Data Interface through Four 3.125Gbps JESD204A compliant
data lanes
Single carrier WCDMA ACLR = 76 dBc @ 80 MHz IF
Analog output: adjustable 8.7 mA to 31.7 mA,
RL = 25 Ω to 50 Ω
2x/4x/8x interpolator/complex modulator allows carrier
placement anywhere in DAC bandwidth
Multi-chip synchronization interface with latency locking
High performance, low noise PLL clock multiplier
Digital inverse sinc filter
56-lead, exposed paddle LFCSP package
The AD9128 is a dual, 16-bit, high dynamic range, digital-toanalog converter (DAC) that provides a sample rate of 1.25
GSPS, permitting a multi-carrier generation up to the Nyquist
frequency. The AD9128 includes features optimized for direct
conversion transmit applications, including complex digital
modulation, and gain and offset compensation. The DAC
outputs can interface seamlessly with Analog Devices’
AD9128
quadrature modulators such as the ADL537x Broadband
QMOD series.
The AD9128 incorporates four high-speed serial data lanes
reducing the interface connections between the DAC and its
digital companion chip compared with CMOS or LVDS parallel
interfaces. The serial interfaces are capable of receiving data
with voltage swings of 200 to 700mV
receiver equalization, the receiver is capable of capturing data
sent across 0 to 20 cm traces on an FR4 board. The AD9128 also
features multi-chip deterministic latency capability, allowing
multiple dual DACs to be in alignment with one another.
A serial port interface provides read/write access to on-chip
registers. Full-scale output current is programmable over a
range of 8.5 mA to 31 mA. TheAD9128 operates on 1.8 V and
3.3 V supply rails.
PRODUCT HIGHLIGHTS
1. Small package size 8mm x 8mm footprint
2. Fewer pins for data input word width with only Four
JESD204A data lines
3. Ultra low noise and intermodulation distortion (IMD)
enables high quality transmission of wideband signals from
baseband to high intermediate frequencies.
4. A proprietary DAC output switching technique enhances
dynamic performance.
. With 3 dB of typical
p-p
TYPICAL SIGNAL CHAIN
COMPLEX I AND
DC
DIGITAL INTERPOLATION FILTERS
FPGA/ASIC/DSP
Rev. PrI
Information furnished by Analog Devices is believed to be accurate and reliable. However, no
responsibility is assumed by Analog Devices for its use, nor for any infri ngements of patents or other
rights of third parties that may result from its use. Specifications subject to change without notice. No
license is granted by implication or otherwise under any patent or patent rights of Analog Devices.
Trademarks and registered trademarks are the property of their respective owners.
SERialiser
DESerialiser
Figure 1. Typical signal chain with simplified block diagram of the AD9128
Input VIN Logic High IOVDD = 1.8 V 1.2 V
Input VIN Logic High IOVDD = 2.5 V 1.6 V
Input VIN Logic High IOVDD = 3.3 V 2.0 V
Input VIN Logic Low IOVDD = 1.8 V 0.6 V
Input VIN Logic Low IOVDD = 2.5 V, 3.3 V 0.8 V
CMOS OUTPUT LOGIC LEVEL
Output V
Output V
Output V
Output V
JESD204 DATA INTERFACE
Number of JESD204A lanes
JESD204A Serial interface speed
DAC sample rate
Input Data rate
RX0P/RX0N, RX1P/RX1N, RX2P/RX2N,
RX3P/RX3N
Input Impedance
DAC CLOCK INPUT (DACCLKP, DACCLKN)
Differential Peak-to-Peak Voltage 100 500 2000 mV
Common-Mode Voltage Self biased input, ac coupled 1.25 V
Receiver Differential Input Impedance, RIN TBD TBD Ω
Maximum Clock Rate 1250 MHz
REFCLK INPUT (REFCLKP, REFCLKN)
Differential Peak-to-Peak Voltage 100 500 2000 mV
Common-Mode Voltage 1.25 V
REFCLK Frequency (PLL Mode) 1 GHz ≤ f
Differential Peak-to-Peak Voltage 100 500 2000 mV
Common-Mode Voltage 1.25 V
SYNCIN INPUT (SYNCINP, SYNCINN)
Input Voltage Range, VIA or VIB 825 1575 mV
Input Differential Threshold, V
Input Differential Hysteresis, V
−100 +100 mV
IDTH
– V
IDTHH
20 mV
IDTHL
Receiver Differential Input Impedance, RIN 80 120 Ω
SYNCOUT OUTPUTS (SYNCOUT1P,
SYNCOUT1N and SYNCOUT2P, SYNCOUT2N)
Output Voltage High, VOA or VOB 1375 mV
Output Voltage Low, VOA or VOB 1025 mV
Output Differential Voltage, |VOD| 150 200 250 mV
Output Offset Voltage, VOS 1150 1250 mV
