Script Adc
Script Adc
On-Board
MCLK Reference
Regulator
FullScale
Control COMP
VCC
FSELECT 2.5V
28 Bit
FREQ0 REG Phase 12 IOUT
10-Bit
Accumulator SIN MUX
MUX DAC
(28 Bit) ROM IOUTB
28 Bit
FREQ1 REG
MSB
MUX DIV BY
12 Bit PHASE0 REG 2
MUX
12 Bit PHASE1 REG
MUX
SIGN BIT OUT
16 Bit Control
Register
DDS SPECIFICATIONS
Dynamic Specifications:
Signal to Noise Ratio 50 dB fMCLK = 50 MHz, fOUT = fMCLK/4096
Total Harmonic Distortion -53 dBc fMCLK = 50 MHz, fOUT = fMCLK/4096
Spurious Free Dynamic Range (SFDR):
Wideband (0 to Nyquist) 50 dBc fMCLK = 50 MHz, fOUT = fMCLK/7
NarrowBand (± 200 kHz) 72 dBc fMCLK = 50 MHz, fOUT = fMCLK/7
Clock Feedthrough –55 dBc
Wake Up Time 1 ms
COMPARATOR
Input Voltage Range 1 V p-p ac-coupled internally
Input Capacitance 10 pF
Input HighPass Cutoff Frequency 4 MHz
Input DC Resistance 1 MΩ
Input DC Current 10 µA
OUTPUT BUFFER
Output Rise/Fall Time 20 ns Using a 15 pF Load
Output Jitter 100 ps rms When DAC data MSB is output
VOLTAGE REFERENCE
Internal Reference 1.116 1.2 1.284 V 1.2 V ± 7%
REFOUT Input Impedance3 1 KΩ
Reference TC 100 ppm/°C
LOGIC INPUTS
VINH, Input High Voltage DVDD –0.9 V +3.6 V to +5.5 V Power Supply
DVDD - 0.5 V +2.7 V to +3.6 V Power Supply
2 V +2.3 V to + 2.7 V Power Supply
VINL, Input Low Voltage 0.9 V +3.6 V to +5.5 V Power Supply
0.5 V +2.3 V to + 3.6 V Power Supply
IINH, Input Current 1 µA
CIN, Input Capacitance 10 pF
REV PrM
–2–
PRELIMINARY TECHNICAL DATA
AD9834
RSET
6.8 K
100nF 10nF
CAP/2.5V REFOUT FS
ADJUST AVDD
10nF
ON-BOARD FULL-SCALE COMP
REGULATOR
REFERENCE CONTROL
12
SIN IOUT
10-BIT DAC
ROM
200R 20pF
AD9834
TIMING CHARACTERISTICS1 (VDD = +2.3 V to +5.5 V; AGND = DGND = 0 V, unless otherwise noted)
t1
MCLK MCLK
t2 t 11 t 11A
t3
FSELECT, VALID DATA VALID DATA VALID DATA
PSELECT
t5 t4
SCLK
t7 t6 t8
FSYNC
t10
t9
REV PrM
–3–
PRELIMINARY TECHNICAL DATA
AD9834
ABSOLUTE MAXIMUM RATINGS* Storage Temperature Range . . . . . . . . . –65°C to +150°C
(TA = +25°C unless otherwise noted) Maximum Junction Temperature . . . . . . . . . . . . . +150°C
AVDD to AGND . . . . . . . . . . . . . . . . . . . . –0.3 V to +6 V TSSOP Package
DVDD to DGND . . . . . . . . . . . . . . . . . . . . –0.3 V to +6 V θJA Thermal Impedance . . . . . . . . . . . . . . . . . . . 143°C/W
AVDD to DVDD . . . . . . . . . . . . . . . . . . –0.3 V to +0.3 V θJC Thermal Impedance . . . . . . . . . . . . . . . . . . . 45°C/W
AGND to DGND. . . . . . . . . . . . . . . . . . –0.3 V to +0.3 V Lead Temperature, Soldering (10 sec) . . . . . . . . . 300°C
CAP/2.5V . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.75 V IR Reflow, Peak Temperature . . . . . . . . . . . . . . . 220°C
Digital I/O Voltage to DGND –0.3 V to DVDD + 0.3 V *Stresses above those listed under “Absolute Maximum Ratings” may cause permanent
Analog I/O Voltage to AGND –0.3 V to AVDD + 0.3 V damage to the device. This is a stress rating only and functional operation of the device
at these or any other conditions above those listed in the operational sections of this
Operating Temperature Range specification is not implied. Exposure to absolute maximum rating conditions for
Industrial (B Version) . . . . . . . . . . . . . . –40°C to +85°C extended periods may affect device reliability.
