The MAX113/MAX117 are microprocessor-compatible,
8-bit, 4-channel and 8-channel analog-to-digital converters (ADCs). They operate from a single +3V supply
and use a half-flash technique to achieve a 1.8µs conversion time (400ksps). A power-down pin (PWRDN)
reduces current consumption to 1µA typical. The
devices return from power-down mode to normal operating mode in less than 900ns, allowing large supplycurrent reductions in burst-mode applications. (In burst
mode, the ADC wakes up from a low-power state at
specified intervals to sample the analog input signals.)
Both converters include a track/hold, enabling the ADC
to digitize fast analog signals.
____________________________Features
♦ +3.0V to +3.6V Single-Supply Operation
♦ 4 (MAX113) or 8 (MAX117) Analog Input Channels
♦ Low Power: 1.5mA (operating mode)
1µA (power-down mode)
♦ Total Unadjusted Error ≤ 1LSB
♦ Fast Conversion Time: 1.8µs per Channel
♦ No External Clock Required
♦ Internal Track/Hold
♦ Ratiometric Reference Inputs
♦ Internally Connected 8th Channel Monitors
Reference Voltage (MAX117)
Microprocessor (µP) interfaces are simplified because
the ADC can appear as a memory location or I/O port
without external interface logic. The data outputs use
latched, three-state buffer circuitry for direct connection
to an 8-bit parallel µP data bus or system input port.
The MAX113/MAX117 input/reference configuration
enables ratiometric operation.
The 4-channel MAX113 is available in a 24-pin DIP or
SSOP. The 8-channel MAX117 is available in a 28-pin
DIP or SSOP. For +5V applications, refer to the
MAX114/MAX118 data sheet.
________________________Applications
Battery-Powered SystemsPortable Equipment
System-Health MonitoringRemote Data Acquisition
Communications Systems
______________Ordering Information
PART
MAX113CNG
MAX113CAG
MAX113C/D0°C to +70°C
MAX113ENG
MAX113EAG
MAX113MRG-55°C to +125°C
Ordering Information continued at end of data sheet.
*Dice are specified at T
**Contact factory for availability.
Stresses beyond those listed under “Absolute Maximum Ratings” may cause permanent damage to the device. These are stress ratings only, and functional
operation of the device at these or any other conditions beyond those indicated in the operational sections of the specifications is not implied. Exposure to
absolute maximum rating conditions for extended periods may affect device reliability.
Note 1: Accuracy measurements performed at V
Note 2: Guaranteed by design.
Note 3: Power-down current increases if logic inputs are not driven to GND or V
Note 4: Input control signals are specified with t
delays get shorter at higher supply voltages. See the Conversion Time vs. Supply Voltage graph in the
Characteristics
to extrapolate timing delays at other power-supply voltages.
= tf= 5ns, 10% to 90% of 3V, and timed from a voltage level of 1.3V. Timing
r
Typical Operating
Note 5: See Figure 1 for load circuit. Parameter defined as the time required for the output to cross 0.66V or 2.0V.
Note 6: See Figure 2 for load circuit. Parameter defined as the time required for the data lines to change 0.5V.
Note 7: Also defined as the Minimum Address-Valid to Convert-Start Time.
Analog Input Channel 6IN6—
Analog Input Channel 5IN5—2
Analog Input Channel 4IN41
Analog Input Channel 3
Analog Input Channel 2IN23
Analog Input Channel 1
Mode Selection Input. Internally pulled low with a 15µA current source. MODE = 0
Three-State Data Output (LSB)D06
Three-State Data Outputs D1, D2, D37, 8, 9
Read Input. RD must be low to access data (see
Interrupt Output. INT goes low to indicate end of conversion (see
section).
GroundGND12
Lower limit of reference span. REF- sets the zero-code voltage. Range is GND ≤
V
< V
REF-
Upper limit of reference span. REF+ sets the full-scale input voltage. Range is
V
< V
REF-
Write-Control Input/Ready-Status Output (see
Chip-Select Input. CS must be low for the device to recognize WR or RD inputs.CS
Three-State Data Outputs D4, D5, D617, 18, 19
Three-State Data Output (MSB)D720
Multiplexer Channel Address Input (MSB)A2—
Multiplexer Channel Address Input A121
Multiplexer Channel Address Input (LSB)A022
Power-Down Input. PWRDN reduces supply current when low.PWRDN
Positive Supply, +3.0V to +3.6VV
Analog Input Channel 7IN7—
The MAX113/MAX117 use a half-flash conversion technique (see
ADC sections achieve an 8-bit result. Using 15 comparators, the flash ADC compares the unknown input
voltage to the reference ladder and provides the upper
four data bits. An internal digital-to-analog converter
(DAC) uses the four most significant bits (MSBs) to
generate both the analog result from the first flash conversion and a residue voltage that is the difference
between the unknown input and the DAC voltage. The
residue is then compared again with the flash comparators to obtain the lower four data bits (LSBs).