Output Impedance, Single-Ended, RO 80 100 120 Ω
Rev. PrI | Page 5 of 67
AD9128 Preliminary Technical Data
SERIAL PERIPHERAL INTERFACE
Maximum Clock Rate (SCLK) 20 MHz
Minimum Pulse Width High (t
Minimum Pulse Width Low (t
Setup Time, SDI to SCLK (tDS) 1.9 ns
Hold Time, SDI to SCLK (tDH) 0.2 ns
Data Valid, SDO to SCLK (tDV) 2.3 ns
Setup Time, CS to SCLK (t
DIGITAL INPUT DATA TIMING
Table 3. Input data timing specifications
Parameter Min Typ Max Unit
LATENCY (DACCLK Cycles)
1× Interpolation (With or Without Modulation) TBD Cycles
2× Interpolation (With or Without Modulation) TBD Cycles
4× Interpolation (With or Without Modulation) TBD Cycles
8× Interpolation (With or Without Modulation) TBD Cycles
Inverse Sinc TBD Cycles
Fine Modulation TBD Cycles
Power-Up Time TBD ms
NOISE SPECTRAL DENSITY (NSD) EIGHT-TONE, 500 kHz TONE SPACING
f
= 200 MSPS, f
DAC
f
= 400 MSPS, f
DAC
f
= 800 MSPS, f
DAC
W-CDMA ADJACENT CHANNEL LEAKAGE RATIO (ACLR), SINGLE CARRIER
f
= 491.52 MSPS, f
DAC
f
= 491.52 MSPS, f
DAC
f
= 983.04 MSPS, f
DAC
W-CDMA SECOND ACLR, SINGLE CARRIER
f
= 491.52 MSPS, f
DAC
f
= 491.52 MSPS, f
DAC
f
= 983.04 MSPS, f
DAC
= 20 MHz TBD dBc
OUT
= 50 MHz TBD dBc
OUT
= 70 MHz TBD dBc
OUT
= 70 MHz TBD dBc
OUT
= 50 MHz TBD dBc
OUT
= 60 MHz TBD dBc
OUT
= 80 MHz TBD dBc
OUT
= 100 MHz TBD dBc
OUT
= 80 MHz TBD dBm/Hz
OUT
= 80 MHz TBD dBm/Hz
OUT
= 80 MHz TBD dBm/Hz
OUT
= 10 MHz TBD dBc
OUT
= 122.88 MHz TBD dBc
OUT
= 122.88 MHz TBD dBc
OUT
= 10 MHz TBD dBc
OUT
= 122.88 MHz TBD dBc
OUT
= 122.88 MHz TBD dBc
OUT
Rev. PrI | Page 6 of 67
Preliminary Technical Data AD9128
ABSOLUTE MAXIMUM RATINGS
Table 5. Absolute Maximum Ratings
With
Respe
Parameter
AVDD33 AVSS,
IOVDD AVSS,
DVDD18, CVDD18, SVDD,
PLLVDD, VTTVDD
AVSS EPAD,
EPAD AVSS,
CVSS AVSS,
DVSS AVSS,
BIAS_RES, REFIO,
IOUT1P/IOUT1N,
IOUT2P/IOUT2N
RXN[3:0]/RXP[15:0],
JESD_FRAMEP/JESD_FRAME
N
DACCLKP/DACCLKN,
REFCLKP/REFCLKN
ct To Rating
−0.3 V to +3.6 V
EPAD,
CVSS,
DVSS
−0.3 V to +3.6 V
EPAD,
CVSS,
DVSS
AVSS,
EPAD,
CVSS,
DVSS
CVSS,
DVSS
CVSS,
DVSS
EPAD,
DVSS
EPAD,
CVSS
AVSS −0.3 V to AVDD33 +
EPAD,
DVSS
CVSS −0.3 V to CVDD18 +
−0.3 V to +2.1 V
−0.3 V to +0.3 V
−0.3 V to +0.3 V
−0.3 V to +0.3 V
−0.3 V to +0.3 V
0.3 V
−0.3 V to DVDD18 +
0.3 V
0.3 V
,
RESET
SDIO, SDO
Junction Temperature 125°C
Storage Temperature
Range
IRQ, CS
, SCLK,
EPAD,
DVSS
−65°C to +150°C
Stresses above those listed under Absolute Maximum Ratings
may cause permanent damage to the device. This is a stress
rating only; functional operation of the device at these or any
other conditions above those indicated in the operational
section of this specification is not implied. Exposure to absolute
maximum rating conditions for extended periods may affect
device reliability.
THERMAL RESISTANCE
θJA is specified for the worst-case conditions, that is, a device
soldered in a circuit board for surface-mount packages.
Table 6. Thermal Resistance
Package Type θJA θJB θJC Unit
56 pin LFCSP TBD TBD TBD °C/W
ESD CAUTION
−0.3 V to IOVDD +
0.3 V
Rev. PrI | Page 7 of 67
AD9128 Preliminary Technical Data
PIN CONFIGURATION AND FUNCTION DESCRIPTIONS
Figure 2.
Table 7. Pin List and Description
Pin Name Description I/O Pin
1 CVDD18 1.8V Clock Supply – Analog Clock Supply I
2 CVDD18 1.8V Clock Supply – Analog Clock Supply I
3 CVDD18 1.8V Clock Supply – PLL Supply (DAC clock generator) I
4 REFCLKP/DACALIGNP If PLL enabled, Reference Clock positive, else DAC align positive1. REFCLKP
can be AC coupled. DACALIGNP cannot be AC coupled and needs 100
ohm (LVDS) resistor
REFLCKN can be AC coupled. DACALIGNN cannot be AC coupled and
needs 100 ohm (LVDS) resistor
6 JESD_FRAMEP JESD204A Compliant frame clock, positive. Can be AC coupled. I
7 JESD_FRAMEN JESD204A Compliant frame clock, negative. Can be AC coupled. I
8 DVDD18 1.8V Digital Supply – for frame clock and timing DLL. I
9 JESD_SYNCOUT1P JESD204A SYNC Signal, Positive. LVDS compliant. O
10 JESD_SYNCOUT1N JESD204A SYNC Signal, Negative. LVDS compliant. O
11 DVDD18 1.8V Digital Supply (Core) I
1
Single Edge DAC Alignment input if PLL is disabled. LVDS resistor required between this pin and REFCLK in align mode.
