The TSA1002 is a 10-bit, 50Msps sampling
frequency Analog to Digital converter using a
CMOS technology combining high performances
and very low power consumption.
The TSA1002 is based on a pipeline structure and
digital error correction to provide excellent static
linearity and guarantee 9.4 effective bits at
Fs=50Msps, and Fin=15MHz .
A voltage reference is int egrated in the circuit to
simplify the design and minimize external
components. It is nevertheless possible to use the
circuit with an external reference.
Especially designed for high speed, low power
applications, the TSA1002 only dissipates 50m W
at 50Msps. A tri-state capability, available on the
output buffers, enables to address several slave
ADCs by a unique master.
The output data can be coded into two d ifferent
formats. A Data Ready signal is raised as the data
is valid on the output and can be used for
synchronization purposes.
The TSA1002 is available in commercial (0 to
+70°C) and extended (-40 t o +85°C) temperat ure
range, in a small 48 pins TQFP package.
ORDER CODE
Part Number
TSA1002CF0°C to +70°CTQFP48TraySA1002C
TSA1002CFT0°C to +70°CTQFP48Tape & ReelSA1002C
TSA1002IF-40°C to +85°CTQFP48TraySA1002I
TSA1002IFT-40°C to +85°CTQFP48Tape & ReelSA1002I
EVAL1002/AAEvaluation board
Temperature
Range
PackageConditioning Marking
PIN CONNECTIONS (top view)
AGND
index
corner
IPOL
VREFP
VREFM
AGND
VIN
AGND
VINB
AGND
INCM
AGND
AVCC
AVCC
AVCC
AVCC
DFSB
474844 43 42 41 40 39 38
46 45
1
2
3
4
5
6
7
8
9
10
11
12
13 14 15 16 17 18 19 20 21 22
DVCC
DVCC
TSA1002
CLK
DGND
DGND
OEB
NC
NC
DGND
VCCB
GNDB
VCCB
NC
DR
37
NC
36
NC
35
NC
34
D0 (LSB)
33
32
D1
31
D2
30
D3
D4
29
D5
28
D6
27
26
D7
25
D8
23 24
GNDB
GNDB
VCCBNCOR
D9 (MSB)
PACKAGE
7 × 7 mm TQFP48
APPLICATIONS
■ Medical imaging and ultrasound
■ Portable instrumentation
■ Cable Modem Receivers
■ High resolution fax and scanners
■ High speed DSP interface
October 2000
1/19
TSA1002
ABSOLUTE MAXIMUM RATINGS
SymbolParameterValuesUnit
AVCC
DVCC
VCCB
Analog Supply voltage
Digital Supply voltage
Digital buffer Supply voltage
IDoutDigital output current-100 to 100mA
TstgStorage temperature+150°C
Electrical Static Discharge
ESD
- HBM
- CDM-JEDEC Standard
1) All voltages values, e xcept differential voltage, are with respect to network g round terminal. The m agnitude of i nput and output volt ages
must neve r exceed -0.3V or VCC+0V
1IPOLAnalog bias current input25D8Digital ou tputCMOS output (2.5V)
2VREFP Top voltage reference1V26D7Digita l outputCMOS output (2.5V)
3VREFM Bottom voltage reference0V27D6Digital outputCMOS output (2.5V)
4AGNDAnalog ground0V 28D5Digital outputCMOS output (2.5V)
5VINAnalog input1Vpp29D4Digital outputCMOS output (2.5V)
6AGNDAnalog ground0V30D3Digital outputCMOS output (2.5V)
7VINBInverted analog input1Vpp31D2Digital outputCMOS output (2.5V)
8AGNDAnalog ground0V32D1Digital outputCMOS output (2.5V)
9INCMInput common mode0.5V33D0(LSB) Least Significant Bit output CMOS output (2.5V)
10AGNDAnalog ground0V34NCNon connected
1 1AVCC Analog p ow er supply2.5V35NCNon c onnected
12AVCC Analog power supply2.5V36NCNon connected
13DVCC Digital power s upply2.5V37NCNon connected
14DVCC Digital power supply2.5V38DRData Ready outputCMOS out put (2.5V)
15DGNDDigital ground0V39VCCBDigital Buffer power supply 2.5V
16CLKClock input2.5V compatible CMOS input40GNDB Digital Buffer ground0V
17DGNDDigital ground0V41VCCBDigital Buffer power supply 2.5V
18NCNon connected42NCNon connected
19DGN D Digita l ground0V43NCNon connected
20GND B Digital buffer ground0V44OEBOutput Enable i nput2.5V compatible CMOS input
21GNDB Digital buffer gro und0V45DFSBData Format Se lect input2.5V compa tible CMOS inpu t
22VCCBDigital buffer power supply 2.5V46AVCCAnalog power supply2.5V
23OROut Of Range outputCMOS output (2.5V)47AVCCAnalog power supply2.5V
24D9(MSB) Most Significant Bit outputCMOS output (2.5V)48AGNDAnalog ground0V
1) Rp ol = 18KΩ. Equiv alent load: Rloa d= 470Ω and Cload= 6pF
Tmin= -40°C to Tmax= 85°C. Not fully tes t ed over the tem perature ra nge. Guarant ed by samplin g.