ORDERING GUIDE
PIN CONFIGURATION
FS ADJUST 1 20 IOUTB
REFOUT 2 AD9834 19 IOUT
COMP 3 18 AGND
AVDD 4 O 17 VIN
TOP VIEW
DVDD 5 (Not to Scale) 16 SIGNBITOUT
CAP/+2.5V 6 15 FSYNC
DGND 7 14 SCLK
MCLK 8 13 SDATA
FSELECT 9 12 SLEEP
PSELECT 10 11 RESET
CAUTION
ESD (electrostatic discharge) sensitive device. Electrostatic charges as high as 4000 V readily
accumulate on the human body and test equipment and can discharge without detection.
Although the AD9834 features proprietary ESD protection circuitry, permanent damage may
occur on devices subjected to high energy electrostatic discharges. Therefore, proper ESD
precautions are recommended to avoid performance degradation or loss of functionality.
REV PrM
–4–
PRELIMINARY TECHNICAL DATA
AD9834
PIN FUNCTIONS DESCRIPTIONS
REV PrM
–5–
PRELIMINARY TECHNICAL DATA
AD9834
Typical Performance Characteristics
TPC 1. Typical Current Consumption TPC 2. Narrow Band SFDR vs. MCLK TPC 3. Wide Band SFDR vs. MCLK
vs. MCLK Frequency Frequency Frequency
TPC 4. Wide Band SFDR vs. fOUT/fMCLK TPC 5. SNR vs. MCLK Frequency TPC 6. SNR vs. fOUT/fMCLK for
for Various MCLK Frequencies Various MCLK Frequencies
REV PrM
–6–
PRELIMINARY TECHNICAL DATA
AD9834
Typical Performance Characteristics
TPC 9. fMCLK = 10 MHz; fOUT = 2.4 kHz; TPC 10. fMCLK = 10 MHz; fOUT = 1.43 kHz TPC 11. fMCLK = 10 MHz; fOUT = 3.33 kHz
Frequency Word = 000FBA9 = fMCLK/7 ; = fMCLK/3 ;
Frequency Word = 2492492 Frequency Word = 5555555
TPC 12. fMCLK = 50 MHz; fOUT = 12 kHz; TPC 13. fMCLK = 50 MHz; fOUT = 120 kHz; TPC 14. fMCLK = 50 MHz; fOUT = 1.2
Frequency Word = 000FBA9 Frequency Word = 009D496 MHz; Frequency Word = 0624DD3
TPC 15. fMCLK = 50 MHz; fOUT = 4.8 TPC 16. fMCLK = 50 MHz; TPC 17. fMCLK = 50 MHz;
MHz; Frequency Word = 189374C fOUT = 7.143 MHz = fMCLK/7 ; fOUT = 16.667 MHz = fMCLK/3 ;
Frequency Word = 2492492 Frequency Word = 5555555
REV PrM
–7–
PRELIMINARY TECHNICAL DATA
AD9834
TERMINOLOGY THEORY OF OPERATION
Integral Nonlinearity Sine waves are typically thought of in terms of their
This is the maximum deviation of any code from a magnitude form a(t) = sin (ωt). However, these are
straight line passing through the endpoints of the transfer nonlinear and not easy to generate except through piece
function. The endpoints of the transfer function are zero wise construction. On the other hand, the angular
scale, a point 0.5 LSB below the first code transition information is linear in nature. That is, the phase angle
(000 . . . 00 to 000 . . . 01) and full scale, a point 0.5 LSB rotates through a fixed angle for each unit of time. The
above the last code transition (111 . . . 10 to 111 . . . 11). angular rate depends on the frequency of the signal by the
The error is expressed in LSBs. traditional rate of ω = 2πf.
Differential Nonlinearity
MAGNITUDE
This is the difference between the measured and ideal 1 +1
LSB change between two adjacent codes in the DAC. A
specified differential nonlinearity of ±1 LSB maximium ensures
0
monotonicity.
Output Compliance
-1
The output compliance refers to the maximum voltage
that can be generated at the output of the DAC to meet PHASE
2π
the specifications. When voltages greater than that speci-
fied for the output compliance are generated, the AD9834
may not meet the specifications listed in the data sheet. 0
REV PrM
–8–
PRELIMINARY TECHNICAL DATA
AD9834
CIRCUIT DESCRIPTION
The AD9834 is a fully integrated Direct Digital Synthesis DAC. This requires the SIN ROM to have two bits of
(DDS) chip. The chip requires one reference clock, one phase resolution more than the 10-bit DAC.
low precision resistor and eight decoupling capacitors to The SIN ROM is enabled using bits MODE and
provide digitally created sine waves up to 25 MHz. In OPBITEN in the control register. This is explained fur-
addition to the generation of this RF signal, the chip is ther in Table 14.