Rev. PrI | Page 8 of 67
I
I
Preliminary Technical Data AD9128
12 DVDD18 1.8V Digital Supply (Core) I
13 JESD_SYNCOUT2P JESD204A SYNC Signal Auxiliary, Positive. LVDS compliant. O
14
15 RXN0 Serial Channel input 0, Negative. CML compliant. 50 ohm-terminated to
16 RXP0 Serial channel input 0, Positive. CML compliant. 50 ohm-terminated to Vtt
17 SVDD18 1.8V Deserializer Supply for RX0 and RX1 I
18 RXN1 Serial channel input 1, Negative. CML compliant. 50 ohm-terminated to
19 RXP1 Serial channel input 1, Positive. CML compliant. 50 ohm-terminated to Vtt
20 VTTVDD18 (SVDD18) 1.8V Deserializer supply (Vtt and bias generation supply) I
21 VTT SERDES Lane Input Termination Voltage. Used for supplying external
22 PLLVDD18 1.8V PLL Supply I
23 DVSS (GND) Tied to ground I
24 RXP2 Serial channel input 2, Positive . CML compliant. 50 ohm-terminated to Vtt
25 RXN2 Serial channel input 2, Negative. CML compliant. 50 ohm-terminated to
26 SVDD18 1.8V Deserializer Supply for Rx2 and RX3 I
27 RXP3 Serial channel input 3, Positive. CML compliant. 50 ohm-terminated to Vtt
28 RXN3 Serial channel input 3, Negative . CML compliant. 50 ohm-terminated to
29 JESD_SYNCINN JESD204A SYNC signal input, Negative. LVDS compliant SYNC input with
30 JESD_SYNCINP JESD204A SYNC signal input, Positive. LVDS compliant SYNC input with
31 DVDD18 1.8V Digital Supply (Core) I
32 DVDD18 1.8V Digital Supply (Core) I
33 DVDD18 1.8V Digital Supply I
34 DVDD18 1.8V Digital Supply I
35 IRQB Interrupt Request. Open Drain, Active Low Output O
36 CSB Serial Port Chip Select. Active Low (CMOS levels w.r.t. IOVDD) I
37 SDO Serial Port Data Output (CMOS levels w.r.t. IOVDD) O
38 SDIO Serial Port Data Input/Output (CMOS levels w.r.t. IOVDD) I/O
39 SCLK Serial Port Clock Input (CMOS levels w.r.t. IOVDD) I
40 IOVDD 1.8V – 3.3V Serial Port Supply I
41 RESETB Reset. Active Low. (CMOS levels w.r.t. IOVDD) I
42 TXENABLE Transmit Enable Function pin, programmable parameters through SPI. I
43 AVDD33 3.3V Analog Supply – DAC supply I
44 IOUT2P Q DAC Positive Current Output O
45 IOUT2N Q DAC Negative Current Output O
46 AVDD33 3.3V Analog Supply O
47 REFIO Voltage Reference. Nominally 1.2V output. Should be decoupled to
48 BIAS_RES External reference resistance. Used to set LVDS swing, DAC full-scale
JESD_SYNCOUT2N JESD204A SYNC Signal Auxiliary, Negative. LVDS compliant. O
I
Vtt pin voltage. Can be AC coupled. Resistance calibrated.
I
pin voltage. Can be AC coupled. Resistance calibrated.
I
Vtt pin voltage. Can be AC coupled. Resistance calibrated.
I
pin voltage. Can be AC coupled. Resistance calibrated.
I
termination voltage. Load should be < 100pF if internal voltage is used.
I
pin voltage. Can be AC coupled. Resistance calibrated.
I
Vtt pin voltage. Can be AC coupled. Resistance calibrated.
I
pin voltage. Can be AC coupled. Resistance calibrated.
I
Vtt pin voltage. Can be AC coupled. Resistance calibrated.
I
current, and deserializer input termination. Place 10K ohm resistor to
analog ground.
Rev. PrI | Page 9 of 67
AD9128 Preliminary Technical Data
49 AVDD33 3.3V Analog Supply I
50 AVDD33 3.3V Analog Supply I
51 IOUT1N I DAC Negative Current Output O
52 IOUT1P I DAC Positive Current Output O
53 AVDD33 3.3V Analog Supply I
54 DACCLKN/DACALIGNN DAC Clock Negative input if PLL disabled. DAC Alignment Negative input
if PLL is enabled. DACCLKN can be AC coupled. DACALIGNN cannot be AC
coupled and needs 100 ohm (LVDS) resistor
55 DACCLKP/DACALIGNP DAC Clock Positive input if PLL disabled. DAC Alignment Positive input if
PLL is enabled. DACCLKP can be AC coupled. DACALIGNP cannot be AC
coupled and needs 100 ohm (LVDS) resistor
56 CVDD18 1.8V Supply – Clock Supply Voltage I
I
I
Rev. PrI | Page 10 of 67
Preliminary Technical Data AD9128
TYPICAL PERFORMANCE CHARACTERISTICS
Figure 3
Figure 4
Figure 6
Figure 7
Figure 5
Rev. PrI | Page 11 of 67
AD9128 Preliminary Technical Data
TERMINOLOGY
Integral Nonlinearity (INL)
INL is the maximum deviation of the actual analog output from
the ideal output, determined by a straight line drawn from zero
scale to full scale.