2)
65.5
68.5
63.4
60
60
60
58.5
58.3
57.4
48
48
48
63.5
67.4
62.5
57
55
57
58.5
58.2
57.0
48
48
48
9.6
9.5
9.3
7.9
7.9
7.9
79.2
77
69
59.5
59.4
59.0
77.8
76
68.1
59.4
59.3
58.5
9.76
9.71
9.60
dBc
dBc
dB
dB
dB
dB
dB
dB
bits
bits
7/19
TSA1002
DEFINITIONS OF SPECIFIED PARAMETERS
STATIC PARAMETERS
Static measurements are performed through
method of histograms on a 2MHz input signal,
sampled at 40Msps, which is high eno ugh to fully
characterize the test frequency response. The
input level is +1dBFS to saturate the signal.
Differential Non Linearity (DNL)
The average devia tion of any outp ut code width
from the ideal code width of 1LSB.
Integral Non linearity (INL)
An ideal converter presents a transfer function as
being the straight line from the starting code to the
ending code. The INL is the deviation for each
transition from this ideal curve.
DYNAMIC PARAMETERS
Dynamic measurements are performed by
spectral analysis, appli ed to an in put sinewave of
various frequencies and sampled at 40Msps.
Spurious Free Dynamic Range (SFDR)
The ratio between the am plitude of fundamental
tone (signal power) and the power of the worst
spurious signal (not always an harmonic) over the
full Nyquist band. It is expressed in dBc.
Total Harmonic Distortion (THD)
The ratio of the rms sum of the first five harmonic
distortion components to the rms value of the
fundamental line. It is expressed in dB.
Signal to Noise Ratio (SNR)
The ratio of the rms value of the fundamental
component to the rms sum of all other spectral
components in the N yquist band (f
/2) excluding
s
DC, fundamental and the first five harmonics.
SNR is reported in dB.
Signal to Noise and Distorsion Ratio (SINAD)
Similar ratio as for SNR but including the harmonic
distortion components in the noise figure (not DC
signal). It is expressed in dB.
From the SINAD, the Effective Number of Bits
(ENOB) can easily be deduced using the formula:
SINAD= 6.02 × ENOB + 1.76 dB.
When the applied signal is not Full Scale (FS), but
has an A
The maximum analog input frequency at which the
spectral response of a full power signal is reduced
by 3dB. Higher values can be achieved with
smaller in put levels.
Effective Resolution Bandwidth (ERB)
The band of input signal frequencies that the ADC
is intended to convert without l oosing linearity i.e.
the maximum analog input frequency at which the
SINAD is decreased by 3dB or the ENOB by 1/2
bit.
Pipeline dela y
Delay between time when the analog input is
initially sampled and time when the corresponding
digital data output is valid on the output bus. Also
called data latency. It is expressed as a number of
clock cycles.