fully capable of a broad range of simple and complex Digital-to-Analog Converter
modulation schemes. These modulation schemes are fully The AD9834 includes a high impedance current source
implemented in the digital domain allowing accurate and 10-bit DAC, capable of driving a wide range of loads.
simple realization of complex modulation algorithms us- Full-scale output current can be adjusted, for optimum
ing DSP techniques. power and external load requirements, through the use of
The internal circuitry of the AD9834 consists of the fol- a single external resistor (RSET).
lowing main sections: a Numerical Controlled Oscillator The DAC can be configured for either single-ended or
(NCO), Frequency and Phase Modulators, SIN ROM, a differential operation. IOUT and IOUTB can be con-
Digital-to-Analog Converter, a Comparator and a nected through equal external resistors to AGND to
Regulator. develop complementary output voltages. The load resis-
Numerical Controlled Oscillator + Phase Modulator tors can be any value required, as long as the full-scale
This consists of two frequency select registers, a phase voltage developed across it does not exceed the voltage
accumulator, two phase offset registers and a phase offset compliance range. Since full-scale current is controlled by
adder. The main component of the NCO is a 28-bit phase RSET, adjustments to RSET can balance changes made to the
accumulator which assembles the phase component of the load resistors.
output signal. Continuous time signals have a phase range Comparator
of 0 to 2. Outside this range of numbers, the sinusoid The AD9834 can be used to generate synthesised digital
functions repeat themselves in a periodic manner. The clock signals. This can be done by using the on-board
digital implementation is no different. The accumulator self-biasing comparator, which converts the DAC's sinu-
simply scales the range of phase numbers into a multibit soidal signal to a square wave. The output from the DAC
digital word. The phase accumulator in the AD9834 is may be filtered externally before being applied to the
implemented with 28 bits. Therefore, in the AD9834, 2 comparator input. The comparator reference voltage is the
= 228. Likewise, the ∆Phase term is scaled into this range time-average of the signal applied to VIN. The comparator
of numbers 0 < ∆Phase < 228 – 1. Making these substitu- can accept a signal of 1 Vpp. As the comparator's input is
tions into the equation above ac-coupled, to operate correctly as a zero crossing
f = ∆Phase x fMCLK/228 dectector, it requires a minimum input frequency of 3
MHz. The comparator's output will be a square wave with
where 0 < ∆Phase < 228 - 1. an amplitude from 0 V to DVDD.
The input to the phase accumulator (i.e., the phase step) To enable the comparator, bits SIGNPIB and OPBITEN
can be selected either from the FREQ0 Register or in the control resister are set to '1'. This is explained fur-
FREQ1 Register and this is controlled by the FSELECT ther in Table 13.
pin or the FSEL bit. NCOs inherently generate Regulator
continuous phase signals, thus avoiding any output The AD9834 has separate power supplies for the analog
discontinuity when switching between frequencies. and digital section. AVDD provides the power supply
Following the NCO, a phase offset can be added to required for the analog section, while DVDD provides the
perform phase modulation using the 12-bit Phase power supply for the digital section. Both of these supplies
Registers. The contents of one of these phase registers is can have a value of +2.3V to +5.5V, and are independant
added to the most significant bits of the NCO. The of each other e.g. the analog section can be operated at 5V
AD9834 has two Phase registers, the resolution of these and the digital section can be operated at 3V or vice versa.
registers being 2π/4096.
SIN ROM The internal digital section of the AD9834 is operated at
To make the output from the NCO useful, it must be 2.5 V. An on-board regulator steps down the voltage ap-
converted from phase information into a sinusoidal value. plied at DVDD to 2.5 V. The digital inteface (serial port)
Since phase information maps directly into amplitude, the of the AD9834 is also operated from DVDD. These digi-
SIN ROM uses the digital phase information as an ad- tal signals are level shifted within the AD9834 to make
dress to a look-up table, and converts the phase them 2.5V compatible.
information into amplitude. Although the NCO contains a When the applied voltage at the DVDD pin of the
28-bit phase accumulator, the output of the NCO is trun- AD9834 is equal to or less than 2.5V, the pins CAP/2.5V
cated to 12 bits. Using the full resolution of the phase and DVDD should be tied together, thus by-passing the
accumulator is impractical and unnecessary as this would on-board regulator.
require a look-up table of 228 entries. It is necessary only
to have sufficient phase resolution such that the errors due
to truncation are smaller than the resolution of the 10-bit
REV PrM
–9–
PRELIMINARY TECHNICAL DATA
AD9834
FUNCTIONAL DESCRIPTION
signal will appear at the DAC output 7 MCLK cycles after
Serial Interface RESET is set to 0.