Differential Nonlinearity (DNL)
DNL is the measure of the variation in analog value, normalized
to full scale, associated with a 1 LSB change in digital input code.
Offset Error
Offset error is the deviation of the output current from the ideal
of 0 mA. For IOUT1P, 0 mA output is expected when all inputs
are set to 0. For IOUT1N, 0 mA output is expected when all
inputs are set to 1.
Gain Error
Gain error is the difference between the actual and ideal output
span. The actual span is determined by the difference between
the output when all inputs are set to 1 and the output when all
inputs are set to 0.
Output Compliance Range
The output compliance range is the range of allowable voltage
at the output of a current output DAC. Operation beyond the
maximum compliance limits can cause either output stage
saturation or breakdown, resulting in nonlinear performance.
Temperature Drift
Temperature drift is specified as the maximum change from
the ambient (25°C) value to the value at either T
MIN
or T
MAX
.
For offset and gain drift, the drift is reported in ppm of fullscale range (FSR) per degree Celsius. For reference drift, the
drift is reported in ppm per degree Celsius.
Signal-to-Noise Ratio (SNR)
SNR is the ratio of the rms value of the measured output signal
to the rms sum of all other spectral components below the Nyquist
frequency, excluding the first six harmonics and dc. The value
for SNR is expressed in decibels.
Interpolation Filter
If the digital inputs to the DAC are sampled at a multiple rate of
f
(interpolation rate), a digital filter can be constructed that
DATA
has a sharp transition band near f
appear around f
(output data rate) can be greatly suppressed.
DAC
/2. Images that typically
DATA
Adjacent Channel Leakage Ratio (ACLR)
ACLR is the ratio in decibels relative to the carrier (dBc) between
the measured power within a channel relative to its adjacent
channel.
Complex Image Rejection
In a traditional two-part upconversion, two images are created
around the second IF frequency. These images have the effect
of wasting transmitter power and system bandwidth. By placing
the real part of a second complex modulator in series with the
first complex modulator, either the upper or lower frequency
image near the second IF can be rejected.
Current Mode Logic (CML)
CML is a differential digital logic family. Signal transmission is
point-to-point, unidirectional and terminated at the destination
with 50 resistors to a voltage, VTT, on both differential lines.
CML is the physical layer for JESD204.
Power Supply Rejection (PSR)
PSR is the maximum change in the full-scale output as the
supplies are varied from minimum to maximum specified
voltages.
Settling Time
Settling time is the time required for the output to reach and
remain within a specified error band around its final value,
measured from the start of the output transition.
Spurious Free Dynamic Range (SFDR)
SFDR is the difference, in decibels, between the peak amplitude
of the output signal and the peak spurious signal within the dc
to Nyquist frequency of the DAC. Typically, energy in this band
is rejected by the interpolation filters. This specification, therefore, defines how well the interpolation filters work and the
effect of other parasitic coupling paths on the DAC output.
Rev. PrI | Page 12 of 67
Preliminary Technical Data AD9128
THEORY OF OPERATION
The AD9128 is a 16-bit Dual DAC with a SERDES interface that
is fully compliant with the JESD204A specifications. Figure 8
shows a top-level diagram of the AD9128. Four high-speed
serial lanes carry data with a maximum speed of 3.125Gbps,
resulting in a 312.5 MSPS (maximum) input data rate for each
of the two DACs. The AD9128 can be configured to operate in
1, 2 or 4 JESD204A lane modes, depending on the required
DAC input data rate. It can also operate in single DAC mode,
with either 1-lane or 2-lane mode.
The two DACs can operate as I and Q channels in a direct
conversion transmitter. Or as two independent DACs running
at the same DAC sampling rate. The digital data-path of the
AD9128 offers four interpolation modes (1X, 2X, 4X or 8X)
through three half-band filters with a maximum DAC sampling
rate of 1.25 GSPS.
F
is the DAC sampling frequency. The input signal data-rate for
DAC
both DACs is
For I/Q applications a Numerically Controlled Oscillator
(NCO) provides a means for modulating the signal with a
programmable carrier signal. The NCO generates the carrier
signal for a complex modulator in the digital data path. The
resolution of the NCO is 32 bits, allowing the signal to be
placed in the output spectrum with very fine resolution. The
AD9128 Startup Sequence). The following sections describe
elements of the AD9128 in detail.
F
÷
the interpolation factor.
DAC
AD9128 also features coarse modulation. Coarse modulation
up converts a digital signal centered at DC center frequency of
F
/4. This option consumes significantly less power compared
DAC
with the NCO modulation approach. Digital gain, offset and
phase compensation are included in the AD9128 to help with
unwanted sideband suppression in direct conversion
transmitters. An inverse Sinc filter is provided to compensate
for DAC output sinc-related roll-off.
The DAC Clock (DACCLK) can be sourced externally. Or
generated on chip using a PLL synthesizer with externally
supplied reference signal.
The AD9128 DAC core provides a fully differential current
output with a nominal full-scale current of 20mA. The full-scale
current is user adjustable between 8.7mA and 31.7mA. The
differential current outputs are complementary.
The AD9128 is capable of multi-chip synchronization and can
both synchronize devices and establish deterministic latency
(latency locking) among multiple AD9128 devices. The latency
for each of the DACs remains constant from link establishment
to link establishment.