8/19
EQUIVALENT CIRCUITS
TSA1002
Figure 1 : Analog Input Circuit
AVCC=2.5V
VIN
(or V IN B )
PAD
CAPACITANCE
7 pF
AGND=0V
Figure 2 : Input clock circuit
DVCC=2.5V
CLK
Req # 33 kΩ
(if Fs=50 MHz)
comm on m ode
Figure 3 : Input buffers
VCC buf=2.5V
Ω278.5Ω208.2Ω355.5
DFS
7 pF
PAD
CAPACITANCE
GND buff=0V
Figure 4 : Tri-state output buffers
VCC buf=2.5V
OE
PAD
CAPACITANCE
7 pF
DGND=0V
DATA
GN D bu ff=0V
GND buff =0V
VCC b u f =2 .5V
OUT
2 mA
OUTPUT
BUFFER
PAD CAPAC ITANCE
7pF
9/19
TSA1002
Static parameter: Integral Non Linearity
Fs=50MSPS; Fin=1MHz; Icc=20mA; N=131072pts
0.8
0.6
0.4
0.2
0
-0.2
INL (LSBs)
-0.4
-0.6
-0.8
020040 060080010 00
O utput Code
Static parameter: Differential Non Linearity
Fs=50MSPS; Fin=1MHz; Icc=20mA; N=131072pts
0.5
0.4
0.3
0.2
0.1
0
-0 .1
DNL (LSBs)
-0 .2
-0 .3
-0 .4
-0 .5
02004006008001000
Linearity vs. AVcc
Fs=50MSPS; Icca=20mA; Fin=1MHz
60
59.5
59
58.5
58
57.5
57
56.5
56
55.5
Dynamic parameters (dB)
55
2.252.352.452.552.65
SNR
SINAD
ENOB
AVCC (V)
Output Code
10
9.9
9.8
9.7
9.6
ENOB (bits)
9.5
9.4
9.3
Distortion vs. AVcc
Fs=50MSPS; Icca=20mA; Fin=1MHz
-69
-71
-73
-75
-77
-79
-81
-83
Dynamic Parameters (dB)
-85
2.252.352.452.552.65
SFDR
THD
AVCC (V)
10/19
TSA1002
Linearity vs. DVcc
Fs=50MSPS; Icca=20mA; Fin=1MHz
59.1
59.05
59
58.95
58.9
58.85
SNR
ENOB
SINAD
Dynamic parameters (dB)
58.8
2.252.352.452.552.65
DVCC (V)
Linearity vs. VccB
Fs=50MSPS; Icca=20mA; Fin=1MHz
59.5
59
58.5
58
57.5
Dynamic parameters (dB)
57
2.252.352.452.552.65
SNR
SINAD
ENOB
VCCB (V)
9.6
9.595
9.59
9.585
9.58
9.575
9.57
10
9.9
9.8
9.7
9.6
9.5
9.4
Distortion vs. DVcc
Fs=50MSPS; Icca=20mA; Fin=1MHz
-65
-67
-69
-71
-73
-75
ENOB (bits)
-77
-79
-81
-83
Dynamic parameters (dB)
-85
2.252.352.452.552.65
SFDR
THD
DVCC (V)
Distortion vs. VccB
Fs=50MSPS; Icca=20mA; Fin=1MHz
-72
-73
-74
-75
-76
ENOB (bits)
-77
-78
-79
Dynamic Parameters (dB)
-80
2.252.352.452.552.65
THD
SFDR
VCCB (V)
Linearity vs. Fs
Icca=20mA; Fin=5MHz
66
61
56
51
Dynamic parameters (dB)
46
253545556575
ENOB
SNR
SINAD
Fs (MHz)
10
9.5
9
8.5
8
7.5
Distortion vs. Fs
Icca=20mA; Fin=5MHz
-50
-55
-60
-65
-70
ENOB (bits)
-75
-80
-85
Dynamic parameters (dB)
-90
253545556575
THD
SFDR
Fs (MHz)
11/19
TSA1002
Linearity vs. Fs
Icca=20mA; Fin=15 MHz
66
61
56
51
Dynamic parameters (dB)
46
253545556575
ENOB
SNR
SINAD
Fs (MHz)
Linearity vs. Fin
Fs=50MSPS; Icca=20mA
64
62
60
58
56
Dynamic parameters (dB)
54
0 204060
Fin (MH z)
SNR
SINAD
ENOB
9.6
9.1
8.6
8.1
7.6
10
9.5
9
8.5
8
7.5
Distortion vs. Fs
Icca=20mA; Fin=15 MHz
-50
-55
-60
-65
-70
ENOB (bits)
-75
-80
-85
Dynamic parameters (dB)
-90
253545556575
THD
SFDR
Fs (MHz)
Distortion vs. Fin
Fs=50MSPS; Icca= 2 0mA
-50
-55
-60
-65
-70
ENOB (bits)
-75
-80
Dynamic parameters (dB)
-85
0 204060
THD
SFDR
Fin (MHz)
Linearity vs.Temperatur e
Fs=50MSPS; Icca=20mA; Fin=5MHz
12/19
64
62
60
58
56
54
52
Dynamic Parameters (dB)
50
-50050100
ENOB
SNR
SINAD
Temperature (°C)
10
9.8
9.6
9.4
9.2
9
8.8
8.6
8.4
8.2
8
Distortion vs. Temperature
Fs=50MSPS; Icca=20mA; Fin=5MHz;
80
75
70
65
60
Dynamic Parameters (dB)
55
-50050100
SFDR
THD
Temperature (°C)
TSA1002 APPLICATION NOTE
DETAILED INFORMATION
The TSA1002 is a High Speed analog to digital
converter based on a pipeline architecture and the
latest deep submicron CMOS process to achieve
the best performances in terms of linearity and
power consumption.