The AD9834 has a standard 3-wire serial interface, which
is compatible with SPI, QSPI, MICROWIRE and DSP Latency
interface standards. Associated with each operation is a latency. When the pins
Data is loaded into the device as a 16-bit word under the FSELECT and PSELECT change value there is a pipe-
control of a serial clock input, SCLK. The timing dia- line delay before control is transfered to the selected
gram for this operation is given in Figure 4. register. When the timing specifications t11 and t11A are
The FSYNC input is a level triggered input that acts as a met (see figure 3) FSELECT and PSELECT have laten-
frame synchronisation and chip enable. Data can only be cies of 7 MCLK cycles. When the timing specifications
transferred into the device when FSYNC is low. To start t11 and t11A are not met, the latency is increased by one
the serial data transfer, FSYNC should be taken low, ob- MCLK cycle.
serving the minimum FSYNC to SCLK falling edge setup Similarly there is a latency associated with each asynchro-
time, t7. After FSYNC goes low, serial data will be shifted nous write operation. If a selected frequency/phase register
into the device's input shift register on the falling edges of is loaded with a new word there is a delay of 7 to 8 MCLK
SCLK for 16 clock pulses. FSYNC may be taken high cycles before the analog output will change. (There is an
after the sixteenth falling edge of SCLK, observing the uncertainty of one MCLK cycle as it depends on the posi-
minimum SCLK falling edge to FSYNC rising edge time, tion of the MCLK rising edge when the data is loaded into
t8. Alternatively, FSYNC can be kept low for a multiple of the destination register.)
16 SCLK pulses, and then brought high at the end of the The negative transition of the RESET and SLEEP func-
data transfer. In this way, a continuous stream of 16 bit tions are sampled on the internal falling edge of MCLK,
words can be loaded while FSYNC is held low, FSYNC therefore also have a latency associated with them.
only going high after the 16th SCLK falling edge of the
last word loaded. The Control Register
The SCLK can be continuous or, alternatively, the SCLK The AD9834 contains a 16-bit control register which sets
can idle high or low between write operations. up the AD9834 as the user wishes to operate it. All control
Powering up the AD9834 bits, except MODE, are sampled on the internal negative
The flow chart in Figure 7 shows the operating routine for edge of MCLK.
the AD9834. When the AD9834 is powered up, the part Table 2, on the following page, describes the individual
should be reset. This will reset appropriate internal regis- bits of the control register. The different functions and the
ters to zero to provide an analog output of midscale. To various output options from the AD9834 are described in
avoid spurious DAC outputs while the AD9834 is being more detail in the section following Table 2.
initialized, the RESET bit/pin should be set to 1 until the To inform the AD9834 that you wish to alter the contents
part is ready to begin generating an output. RESET does of the Control register, D15 and D14 must be set to '0' as
not reset the phase, frequency or control registers. These shown below.
registers will contain invalid data and, therefore, should be
set to a known value by the user. The RESET bit/pin Table 1. Control Register
should then be set to 0 to begin generating an output. A
D15 D14 D13 D0
0 0 CONTROL BITS
SLEEP12
AD9834
SLEEP1
SIN (Low Power) IOUT
0
Phase ROM MUX
Accumulator 1 10 - Bit DAC IOUTB
(28 Bit)
COMPARATOR VIN
MODE + OPBITEN
1
Div 1 MUX DIGITAL
by 2 MUX 0 OUTPUT SIGN BIT OUT
0 (enable)
DIV2
SIGNPIB
OPBITEN
REV PrM
–10–
PRELIMINARY TECHNICAL DATA
AD9834
Table 2. Description of bits in the Control Register
REV PrM
–11–
PRELIMINARY TECHNICAL DATA
AD9834
The Frequency and Phase Resisters
The AD9834 contains 2 frequency registers and 2 phase The FSELECT and PSELECT pins are sampled on the
registers. These are described in Table 3 below. internal falling edge of MCLK. It is recommended that
the data on these pins does not change within a time win-
Table 3. Frequency/Phase Registers dow of the falling edge of MCLK (see Figure 3 for
timing). If FSELECT/PSELECT changes value when a
Register Size Description falling edge occurs, there is an uncertainty of one MCLK
cycle as to when control is transferred to the other fre-
FREQ0 28 Bits Frequency Register 0. When FSEL quency/phase register.
bit or FSELECT pin = 0, this regis-
ter defines the output frequency as a The flow charts in Figures 8 and 9 show the routine for
fraction of the MCLK frequency. selecting and writing to the frequency and phase registers
FREQ1 28 Bits Frequency Register 1. When FSEL of the AD9834.
bit or FSELECT pin = 1, this regis- Writing to a Frequency Register:
ter defines the output frequency as a When writing to a frequency register, bits D15 and D14
fraction of the MCLK frequency. give the address of the frequency register.