A SPI interface provides read and write access to registers. The
various functional blocks and the data interface need to be
setup in a specific sequence for proper operation (See section
Rev. PrI | Page 13 of 67
AD9128 Preliminary Technical Data
Figure 8. AD9128 Functional Block Diagram
Rev. PrI | Page 14 of 67
Preliminary Technical Data AD9128
T
A
HIGH SPEED SERIAL DATA INTERFACE
The AD9128 has four JESD204A data ports that receive data for
both I and Q transmit paths. Figure 9 describes the
communication layers implemented in the AD9128 for each
high speed serial data interface to recover the clock, descramble
and deserialize the data before it is sent to the Digital Signal
Processing section of the AD9128. If a lower data speed is
SYNC
RX P/N
FRAME
RECEIVERLINK LAYERDESCRAMBLER
Figure 9. Functional block diagram of Serial receiver
As the AD9128 can operate with more than one active high
speed serial data lane, both achieving synchronization and
handling loss of synchronization of the lanes are very
important. To simplify the interface to the companion digital
chip, theAD9128 designates one master signal (SYNCb) as far
as multiple lane synchronization is concerned. If one lane loses
synchronization, a resynchronization request is sent to the
transmitter and the transmitter stops sending data to all lanes
until resynchronization has been achieved.
RECEIVER CIRCUIT
The AD9128 provides four 1.8V differential serial input
interfaces compliant with the JESD204A specifications. These
interfaces can accept signals at frequencies up to 3.125Gbps
using the input topology in Figure 10.
RXP
Z
2
RDIFF
Z
RDIFF
RXN
Figure 10. Receiver line termination
The receiver eye mask in Figure 11 specifies the signal
amplitude and jitter tolerance for the AD9128 High Speed Serial
Data Interface receiver.
Z
TT
2
V
T
acceptable the part can be configured to operate with either two
or one JESD204A lanes. The maximum data speed is directly
linked to the number of lanes used. In the 4 lane, 2 lane and 1
lane configurations, the maximum supported data speeds are
312.5 MSPS, 156.25 MSPS and 78.125 MSPS respectively.
I DSP + DAC
TRANSPORT
LAYER
Q DSP + DAC
500mV
87.5mV
DIFFERENTI
VOLTAG E
L
0
- 87.5mV
-500mV
0.22 / DATA RATE
0.44 / DATA RATE
Figure 11. Receiver Eye Mask
The receiver is equipped with a Clock/Data Recovery circuit
(CDR) based on a PLL. The PLL effectively multiplies the Frame
clock input by 5 X F (F=Number of bytes per frame) and the
CDR synchronizes the phase used to sample the data on each
serial lane independently. This independent phase adjustment
per serial interface ensures accurate data sampling and eases the
implementation of multiple serial interfaces on a PCB. A byte
rate PLL clock is then used in the link layer, descrambler and
transport layers to deserialize the serial input and provide data
to the DAC inputs.
LINK LAYER
The AD9128 can operate with more than one active high speed
serial data interface. Link layer communications such as code
group synchronization, frame alignment and frame
synchronization are handled by all four lanes. However, the
configuration data is always checked only on a single logical
high speed serial data interface: LN0. This logical serial
interface can be connected to any of the four JESD204A
physical receivers RXn. It is important to note that logical LN0
must be active in all modes of operation.
The AD9128 decodes 8B/10B control characters allowing
marking of start and end of frame and alignment between serial
lanes. The AD9128 serial interface can issue a synchronization
Rev. PrI | Page 15 of 67
AD9128 Preliminary Technical Data
request by setting the SYNCb pin low. The synchronization
protocol follows the JESD204A standard. When a stream of 8
consecutive /K/ symbols is received, the AD9128 deactivates the
synchronization request by setting SYNCb pin high and waits
for the transmitter to issue an Initial Lane Alignment Sequence
(ILAS).
DESCRAMBLER
The AD9128 provides an optional descrambler block using a
self-synchronous descrambler with polynomial: 1 + x
Data scrambling can be selected at the transmitter to reduce
spectral peaks that would be produced when the same data
octets repeat from frame to frame. Another advantage of
scrambling is that it makes the spectrum data independent so
that possible frequency-selective effects on the electrical
interface will not cause data-dependent errors.
14
+ x15.
TRANSPORT LAYER
The transport layer maps the incoming descrambled data to
DAC samples. It provides control of the JESD204A parameters
shown in Table 8.
all 4 lanes cannot be used in single-DAC mode). The maximum
input data rate in single-DAC mode is the same as the dual
DAC mode (312.5MHz).
Table 10. AD9128 interface speeds
Parameter Value
Serial interface
speed
Effective serial
interface speed
3.125Gbps
2.5Gbps
# of lanes used 4 2 1
DAC Data update
rate (MHz)
312.5 156.25 78.125
Table 8. JESD204A Transport Layer Parameters
Parameter Description
F Number of bytes per frame: 1, 2 or 4 depending on
L
K Number of frame per multi-frame: K = 32 if F=1,
K=16 otherwise.
L Number of lanes per converter device: 1, 2 or 4
M Number of converter per device = 1, 2
Since the AD9128 uses the Frame input as the reference clock
for the deserializer PLL, the Frame input needs to be greater
than 50 MHz. A number of transport layer configurations are
defined to fit the AD9128 definition, as shown in
Ta bl e 9.
Table 9. JESD204A Configuration Parameters
Parameter Description
CF
Number of control words per frame clock per link
= 0. No control word is embedded with samples
CS Number of control bits per conversion sample = 0
High density user data format. Used when samples
HD
need to be split across lanes.