The pipeline structure consists of 9 internal
conversion stages in which the analog signal is
fed and sequencially converted into digital data.
Each 8 first stages consists of an Analog to Digital
converter, a Digital to Analog converter, a Sample
and Hold and a gain of 2 amplifier. A 1.5bit
conversion resolution is achieved in each stage.
The latest stage simply is a comparator. Each
resulting LSB-MSB c ouple is then time shifted to
recover from the conversion delay. Digital data
correction completes the processing by
recovering from the redundancy of the (LSB-MSB)
couple for each stage. The corrected data are
outputed through the digital buffers.
Signal input is sampled on the rising edge of the
clock while digital outputs are delivered on the
falling edge of the Data Ready signal.
The advantages of such a convert er reside in the
combination of pipeline architec ture and the most
advanced technologies. The highest dynamic
performances are achieved while consumption
remains at the lowest level.
Some functionalities hav e been added i n order to
simplify as much as possible the application
board. These operational m odes are described in
the following table.
The TSA1002 is pin to pin compatible with the
8bits/40Msps TSA0801, the 10bits/25Msps
TSA1001 and t he 12bits/50Msps TSA1201. T his
ensures a conformity within the product family and
above all, an easy upgrade of the application.
OPERATIONAL MODES DESCRIPTION
InputsOutput s
Analog input differential levelDFSBOEBORDRMost Significant Bit (MSB)
(VIN-VINB)>RANGEHLHCLKD9
-RANGE>(VIN-VINB)HLHCLKD9
RANGE>(VIN-VINB)>-RANGEHLLCLKD9
(VIN-VINB)>RANGELLHCLKComplemented D9
-RANGE>(VIN-VINB)LLHCLKComplemented D9
RANGE>(VIN-VINB)>-RANGELLLCLKComplemented D9
XXHHZHZHZ
Data Format Select (DFSB)
When set to low level (VIL), the digital input DFSB
provides a tw o’s complement d igital output MSB.
This can be of interest when performing some
further signal processing.
When set to high level (VIH), DFSB provides a
standard binary output coding.
Output Enable (OEB)
When set to low level (VIL), all digital outputs
remain active and are in low impedance state.
When set to high level (VIH), all digital outputs
buffers are in high impedance state. This results in
13/19
lower consumption while the converter goes on
sampling.
When OEB is set to low level again, , the data is
then valid on the output with a very short Ton
delay.
The timing diagram summarizes this operating
cycle.
Out of Range (OR)
This function is im plemented on the output stage
in order to set up an "Out of Range" flag whenever
the digital data is over the full scale range.
TSA1002
Typically, there is a detection of all the dat a bei ng
at ’0’ or all the data being at ’1’. This ends up with
an output signal OR which is in low level state
(VOL) when the data stay within the range, or in
high level state (VOH) when the data are out of the
range.
Data Ready (DR)
The Data Ready output is an image of the clock
being synchronized on the output data (D0 to D9).
This is a very helpful signal that simplifies the
synchronization of the measurement equipment or
the cont ro llin g DSP.
As digital output, DR goes in high impedance state
when OEB is as serted to High level as described
in the timing diagram.