An internal analog multiplexer enables the devices to
read four (MAX113) or eight (MAX117) different analog
voltages under microprocessor (µP) control. One of the
MAX117’s analog channels, IN8, is internally hardwired and always reads V
In burst-mode or low-sample-rate applications, the
MAX113/MAX117 can be shut down between conversions, reducing supply current to microamp levels (see
Typical Operating Characteristics
PWRDN pin shuts the devices down, reducing supply
current typically to 1µA when powered from a single
+3V supply. A logic high on PWRDN wakes up the
MAX113/MAX117, and the selected analog input enters
the track mode. The signal is fully acquired after 900ns
(this includes both the power-up delay and the
track/hold acquisition time), and a new conversion can
Functional Diagram
REF+
) in which two 4-bit flash
when selected.
Power-Down Mode
). A logic low on the
V
DD
TO HIGH-Z
OL
3k
10pF
DATA
OUTPUTS
3k
a) V
TO HIGH-Z b) V
OH
Figure 2. Load Circuits for Data-Hold Time Test
DATA
OUTPUTS
10pF
be started. If the power-down feature is not required,
connect PWRDN to VDD. For minimum current consumption, keep digital inputs at the supply rails in
power-down mode. Refer to the
Reference
section for
information on reducing the reference current during
power-down.
___________________Digital Interface
The MAX113/MAX117 have two basic interface modes,
which are set by the MODE pin. When MODE is low,
the converters are in read mode; when MODE is high,
the converters are set up for write-read mode. The A0,
A1, and A2 inputs control channel selection, as shown
in Table 1. The address must be valid for a minimum
time, t
+3V, 400ksps, 4/8-Channel,
8-Bit ADCs with 1µA Power-Down
Read Mode (MODE = 0)
In read mode, conversions and data access are controlled by the RD input (Figure 3). The comparator
inputs track the analog input voltage for the duration of
t
. A conversion is initiated by driving CS and RD
ACQ
low. With µPs that can be forced into a wait state, hold
RD low until output data appears. The µP starts the
conversion, waits, and then reads data with a single
read instruction.
In read mode, WR/RDY is configured as a status output
(RDY), so it can drive the ready or wait input of a µP.
RDY is an open-collector output (no internal pull-up)
that goes low after the falling edge of CS and goes high
at the end of the conversion. If not used, the WR/RDY
pin can be left unconnected. The INT output goes low
at the end of the conversion and returns high on the ris-
MAX113/MAX117
ing edge of CS or RD.
Write-Read Mode (MODE = 1)
Figures 4 and 5 show the operating sequence for writeread mode. The comparator inputs track the analog
input voltage for the duration of t
initiated by a falling edge of WR. When WR returns
high, the result of the four-MSBs flash is latched into the
output buffers and the conversion of the four-LSBs flash
starts. INT goes low, indicating conversion end, and the
lower four data bits are latched into the output buffers.
The data is then accessible after RD goes low (see
Timing Characteristics
A minimum acquisition time (t
).
ACQ
going low to the start of another conversion (WR going
low).
Options for reading data from the converter include
using internal delay, reading before delay, and pipelined
operation (discussed in the following sections).
. The conversion is
ACQ
) is required from INT
t
ADDRESS VALID
t
ACQ
t
RDY
UP
t
CSS
(N)
t
AH
WITH EXTERNAL
PULL-UP
t
CRD
t
ACCO
PWRDN
CS
RD
A0–A2
RDY
INT
D0–D7
Figure 3. Read Mode Timing (Mode = 0)
CS
t
CSH
t
WR
t
AH
t
CSS
t
RD
t
INTL
t
ACC2
WR
A0–A2
INT
D0–D7
t
CSS
t
ACQ
ADDRESS
VALID (N)
RD
Figure 4. Write-Read Mode Timing (tRD> t
t
CSH
t
ACQ
ADDRESS VALID (N + 1)
t
INTH
t
(N)
t
ACQ
t
READ2
VALID DATA
(N)
) (Mode = 1)
DH
t
DH
VALID DATA
ADDRESS VALID (N + 1)
INTL
t
t
CSH
AH
t
INTH
Using Internal Delay
The µP waits for the INT output to go low before reading
the data (Figure 4). INT goes low after the rising edge of
WR, indicating that the conversion is complete and the
result is available in the output latch. With CS low, data
outputs D0–D7 can be accessed by pulling RD low. INT
is then reset by the rising edge of CS or RD.