PHASE0 12 Bits Phase Offset Register 0. When PSEL
bit or PSELECT pin = 0, the con- Table 6. Frequency Register Bits
tents of this register are added to the
output of the phase accumulator. D15 D14 D13 D0
PHASE1 12 Bits Phase Offset Register 1. When PSEL 0 1 MSB 14 FREQ0 REG BITS LSB
bit or PSELECT pin = 1, the con- 1 0 MSB 14 FREQ1 REG BITS LSB
tents of this register are added to the
output of the phase accumulator. If the user wishes to alter the entire contents of a fre-
quency register, two consecutive writes to the same
address must be performed, as the frequency registers are
The analog output from the AD9834 is 28 bits wide. The first write will contain the 14 LSBs
fMCLK/228 x FREQREG while the second write will contain the 14 MSBs. For this
where FREQREG is the value loaded into the selected mode of operation, the control bit B28 (D13) should be
frequency register. This signal will be phase shifted by set to 1. An example of a 28-bit write is shown in Table 7
2π/4096 x PHASEREG below.
where PHASEREG is the value contained in the selected
phase register. Table 7: Writing 3FFF0000 to FREQ0 REG
Access to the frequency and phase registers is controlled
by both the FSELECT/PSELECT pins and the FSEL/ SDATA input Result of input word
PSEL control bits. If the control bit PIN/SW = 1, the
0010 0000 0000 0000 Control word write (D15, D14 = 00);
pins controls the function, whereas if PIN/SW = 0, the
B28 (D13) = 1; HLB (D12) = X
bits control the function. This is outlined in tables 4 and 5
0100 0000 0000 0000 FREQ0 REG write (D15, D14 = 01);
below. If the FSEL/PSEL bits are being used, the pins
14 LSBs = 0000
should preferably be held at CMOS logic high or low.
0111 1111 1111 1111 FREQ0 REG write (D15, D14 = 01);
Control of the frequency/phase registers can be inter-
14 MSBs = 3FFF
changed from the pins to the bits.
In some applications, the user does not need to alter all 28
Table 4: Selecting a Frequency Register
bits of the frequency register. With coarse tuning, only
FSELECT FSEL PIN/SW Selected Register the 14 MSBs are altered while with fine tuning, only the
14 LSBs are altered. By setting the control bit B28 (D13)
0 X 1 FREQ0 REG to 0, the 28-bit frequency register operates as 2 14-bit
1 X 1 FREQ1 REG registers, one containing the 14 MSBs and the other con-
X 0 0 FREQ0 REG taining the 14 LSBs. This means that the 14 MSBs of the
X 1 0 FREQ1 REG frequency word can be altered independent of the 14 LSBs
and vice versa. Bit HLB (D12) in the control register
identifies which 14 bits are being altered. Examples of this
Table 5: Selecting a Phase Register are shown over.
PSELECT PSEL PIN/SW Selected Register
0 X 1 PHASE0 REG
1 X 1 PHASE1 REG
X 0 0 PHASE0 REG
X 1 0 PHASE1 REG
REV PrM
–12–
PRELIMINARY TECHNICAL DATA
AD9834
Table 8: Writing 3FFF to the 14 LSBs of FREQ1 REG The Sleep Function
SDATA input Result of input word Sections of the AD9834 which are not in use can be pow-
ered down minimise power consumption. This is done
0000 0000 0000 0000 Control word write (D15, D14 = 00); using the Sleep Function. The parts of the chip that can
B28 (D13) = 0; HLB (D12) = 0, i.e. LSBs be powered down are the Internal clock and the DAC.
1011 1111 1111 1111 FREQ1 REG write (D15, D14 = 10); The DAC can be powered down through hardware or
14 LSBs = 3FFF software. The pin/bits required for the Sleep Function are
outlined in Table 12.
Table 9: Writing 3FFF to the 14 MSBs of FREQ0 REG Table 12: Applying the SLEEP Function
SDATA input Result of Input word SLEEP SLEEP1 SLEEP12 PIN/SW Result
0001 0000 0000 0000 Control word write (D15, D14 = 00); pin bit bit bit
B28 (D13) = 0; HLB (D12) = 1, i.e. MSBs 0 X X 1 No powerdown
0111 1111 1111 1111 FREQ0 REG write (D15, D14 = 01); 1 X X 1 DAC Powered Down
14 MSBs = 3FFF X 0 0 0 No powerdown
X 0 1 0 DAC Powered Down
Writing to a Phase Register: X 1 0 0 Internal Clock disabled
When writing to a phase register, bits D15 and D14 are X 1 1 0 Both the DAC powered
set to 11. Bit D13 identifies which phase register is being down and the Internal
loaded. Clock disabled
Table 10. Phase Register Bits DAC Powered Down: This is useful when the AD9834 is
used to output the MSB of the DAC data only. In this
D15 D14 D13 D12 D11 D0
case, the DAC is not required so it can be powered down
1 1 0 X MSB 12 PHASE0 BITS LSB to reduce power consumption.