DRIVING THE ANALOG INPUT
Differentia l inp u t s
The TSA1002 has been designed to obtain
optimum performances when being differentially
driven. An RF transformer is a good way to
achieve such performances.
Figure 5 describes the schematics. The input
signal is fed to the primary of the transformer,
while the secondary drives both ADC inputs. The
common mode voltage of the ADC (INCM) is
connected to the center-tap of the second ary of
the transformer in order to bias the input signal
around this common voltage, internally set to
0.56V. The INCM is decoupled to maintain a low
noise level on this node. Our evaluation b oard is
mounted with a 1:1 ADT1-1 transformer from
Minicircuits. You might also use a higher
impedance ratio (1:2 or 1:4) to reduce the driving
requirement on the analog signal source.
Each analog input can drive a 1Vpp amplitude
input signal, so the resultant differential amplitude
is 2Vpp.
Figure 5 : Differential input configuration
Single-ended input configuration
Some applications may require a single-ended
input which is easily achieved with the
configuration reported on Figure 6.
In this case, it is recommended to use an
AC-coupled analog input and connect the other
analog input to the common mode voltage of the
circuit (INCM) so as to properly bi as the ADC. The
INCM may remain at the same internal level
(0.56V) thus driving only a 1Vpp i nput amplitude,
or it must be increased to 0.9V to drive a 2Vpp
input amplitude. You wi ll get higher pe rform ances
using a 2Vpp signal.
Figure 6 : Single-ended input configuration
Signal so urce
50Ω
100nF
330pF
VIN
TSA1002
VINB
INCM
10nF
470nF
0.9V
Dynamic characteristics, while not being as
remarkable as for differential configuration, are
still of very goo d quality. Measurements done at
50Msps, 2MHz input frequency, -1dBFS input
level sum up these performances. An SFDR of
-64.5dBc, a SNR of 57.8dB and an ENOB Full
Scale of 9.3bits are achieved.
REFERENCE CONNECTION
Inte rnal ref erence
In the standard configuration, the ADC is bi ased
with the internal reference voltage. VREFM pin is
connected to Analog Ground while VREFP is
internally set to a voltage of 1.03V. It is
recommended to d ecouple the V R EF P i n order to
minimize low and high frequency noise. Refer to
Figure 7 for the schematics.
Analog source
14/19
50Ω
ADT1-1
1:1
330pF
100pF
10nF
VIN
TSA1002
VINB
INCM
470nF
Figure 7 : Internal reference setting
VIN
1.03V
VREFP
330pF
TSA1002
VINB
VREFM
10nF
470nF
TSA1002
Exte rnal ref eren ce
It is possible to use an external reference vo ltage
instead of the internal one for specific applications
requiring even better linearity or enhanced
temperature behaviour. In this case, the amplitude
of the external voltage mu st be at least equal to
the internal one (1.03V). Using the
STMicroelectronics Vref TS821 leads to optim um
performances when configured as shown on
Figure 8.
Figure 8 : External reference setting
1k
Ω
10nF
470nF
VCCA
VIN
TSA1002
VINB
VREFP
VREFM
330pF
TS821
external
reference
At 15Msps sampling frequency, 1MHz input
frequency and -1dBFS amplitude signal,
performances can be improved of up to 2dBc on
SFDR and 0.3dB on SINAD. At 50Msps sam pling
frequency, 1MHz input frequency and -1dBFS
amplitude signal, performanc es can be improved
of up to 1dBc on SFDR and 0.6dB on SINAD.
This can be very helpful for example for
multichannel application to kee p a good matching
among the sampling frequency range.
Clock input
The quality of your converter is very dependant on
your clock input accuracy, in terms of aperture
jitter; the use of low jitter crystal controlled
oscillator is recommended.
The duty cycle must be between 45% and 55%.
The clock power supplies must be separated from
the ADC output ones to avoid digital noise
modulation at the output.
It is recommended to always keep the circuit
clocked, even at the lowest specified sampling
frequency of 0.5Msps, bef ore applying the supply
voltages.
Power co nsumption
The internal architecture of the TSA1 002 enables
to optimize the power consumption according to
the sampling frequency of the application. For this
purpose, a resistor is placed between I POL and
the analog Ground pins.