Set to 1 when F=1, 0 otherwise.
N Converter resolution = 16
N’ Total number of bits per sample = 16
Since the AD9128 has four high speed serial data interfaces,
several combinations of lanes per converter can be used
depending on the data rate desired.
The AD9128 can also operate in real single-DAC mode. In this
case, it can be configured in either 2-lane or 1-lane mode (Note:
Figure 12. Serial data interface with 4 lanes active
Figure 13. Serial data interface with 2 lanes active
FRAME INPUT
Figure 14. Serial data interface with 1 lane active
IDACn[7:0] IDAC
IDACn[15:8]
[15:8] IDAC
n+1
[15:8]QDAC
n+1
IDACn[7:0] QDACn[15:8] QDACn[7:0]
n+1
n+1
[7:0]
[7:0]
JESD204A SERIAL LINK ESTABLISHMENT
A brief summary of the high speed serial link establishment
process is given below. Please see the JESD204A Specifications
document (reference) for complete details.
1. Code group synchronization
a. Each receiver must locate K (K28.5) characters in its
input data stream
Rev. PrI | Page 16 of 67
Preliminary Technical Data AD9128
b. Once 8 consecutive K characters have been detected
on all link lanes, the receiver block de-asserts the
SYNCb signal to the transmitter block.
c. The transmitter captures the change in SYNCb and
after a fixed number of frame clocks, starts the Initial
Lane Alignment Sequence (ILAS).
2. Initial Lane Alignment Sequence
a. The main purposes of this phase are to align all the
lanes of the link and verify the parameters of the link.
b. Before the link is established, each of the link
parameters is written to the receiver device to
designate how data will be sent to the receiver block.
c. ILAS consists of 4 or more multi-frames. The last
character or each multi-frame is a multi-frame
alignment character /A/
d. The first, third, and fourth multi-frames are populated
with pre-determined data values. The de-framer uses
the final /A/ of each lane to align the ends of the
multi-frames within the receiver.
e. The second multi-frame contains an R (K.28.0),
Q(K.28.4), and then data corresponding to the link
parameters.
f. Additional multi-frames can be added to ILAS if
needed by the receiver. The AD9128 uses 8 multiframes in its ILAS. (When alignment scheme or
deterministic latency are used.)
g. After the last /A/ character of the last ILAS multi-
frame data begins to be streamed.
3. Data Streaming
a. In this phase data is streamed from the transmitter
block to the receiver block.
b. Data can be optionally scrambled. Scrambling does
not start until the very first octet following the ILAS.
c. The receiver block processes and monitors the data it
receives for errors including:
i. Bad running disparity (8b/10b error)
ii. Not in Table (8b/10b error)
iii. Unexpected control-character
iv. Bad ILAS
v. Inter-lane skew error (through character
replacement)
d. If any of these errors exists, it is reported back to the
transmitter in one of a few ways
i. SYNCb assertion: Resynchronization (SYNCb
pulled low) is called for at each error. For the first
FIFO OPERATION
The AD9128 contains several stages of FIFO to deal with the
high speed serial data interface protocol and to synchronize the
data input with the DAC clock input (See Figure 15).
The FIFO in the SERDES deframer interface is used to
synchronize the samples sent on the high speed serial data
interface with the deframer clock. This FIFO absorbs timing
variations between the data source and the deframer. When the
FIFO reaches either full or empty state, it is recommended that
the user reset it through Register 35 bit 0 and, if necessary, reestablish the SERDES data link. Note that resetting the SERDES
link does not reset the FIFO to half-full automatically.
A second 2 channel x 16-bit wide, 8- word deep FIFO exists in
the DAC (datapath FIFO) to absorb timing variations between
the DAC clock and the Deframer clock. Figure 16 shows the
block diagram of the data path through the FIFO. The data is
latched into the device, formatted and then written into the
FIFO register determined by the FIFO write-pointer. The value
of the write-pointer is incremented every time a new word is
loaded into the FIFO. Meanwhile, data is read from the FIFO
register determined by the read-pointer and fed into the digital
datapath. The value of the read-pointer is updated every rising
edge of the internal DAC based data clock. The one exception
to this occurs when a resynchronization request is in progress:
the write side of the FIFO does not increment and the read side
is held in reset at a fixed value. Once the ILAS is completed in
the AD9128, then the Datapath FIFO is automatically reset to
“half full”. During a synchronization request, the DAC outputs
are forced to mid-scale and the datapath is flushed. This is done
to prevent corrupted data from passing from the DAC.
Valid data will be transmitted through the FIFOs as long as the
FIFOs do not overflow or become empty. Nominally, data will
be written to the FIFO at the same rate as data is read from the
FIFO. This keeps the data level in the FIFO constant. If data is
written to the FIFO faster than data is read, the data level in the
FIFO increases. If the data is written to the device slower than
data is read, the data level in the FIFO decreases.
three errors, SYNCb is asserted after an error
counter reaches a given error threshold.
ii. SYNCb reporting: SYNCb is pulsed low for a frame
clock period if an error occurs
iii. Reporting may also be done via interrupt (not
covered by the JESD204A specification). See (i) for
error thresholds.