Fastest Conversion:
Reading Before Delay
An external method of controlling the conversion time is
shown in Figure 5. The internally generated delay
(t
) varies slightly with temperature and supply volt-
INTL
age, and can be overridden with RD to achieve the
fastest conversion time. RD is brought low after the rising edge of WR, but before INT goes low. This completes the conversion and enables the output buffers
that contain the conversion result (D0–D7). INT also
goes low after the falling edge of RD and is reset on the
rising edge of RD or CS. The total conversion time is
therefore: tWR+ tRD+ t
Besides the two standard write-read-mode options,
“pipelined” operation can be achieved by connecting
WR and RD together (Figure 6). With CS low, driving
WR and RD low initiates a conversion and concurrently
reads the result of the previous conversion.
_____________Analog Considerations
Figures 7a, 7b, and 7c show typical reference connections. The voltages at REF+ and REF- set the ADC’s
analog input range (Figure 10). The voltage at REFdefines the input that produces an output code of all
zeros, and the voltage at REF+ defines the input that
produces an output code of all ones.
The internal resistance from REF+ to REF- can be as
low as 1kΩ, and current will flow through it even when
the MAX113/MAX117 are shut down. Figure 7d shows
how an N-channel MOSFET can be connected to REFto break this current path during power-down. The FET
should have an on-resistance of less than 2Ω with a 3V
gate drive. When REF- is switched, as in Figure 7d, a
new conversion can be initiated after waiting a time
equal to the power-up delay (tUP) plus the N-channel
FET’s turn-on time.
Although REF+ is frequently connected to VDD, the circuit of Figure 7d uses a low-current, low-dropout, 2.5V
voltage reference: the MAX872. Since the MAX872
cannot continuously furnish enough current for the ref-
t
CSH
t
WR
t
ACQ
ADDRESS
VALID (N + 1)
t
IHWR
t
INTL
t
ID
NEW DATA (N)
= 1800ns.
ACC1
Pipelined Operation
Reference
+3V
4.7µF
0.1µF
V
IN+
V
IN-
IN_
GND
V
REF+
REF-
DD
MAX113
MAX117
Figure 7a. Power Supply as Reference
V
IN+
V
IN-
IN_
GND
0.1µF
V
DD
REF+
REF-
MAX113
MAX117
+3V
4.7µF
0.1µF
7
6
8
1
3
LM10
4
2
+2.5V
34.8k
3.01k
Figure 7b. External Reference, 2.5V Full Scale
0.1µF
IN_
GND
V
DD
REF+
REF-
MAX113
MAX117
+3V
4.7µF
* CURRENT PATH MUST STILL
EXIST FROM V
0.1µF
TO GND
IN-
V
IN+
+2.5V
V
IN-
0.1µF
R*
Figure 7c. Input Not Referenced to GND
erence resistance, this circuit is intended for applications where the MAX113/MAX117 are normally in standby and are turned on in order to make measurements
at intervals greater than 100µs. C1 (the capacitor connected to REF+) is slowly charged by the MAX872 during the standby period, and furnishes the reference
current during the short measurement period.
The 4.7µF value of C1 ensures a voltage drop of less
than 1/2LSB when performing four to eight successive
conversions. Larger capacitors reduce the error still further. Use ceramic or tantalum capacitors for C1.
+3V, 400ksps, 4/8-Channel,
8-Bit ADCs with 1µA Power-Down
+3V
0.1µF
V
DD
MAX113
REF+
MAX872
PWRDN
* IRML2402
Figure 7d. An N-channel MOSFET switches off the reference
MAX113/MAX117
load during power-down
N-FET*
C1
4.7µF
0.1µF
MAX117
REF-
PWRDN
Initial Power-Up
When power is first applied, perform a conversion to
initialize the MAX113/MAX117. Disregard the output
data.
Bypassing
Use a 4.7µF electrolytic in parallel with a 0.1µF ceramic
capacitor to bypass VDDto GND. Minimize capacitor
lead lengths.
Bypass the reference inputs with 0.1µF capacitors, as
shown in Figures 7a, 7b, and 7c.
MAX113
MUX
V
IN2
R
.
IN
.
.
MAX117
R
ON
Figure 8. Equivalent Input Circuit
R
V
IN_
1
2k
V
IN
22pF
T/H
10pF
MAX113
MAX117
Analog Inputs
Figure 9. RC Network Equivalent Input Model
Figure 8 shows the equivalent circuit of the MAX113/
MAX117 input. When a conversion starts and WR is
low, V
is connected to sixteen 0.6pF capacitors.
