altered
documentation. Information of third parties may be subject to additional restrictions.
Resale of TI components or services with statements different from or beyond the parameters stated by TI for that component or service
voids all express and any implied warranties for the associated TI component or service and is an unfair and deceptive business practice.
TI is not responsible or liable for any such statements.
Buyer acknowledges and agrees that it is solely responsible for compliance with all legal, regulatory and safety-related requirements
concerning its products, and any use of TI components in its applications, notwithstanding any applications-related information or support
that may be provided by TI. Buyer represents and agrees that it has all the necessary expertise to create and implement safeguards which
anticipate dangerous consequences of failures, monitor failures and their consequences, lessen the likelihood of failures that might cause
harm and take appropriate remedial actions. Buyer will fully indemnify TI and its representatives against any damages arising out of the use
of any TI components in safety-critical applications.
In some cases, TI components may be promoted specifically to facilitate safety-related applications. With such components, TI’s goal is to
help enable customers to design and create their own end-product solutions that meet applicable functional safety standards and
requirements. Nonetheless, such components are subject to these terms.
No TI components are authorized for use in FDA Class III (or similar life-critical medical equipment) unless authorized officers of the parties
have executed a special agreement specifically governing such use.
Only those TI components which TI has specifically designated as military grade or “enhanced plastic” are designed and intended for use in
military/aerospace applications or environments. Buyer acknowledges and agrees that any military or aerospace use of TI components
which have not been so designated is solely at the Buyer's risk, and that Buyer is solely responsible for compliance with all legal and
regulatory requirements in connection with such use.
TI has specifically designated certain components which meet ISO/TS16949 requirements, mainly for automotive use. Components which
have not been so designated are neither designed nor intended for automotive use; and TI will not be responsible for any failure of such
components to meet such requirements.
Products
Applications
Audio
www.ti.com/audio
Automotive and Transportation
www.ti.com/automotive
Amplifiers
amplifier.ti.com
Communications and Telecom
www.ti.com/communications
Data Converters
dataconverter.ti.com
Computers and Peripherals
www.ti.com/computers
DLP® Products
www.dlp.com
Consumer Electronics
www.ti.com/consumer-apps
DSP
dsp.ti.com
Energy and Lighting
www.ti.com/energy
Clocks and Timers
www.ti.com/clocks
Industrial
www.ti.com/industrial
Interface
interface.ti.com
Medical
www.ti.com/medical
Logic
logic.ti.com
Security
www.ti.com/security
Power Mgmt
power.ti.com
Space, Avionics and Defense
www.ti.com/space-avionics-defense
Microcontrollers
microcontroller.ti.com
Video and Imaging
www.ti.com/video
RFID
www.ti-rfid.com
OMAP Applications Processors
www.ti.com/omap
TI E2E Community
e2e.ti.com
Wireless Connectivity
www.ti.com/wirelessconnectivity
Mailing Address: Texas Instruments, Post Office Box 655303, Dallas, Texas 75265
Copyright © 2012, Texas Instruments Incorporated