Rev. PrI | Page 17 of 67
AD9128 Preliminary Technical Data
R
CHARACTER
ALIGNMENT
CHARACTER
ALIGNMENT
CHARACTER
ALIGNMENT
CHARACTER
ALIGNMENT
RECEIVER
FIFO
FIFO
XBAR
FIFO
FIFO
Figure 15. Block Diagram of AD9128 FIFOs
INPUT
LATCH
DATA
ASSEMBLER
Figure 16 – Block Diagram of Datapath FIFO
FRAME CLOCKING
The frame clock is the master reference for the high speed serial
interface of the AD9128. It drives a PLL in the JESD204A part of
the system and needs to be set to the input data rate of the
system. The user has three options for the frame clock in
AD9128:
Externally sourced through pins JESD_FRAMEP/N: the
input should be AC coupled and will be self-biased
internally.
Externally sourced through REFCLKP/N: this is possible
only if the internal DAC PLL is used and the supplied
reference clock supplied to the PLL (via REFCLKP/N) is at
the data rate of the system. (abd equal to the FRAME rate)
The input should be AC coupled and will be self-biased
internally.
Internally sourced by using a divided down version of the
DAC clock: this helps minimize the number of low
frequency clocks in the user system.
The frame clock source is controlled and monitored through
register 0x001D.
SERDES PLL
The SERDES PLL generates clocks at half the rate of the serial
data rate and supplies them to the Clock and Data Recovery
Rev. PrI | Page 18 of 67
10-BIT/
8-BIT
10-BIT/
8-BIT
10-BIT/
8-BIT
10-BIT/
8-BIT
WRITE POINTER
READ POINTE
REG 0
REG 1
REG 2
REG 3
REG 4
REG 5
REG 6
REG 7
64-BITS
FIFO
DESCRAMBLE
DESCRAMBLE
DESCRAMBLE
DESCRAMBLE
16
DATA
PAT HS
SAMPLE RECONSTRUCTION
16
DACS
FIFO
FIFO
DAC
DAC
(CDR) block. The SERDES PLL settings are controlled and
monitored in the register 0x01E. The PLL divide ratio (register
0x01E, bits [3:0]) is dependent on the F value (number of bytes
per frame) of the JESD204A link. The F value of the link (1,2 or
4) should be written to this register. The SERDES PLL can be
monitored for lock by reading register 0x01E, bit 6.
The SERDES PLL lock can also be accessed through the
interrupt controller by writing Register 0x006, bits 7 and/or 6
high. Bit7 enables the interrupt if the SERDES PLL has lost lock,
and Bit6 enables the interrupt if the SERDES PLL is locked.
These interrupts can be found by reading register 0x009 bits 7
and 6 (when the interrupt output of the AD9128 falls,).
Note that the SERDES PLL must lock before parameters can be
written to the deframer.
CONFIGURING THE JESD204A SERIAL INTERFACE
After the SERDES PLL has been successfully locked, the
Deserializer SPI is available and can be verified by reading
register 0x02 bit 0. The Deserializer SPI is a synchronous
read/write SPI (See section Serial Peripheral Interface for SPI
interface details). It is addressed through the long addressing
mode (default for the AD9128). The addresses for this part of
the circuit range from 0x100-0x17F.
Preliminary Technical Data AD9128
Input termination
The AD9128 will auto-calibrate to 50 ohms termination on
power-up as register 0x010 bit 5 has a default setting of high.
The auto-calibrated value found will be held constant until bit 5
is disabled. Alternatively, a manual calibration value can be
entered through register 0x011 bit 3:0 (highest resistor value is
0000 and lowest value is 1111). Manual calibration requires
register 0x10 bit5 to be low and 0x11 bit4 to be high. All settings
for input termination can be setup and controlled through
registers 0x010 and 0x011.
The input termination voltage of the DAC can be sourced either
externally or internally:
External: An external voltage can be driven through the
VTT pin. In order to support DC compliance, its value
should match the common mode voltage of the CML
driver at the transmitter. It may be bypassed at the pin
to local ground.
Internal: The termination voltage can be supplied
internally by enabling register 0x010 bit 4. The VTT
buffer drives both the internal VTT termination and
the VTT pin. The termination voltage value can be set
through register 0x010 its 3:0. In this case, the VTT pin
should not be bypassed to ground. As in option 1, to
meet DC compliance, the value of the voltage should be
chosen to be close to the value of the CML driver
output common-mode. This will ensure minimum
power consumption.
For AC coupled systems, in order to minimize power
consumption, VTT should be set close to 600mV.
Clock Data Recovery (CDR)
The CDR circuits for the four lanes of the high speed serial
interface can be enabled through register 0x012 bits 3:0. For
two-lane or one-lane operation, any of the two or any one lane
can be chosen. Unused lanes, if enabled, will consume
unnecessary power.
Logical Lane Mapping/Enabling
Each of the four physical high speed serial interface lanes, if
used, must be mapped to an appropriate logical lane. For
example, if four physical lanes are enabled for use with two
converters then each of the four logical lanes are mapped to a
distinct physical lane. Logical lanes 0, 1, 2 and 3 will contain
IMSB, ILSB, QMSB, QLSB respectively. The logical lanes are
enabled through register 0x17D and their mapping is controlled
through register 0x016, as shown in Tab l e 11.
Table 11. Logical lane mapping for JESD204A link
Rev. PrI | Page 19 of 67
AD9128
configuration
# of
Lanes
# of
DAC
4 2 Each physical lane is mapped to a
2 2 Logical 0 should be mapped to the
data I serial input and logical 2 to
the data Q serial input. Logical
lanes 3 and 4 are unused in this
2 1 Logical 0 should be mapped to
data MSBs and logical 2 to the data
1 2 Logical 0 should to be mapped to
the one serial link and others are
1 1 Logical 0 should both be mapped
to the input lane carrying data
Description Reg.