1 1 1 X MSB 12 PHASE1 BITS LSB Internal Clock disabled: When the internal clock of the
AD9834 is disabled the DAC output will remain at its
present value as the NCO is no longer accumulating. New
The RESET Function frequency, phase and control words can be written to the
The RESET function resets appropriate internal registers part when the SLEEP1 control bit is active. The
to zero to provide an analog output of midscale. RESET synchronising clock is still active which means that the
does not reset the phase, frequency or control registers. selected frequency and phase registers can also be changed
When the AD9834 is powered up, the part should be re- either at the pins or by using the control bits. Setting the
set. To reset the AD9834, set the RESET pin/bit to 1. To SLEEP1 bit equal to 0 enables the MCLK. Any changes
take the part out of reset, set the pin/bit to 0. A signal will made to the registers while SLEEP1 was active will be
appear at the DAC output 7 MCLK cycles after RESET is seen at the output after a certain latency.
set to 0. The effect of asserting the SLEEP pin is seen immediately
The RESET function is controlled by both the RESET at the output, i.e. the zero to one transition of this pin is
pin and the RESET control bit. If the control bit PIN/SW not sampled. However, the negative transition of SLEEP
= 0, the RESET bit controls the function, whereas if PIN/ is sampled on the internal falling edge of MCLK.
SW = 1, the pin control the function.
The SIGN BIT OUT Pin
Table 11: Applying RESET
The AD9834 offers a variety of outputs from the chip.
RESET pin RESET bit PIN/SW Result The digital outputs are available from the SIGN BIT
OUT pin. The available outputs are the comparator out-
0 X 1 No Reset Applied put or the MSB of the DAC data.
1 X 1 Internal Registers Reset This pin must be enabled before use. The enabling/dis-
X 0 0 No Reset Applied abling of this pin is controlled by the bit OPBITEN (D5)
X 1 0 Internal Registers Reset in the control register. When OPBITEN = 1, this pin is
enabled. Note that the MODE bit (D1) in the control
The effect of asserting the RESET pin is seen immedi- register should be set to '0' if OPBITEN = '1'.
ately at the output, i.e. the zero to one transition of this
pin is not sampled. However, the negative transition of
RESET is sampled on the internal falling edge of MCLK.
REV PrM
–13–
PRELIMINARY TECHNICAL DATA
AD9834
Comparator Output: The AD9834 has an on-board APPLICATIONS
comparator. To connect this comparator to the SIGN Because of the various output options available from the
BIT OUT pin, the SIGNPIB (D4) control bit must be set part, the AD9834 can be configured to suit a wide variety
to 1. After filtering the sinusoidal output from the DAC, of applications.
the waveform can be applied to the comparator to generate One of the areas where the AD9834 is suitable is in modu-
a square waveform. lation applications. The part can be used to perform
MSB from the NCO: The MSB from the NCO can be simple modulation such as FSK. More complex modula-
output from the AD9834. By setting the SIGNPIB (D4) tion schemes such as GMSK and QPSK can also be
control bit to 0, the MSB of the DAC data is available at implemented using the AD9834.
the SIGN BIT OUT pin. This is useful as a coarse clock In an FSK application, the two frequency registers of the
source. This square wave can also be divided by 2 before AD9834 are loaded with different values; one frequency
being output. The bit DIV2 (D3) in the control register will represent the space frequency while the other will
controls the frequency of this output from the SIGN BIT represent the mark frequency. The digital data stream is
OUT pin. fed to the FSELECT pin which will cause the AD9834 to
modulate the carrier frequency between the two values.
Table 13: Various Outputs from SIGN BIT OUT The AD9834 has two phase registers; this enables the part
to perform PSK. With phase shift keying, the carrier fre-
OPBITEN MODE SIGNPIB DIV2 SIGN BIT OUT quency is phase shifted, the phase being altered by an
Bit Bit Bit Bit Pin amount which is related to the bit stream being input to
the modulator.
0 X X X High Impedance
1 0 0 0 DAC data MSB / 2 The AD9834 is also suitable for signal generator applica-
1 0 0 1 DAC data MSB tions. With the on-board comparator, the device can be
1 0 1 0 Reserved used to generate a square wave.
1 0 1 1 Comparator Output With its low current consumption, the part is suitable for
1 1 X X Reserved applications in which it can be used as a local oscillator.
0 0 Sinusoid
0 1 Up/Down Ramp
1 0 Sinusoid
1 1 Reserved
REV PrM
–14–
PRELIMINARY TECHNICAL DATA
AD9834
DATA WRITE
See Figure 9
SELECT DATA
SOURCES
See Figure 10
INITIALISATION
See Figure 8 below
WAIT 7/8 MCLK
CYCLES
See Timing Diagram Fig. 2
DAC OUTPUT
VOUT = VREFOUT * 18 * RLOAD/RSET * (1+ (SIN(2p(FREQREG * FMCLK * t/228 + PHASEREG/212)))
YES
CHANGE PHASE? CHANGE PSEL/ YES
PSELECT?