The TSA1002 wi ll combine highest pe rformances
and lowest consumption at 50Msps when Rpol is
in the range of 12kΩ to 20kΩ.
At lower sampling frequency, this value of resistor
may be changed and the consumption will
decrease as wel l.
The figure 9 sums up the relevant data.
Figure 9 : Analog Current consumption vs. Fs
According value of Rpol polarization resistance
60
50
40
30
Icca (mA)
20
10
0
253545556575
RPOL
ICCA
Fs (MHz)
20
18
16
14
12
10
8
Rpol (kOhms)
6
4
2
0
Layout precautions
To use the ADC circuits in the best manner at high
frequencies, some precautions have to be taken
for power supplies:
- First of all, the implementation of 4 separate
proper supplies and ground planes (analog,
digital, internal and external buffer ones) on the
PCB is mandatory for high speed circuit
applications to provide low inductance and low
resistance common return.
The separation of the analog signal from the
digital part is essential to prevent noise from
coupling onto the input signal.
- Power supply bypass capacitors must be placed
as close as possible to the IC pins in order to
improve high frequency bypassing and reduce
harmonic distortion.
- Proper termination of all inputs and outputs must
be incorporated with output termination resistors;
then the amplifier load wi ll be only resistive and
the stability of the amplifier will be improved. All
leads must be wide and as short as possible
especially for the analog input in order to decrease
parasitic capacitance and inductance.
- To keep the capacitive loading as low as
possible at digital outputs, short lead lengths of
routing are essential to minimize currents when
the output changes. To minimize this output
15/19
TSA1002
capacitance, buffers or latches close to the output
pins will relax this constraint.
- Choose component sizes as small as possible
(SMD).
EVAL1002 evaluation board
The characterization of the board has been made
with a fully ADC devoted test bench as shown on
Figure 10. The analog signal must be filtered t o be
very pure.
The dataready signal is the acquisition clock of the
logic analyzer.
The ADC digital outputs are latched by the octal
buffers 74LCX573.
All characterization measurements have been
made with:
SFSR=+0.2dB for static parameters.SFSR=-0.5dB for dynamic parameters.
Figure 10 : Analog to Digital Converter characterization bench
Type
330pF C33603470nFC7805A VCCJ12FICHE2M M
330pF C20603470nFC16805CLJ/SM BJ4SM B/H
330pF C8603470nFC19805AGNDJ19FICHE2MM
330pF C2603470nFC3805DFSBJ9FICHE2MM
330pF C5603
330pF C11603
330pF C30603
330pF C17603
330pF C14603
47uFC36C A P
47uFC34C A P
47uFC35C A P
47uFC42C A P
470nF C22805
470nF C32805
470nF C37805
470nF C38805
470nF C13805
470nF C28805
470nF C10805C O N 2J16SIP2
Type
47K
47K
47K
47K
47K
47K
47K
47K
47K
47K
47K
50
50
ator
ΩR12603DGNDJ20FICHE2MM
Ω
R14603DVCCJ15FICHE2M M
ΩR11603GndB1J22FICHE2MM
Ω
Raj1 VR5GndB2J21FICHE2MM
ΩR 10603M es co m m ode J8F IC H E2M M
Ω
R19603OEBJ10FICHE2MM
ΩR13603
Ω
R15603T2-A T1-1WTT2ADT
Ω
R16603T2-A T1-1WTT1ADT
Ω
R17603VccB1J18FICHE2M M
Ω
R18603VDDBUFF3VJ17FICHE2M M
Ω
R3603VinJ1SM B/H
Ω
R1603VrefMJ5FICHE2M M
U3T S SO P 20VrefPJ2F ICH E 2M M
U2TS S OP 20T SA 1002U1T Q F P 48
Information furnished is bel ieved to be accurate and reliable. However, STMicroe lectronics assumes no responsibility for the
consequences of use of such information nor for any infringement of patents or other rights of third parties which may result from
its use. No li cense is granted by imp lica tion or otherwise under any patent or patent rig hts of STMicroelectronics. Specificat ions
mentioned in this publication ar e subject to change without notice. This publication supersedes and replaces all information
previously supplied. S TMicroelectronics products are not authorized for use as critica l components in life suppo rt devices or
systems without express written approval of STMicroelectronics.