0x017
D value
0x0F
distinct logical lane
0x05
case
0x05
LSBs
0x01
ignored
0x01
Each of the input lanes can be individually controlled as far as
serial symbol mapping is concerned. Both the ordering of the
bits (MSB to LSB or LSB to MSB) on bits 7:4, and the individual
polarities on bits 3:0 are controlled through register 0x017.
A few mode bits are required in order to operate the non default
mode of 4 Lanes or 2 Lanes and F = 1. These are contained in
the Register 0x177. They enable sub-modes of the base
configuration of the deframer:
If F=2, Register 0x177 bits 5:2 must be set to 0111 (1 lane
per DAC) and 0x176 must be set to 2.
If F=4, Register 0x177 bits 5:2 must be set to 1011 (2 DACs
1 line) and 0x176 must be set to 4.
If F=1, Register 0x177 bits 5:2 must be set to 0000
Programming the JESD204A link parameters
This section provides details of the link parameters with respect
to the modes of operations supported by the AD9128. The link
parameters are programmed through registers 0x150 to 0x15D.
In order to achieve an accurate comparison, all the register
values must be programmed the same at the transmitter and
receiver end of the link.
1. 0x150: Provides the DID (Device ID or link ID). This is a
comparison only value to identify the link name.
2. 0x151 bits 3:0: Provides the BID (Bank ID). It is an
extension of the DID and meets the same requirements as
the DID.
3. 0x152 bits 3:0: Provides the LID0 (lane ID for lane 0 within
a link). The AD9128 will check the lane identification
values on lane 0 only.
4. 0x153 bit7: Enables the scrambling function on the link.
5. 0x153 Bits4:0: Provides the number of lanes of the link
associated with DID. This value L will be set based on the
number of lanes used and is programmed as one less than
AD9128 Preliminary Technical Data
the number of lanes. The possible values for different DAC
modes are:
15. 0x15B, 0x15C: Reserved fields. Should be set to 0 on both
receiver and transmitter ends.
16. 0x15D: Checksum value equal to the sum of all the
registers from 0x150 to 0x15C modulo 256.
Programming ILAS (Initial Lane Alignment Sequence)
length:
In the AD9128 the length of the ILAS is programmed in register
0x178. It is programmed as the actual number of multi-frames
times four (for a value of 1, the ILAS will be 4 multi-frames
long). The AD9128 uses 8 multi-frames during the ILAS to
accomplish multichip alignment (or 4 if multi-chip alignment is
not needed).
In order to enable multichip alignment or latency locking,
register 0x178 should be set to 0x02 and register 0x17B bit 0
should be set to 1. When latency locking/alignment is not
needed in the system, register 0x178 should be set to 1 and
0x17B bit 0 to 0.
Register
0x015A
value
INTERRUPTS AND SYNCB CONTROL
The deframer monitors the link for errors, and in the AD9128
these errors can be reported back to the transmitter through
different methods:
Through interrupts
Through the SYNCb signal as frame width assertion
pulses on the line
Through the SYNCb interface as forced SYNC requests.
Figure 27 shows a block diagram of the AD9128 interrupt
control. Errors are counted on a lane by lane basis and either an
error interrupt or a SYNCb event is triggered as the count
reaches an Error Threshold. This threshold is programmed in
Register 0x17C. Error counts for each lane can be monitored
through the use of Registers 0x16D – 0x16F. The errors that the
deframer will detect are: Bad Disparity Error, Not in Table
Error, Unexpected control character, Alignment issue, Bad ILS
Sequence, Configuration mismatch
The Interrupt request is masked by bits in registers 0x17A and
0x17B as follows:
0x17A, Bit7 – Bad Disparity (set high to trigger Interrupt
request)
0x17A, Bit 6 – Not in Table
0x017A, Bit 5 – Unexpected control character
0x017A, Bit 4 – Interlane Alignment good
Rev. PrI | Page 21 of 67
0x017A, Bit 3 – Good ILAS sequence
0x017A, Bit2 – Good Checksum
0x017A, Bit1 – Good Frame Sync
0x017A, Bit0 – Good Code Group Sync
The SYNCb frame width error reporting can be enabled by
setting bit 1 of Register 0x175 high (Bad disparity, not in Table,
and Unexpected control character will actuate this error
reporting mode).
The SYNC force is masked by bits in 0x17B as follows:
0x17B, Bit7 – Bad disparity error (set high for error to force
SYNCb high upon error threshold)
0x17B, Bit6 – Not in Table error
0x17B, Bit5 - Unexpected control character
ENABLING THE LINK
Once SYNCb setup/calibration is completed and clocks have
settled, the link is ready to be established. The link can be
established by setting bit 7 of Register 0x00A high. A startup
sequence is performed for the delay path from DAC clock to
SYNCb.
1. The SYNCb phase selector is reset to zero
2. The SYNCb FIFO is reset.
3. The previously programmed value of the SYNCb
launch phase is programmed back to the SYNCb
phase selector.
Once the startup sequence concludes inside the DAC, the
SYNCb signal is allowed to fall. Conditional upon the link
parameters being consistent at both ends and the DAC being
able to capture data, the link will be established.
Loading...
+ 46 hidden pages
You need points to download manuals.
1 point = 1 manual.
You can buy points or you can get point for every manual you upload.