NO NO
YES CHANGE FSEL/ YES CHANGE PHASE
CHANGE FREQUENCY?
FSELECT? REGISTER?
NO NO YES
CHANGE FREQ
REGISTER? CHANGE DAC OUTPUT
YES FROM SIN TO RAMP?
YES
NO
CONTROL CHANGE OUTPUT AT
REGISTER SIGN BIT OUT PIN?
WRTE? YES
NO
INITIALISATION
APPLY RESET
USING CONTROL USING PIN
BIT
SET RESET = 0
SELECT FREQUENCY REGISTERS
SELECT PHASE REGISTERS
USING CONTROL
USING PIN
BIT
Figure 8. Initialisation
REV PrM
–15–
PRELIMINARY TECHNICAL DATA
AD9834
DATA WRITE
NO
NO WRITE TO PHASE
WRITE A FULL 28-BIT WORD WRITE 14 MSBs OR LSBs
TO A FREQUENCY REGISTER? TO A FREQUENCY REGISTER? REGISTER?
YES YES
YES
(CONTROL REGISTER WRITE)
(CONTROL REGISTER WRITE)
B28 (D13) = 0
B28 (D13) = 1
HLB (D12) = 0 / 1 (16 - Bit Write)
D15, D14 = 11
D13 = 0/1 (chooses the
phase register)
WRITE 2 CONSECUTIVE WRITE A 16-BIT WORD D12 = X
16-BIT WORDS D11 ... D0 = Phase Data
(See Tables 8 & 9 for
(See Table 7 for Example) examples)
WRITE TO ANOTHER
WRITE ANOTHER FULL WRITE 14 MSBs OR LSBs PHASE REGISTER?
YES YES
28 BITS TO A TO A
FREQUENCY REGISTER? FREQUENCY REGISTER? YES
NO
NO NO
YES
FSELECT AND PSELECT SET FSELECT
PINS BEING USED? AND PSELECT
NO
PIN/SW = 0 PIN/SW = 1
SET FSEL Bit
SET PSEL Bit
REV PrM
–16–
PRELIMINARY TECHNICAL DATA
AD9834
GROUNDING AND LAYOUT INTERFACING TO MICROPROCESSORS
The printed circuit board that houses the AD9834 should The AD9834 has a standard serial interface which allows
be designed so that the analog and digital sections are the part to interface directly with several microprocessors.
separated and confined to certain areas of the board. This The device uses an external serial clock to write the data/
facilitates the use of ground planes which can be separated control information into the device. The serial clock can
easily. A minimum etch technique is generally best for have a frequency of 40 MHz maximum. The serial clock
ground planes as it gives the best shielding. Digital and can be continuous or, it can idle high or low between
analog ground planes should only be joined in one place. write operations. When data/control information is being
If the AD9834 is the only device requiring an AGND to written to the AD9834, FSYNC is taken low and is held
DGND connection, then the ground planes should be low while the 16 bits of data are being written into the
connected at the AGND and DGND pins of the AD9834. AD9834. The FSYNC signal frames the 16 bits of infor-
If the AD9834 is in a system where multiple devices re- mation being loaded into the AD9834.
quire AGND to DGND connections, the connection AD9834 to ADSP-21xx Interface
should be made at one point only, a star ground point that
should be established as close as possible to the AD9834. Figure 12 shows the serial interface between the AD9834
and the ADSP-21xx. The ADSP-21xx should be set up to
Avoid running digital lines under the device as these will operate in the SPORT Transmit Alternate Framing Mode
couple noise onto the die. The analog ground plane should (TFSW = 1). The ADSP-21xx is programmed through
be allowed to run under the AD9834 to avoid noise cou- the SPORT control register and should be configured as
pling. The power supply lines to the AD9834 should use follows:
as large a track as is possible to provide low impedance Internal clock operation (ISCLK = 1)
paths and reduce the effects of glitches on the power sup- Active low framing (INVTFS = 1)
ply line. Fast switching signals such as clocks should be 16-bit word length (SLEN = 15)
shielded with digital ground to avoid radiating noise to Internal frame sync signal (ITFS = 1)
other sections of the board. Avoid crossover of digital and Generate a frame sync for each write (TFSR = 1).
analog signals. Traces on opposite sides of the board Transmission is initiated by writing a word to the Tx reg-
should run at right angles to each other. This will reduce ister after the SPORT has been enabled. The data is
the effects of feedthrough through the board. A microstrip clocked out on each rising edge of the serial clock and
technique is by far the best but is not always possible with clocked into the AD9834 on the SCLK falling edge.
a double-sided board. In this technique, the component
side of the board is dedicated to ground planes while sig-
nals are placed on the other side.