IN_
During this acquisition phase, the input capacitors
charge to the input voltage through the resistance of
the internal analog switches. In addition, about 22pF of
stray capacitance must be charged. The input can be
OUTPUT CODE
11111111
11111110
11111101
FULL-SCALE
TRANSITION
modeled as an equivalent RC network (Figure 9). As
source impedance increases, the capacitors take
longer to charge.
The typical 32pF input capacitance allows source resistance as high as 1.5kΩ without setup problems. For
larger resistances, the acquisition time (t
increased.
Internal protection diodes, which clamp the analog
input to VDDand GND, allow the channel input pins to
swing from GND - 0.3V to VDD+ 0.3V without damage.
However, for accurate conversions near full scale and
zero scale the inputs must not exceed VDDby more
than 50mV or be lower than GND by 50mV.
If the analog input exceeds 50mV beyond the supplies, limit the input current to no more than two
milliamperes, as excessive current will degrade the
conversion accuracy of the on channel.
Track/Hold
The track/hold enters hold mode when a conversion
starts (RD low or WR low). INT goes low at the end of
the conversion, at which point the track/hold enters
track mode. The next conversion can start after the
minimum acquisition time, t
ACQ
.
Transfer Function
Figure 10 shows the MAX113/MAX117’s nominal transfer function. Code transitions occur halfway between
successive-integer LSB values. Output coding is binary
with 1LSB = (V
REF+
- V
REF-
) / 256.
Conversion Rate
The maximum sampling rate (f
MAX117 is achieved in write-read mode (tRD< t
and is calculated as follows:
f=
MAX
t + t + t + t
WRRDRIACQ
f
=
MAX
f 465kHz
MAX
600ns 800ns 300ns 450ns
=
1
+++
where tWR= the write pulse width, tRD= the delay
between write and read pulses, tRI= RD to INT delay,
and t
= minimum acquisition time.
ACQ
) for the MAX113/
MAX
1
INTL
Signal-to-Noise Ratio and
Effective Number of Bits
Signal-to-noise plus distortion (SINAD) is the ratio of the
fundamental input frequency’s RMS amplitude to all
other ADC output signals. The output spectrum is limited to frequencies above DC and below one-half the
ADC sample rate.
The theoretical minimum analog-to-digital noise is
caused by quantization error, and results directly from
the ADC’s resolution: SNR = (6.02N + 1.76)dB, where
N is the number of bits of resolution. Therefore, a perfect 8-bit ADC can do no better than 50dB.
The FFT Plot (see
Typical Operating Characteristics
shows the result of sampling a pure 30.27kHz sinusoid
at a 400kHz rate. This FFT plot of the output shows the
output level in various spectral bands.
The effective resolution (or “effective number of bits”)
the ADC provides can be measured by transposing the
equation that converts resolution to SNR: N = (SINAD -
1.76) / 6.02 (see
Typical Operating Characteristics
Total Harmonic Distortion
Total harmonic distortion (THD) is the ratio of the RMS
sum of all harmonics of the input signal (in the frequency band above DC and below one-half the sample rate)
to the fundamental itself. This is expressed as:
)
2
242
+++
VVV...V
2
THD = 20log
3
V
1
where V1is the fundamental RMS amplitude, and V
through VNare the amplitudes of the 2nd through Nth
harmonics.
Spurious-Free Dynamic Range
Spurious-free dynamic range (SFDR) is the ratio of the
fundamental RMS amplitude to the amplitude of the
next largest spectral component (in the frequency band
above DC and below one-half the sample rate). Usually
the next largest spectral component occurs at some
harmonic of the input frequency. However, if the ADC is
exceptionally linear, it may occur only at a random
peak in the ADC’s noise floor. See the Signal-to-Noise
Ratio graph in
Maxim cannot assume responsibility for use of any circuitry other than circuitry entirely embodied in a Maxim product. No circuit patent licenses are
Maxim cannot assume responsibility for use of any circuitry other than circuitry entirely embodied in a Maxim product. No circuit patent licenses are
implied. Maxim reserves the right to change the circuitry and specifications without notice at any time.
implied. Maxim reserves the right to change the circuitry and specifications without notice at any time.
12
__________________Maxim Integrated Products, 120 San Gabriel Drive, Sunnyvale, CA 94086 (408) 737-7600
12
__________________Maxim Integrated Products, 120 San Gabriel Drive, Sunnyvale, CA 94086 (408) 737-7600