Good decoupling is important. The analog and digital
supplies to the AD9834 are independent and separately
pinned out to minimize coupling between analog and digi-
tal sections of the device. All analog and digital supplies
should be decoupled to AGND and DGND respectively
with 0.1 µF ceramic capacitors in parallel with 10 µF
tantalum capacitors. To achieve the best from the
decoupling capacitors, they should be placed as close as
possible to the device, ideally right up against the device.
Figure 11. ADSP2101/ADSP2103 to AD9834 Interface
In systems where a common supply is used to drive both
the AVDD and DVDD of the AD9834, it is recommended AD9834 to 68HC11/68L11 Interface
that the system’s AVDD supply be used. This supply Figure 13 shows the serial interface between the AD9834
should have the recommended analog supply decoupling and the 68HC11/68L11 microcontroller. The
between the AVDD pins of the AD9834 and AGND and microcontroller is configured as the master by setting bit
the recommended digital supply decoupling capacitors MSTR in the SPCR to 1 and, this provides a serial clock
between the DVDD pins and DGND. on SCK while the MOSI output drives the serial data line
SDATA. Since the microcontroller does not have a dedi-
cated frame sync pin, the FSYNC signal is derived from a
port line (PC7). The set up conditions for correct opera-
tion of the interface are as follows:
SCK idles high between write operations (CPOL = 0)
data is valid on the SCK falling edge (CPHA = 1).
When data is being transmitted to the AD9834, the
FSYNC line is taken low (PC7). Serial data from the
68HC11/68L11 is transmitted in 8-bit bytes with only 8
falling clock edges occuring in the transmit cycle. Data is
transmitted MSB first. In order to load data into the
AD9834, PC7 is held low after the first 8 bits are trans-
ferred and a second serial write operation is performed to
the AD9834. Only after the second 8 bits have been trans-
REV PrM
–17–
PRELIMINARY TECHNICAL DATA
AD9834
ferred should FSYNC be taken high again.
being applied to the AD9834. The interface to the
DSP56000/DSP56001 is similar to that of the DSP56002.
REV PrM
–18–
PRELIMINARY TECHNICAL DATA
AD9834
DVDD AVDD
C1 C2
0.1µ F 0.1µ F
C13
1 0.01µ F LK4
2
SCLK AVDD
3
SDATA
4 6 5 4 C3
FSYNC DVDD C14
5 0.01µ F CAP DVDD AVDD 10nF
C6 3
J1 COMP
6 0.1µ F DVDD
1 2 C4
7 REFOUT
2 18 14 SCLK 0.1µ F
8 SCLK
9 4 16 13 SDATA 12 LK5
SDATA SLEEP
10
6 14 15 FSYNC 1 R4
11 FSYNC FSADJUST
12 8 12 11 6.8k
RESET RESET
13
14 C11 IOUTB
RESET 20
15
U2 U1 IOUTB
16 R1 R2 R6 C12
10K 10K AD9834 IOUT 200R
17 19
18 IOUT
PSELECT LK1 R5
19 10
PSELECT 200R C11
20
FSELECT 9 VIN 17 R7
21 FSELECT
LK2 300R
22
C11
23
8 16
24 MCLK SBOUT
SIGNBITOUT
25
26 SW DVDD DGND AGND
27 7 18 DVDD
J2 J3 AVDD
28
C8 C7 C9 C10
29
10µF 0.1µ F 0.1µ F 10µF
30 MCLK
LK3
31
DVDD
32 R3 50R 14
33 C5 DVDD
34 0.1µ F
35 U3
36 8
OUT
DGND
7
Switch
Capacitors SW End Stackable Switch (SDC
C1 C2 C5 C6 C7 C9 C14 100nF Ceramic Capacitor Double Throw)
C3 C4 C13 10nF ceramic Capacitor
C8 C10 10uF Tantalum Capacitor
C11 C12 C15 C16 Option for extra Sockets
decoupling capacitors PSEL1; FSEL1; CLK1; Sub Minature BNC
IOUT; IOUTB; SBOUT; Connector
Connectors
Resistors J1 36-Pin Edge Connector
R1 R2 10 KΩ Resistor J2, J3 PCB Mounting Terminal
R3 51 Ω Resistor Block
R4 6.8 kΩ Resistor
R5 R6 200 Ω Resistor
R7 300 Ω Resistor
REV PrM
–19–
PRELIMINARY TECHNICAL DATA
AD9834
OUTLINE DIMENSIONS
Dimensions shown in inches and (mm).
0.260 (6.60)
0.252 (6.40)
20 11
0.177 (4.50)
0.169 (4.30)
0.256 (6.50)
0.246 (6.25)
1
10
REV PrM
–20–