LTC2314-14
10
231414fa
For more information www.linear.com/LTC2314-14
APPLICATIONS INFORMATION
Overview
The LTC2314-14 is a low noise, high speed, 14-bit succes-
sive approximation register (SAR) ADC. The LTC2314-14
operates over a wide supply range (2.7V to 5.25V) and
provides a low drift (20ppm/°C maximum), internal refer-
ence and reference buffer. The internal reference buffer is
automatically configured to a 2.048V span in low supply
range (2.7V to 3.6V) and to a 4.096V span in the high
supply range (4.75V to 5.25V). The LTC2314-14 samples
at a 4.5Msps rate and supports an 87.5MHz data clock.
The LTC2314-14 achieves excellent dynamic performance
(77dB SINAD, 85dB THD) while dissipating only 31mW
from a 5V supply at the 4.5Msps conversion rate.
The LTC2314-14 outputs the conversion data with one
cycle of conversion latency on the SDO pin. The SDO pin
output logic levels are supplied by the dedicated OVDD
supply pin which has a wide supply range (1.71V to 5.25V)
allowing the LTC2314-14 to communicate with 1.8V, 2.5V,
3V or 5V systems.
The LTC2314-14 provides both nap and sleep power-down
modes through serial interface control to reduce power
dissipation during inactive periods.
Serial Interface
The LT2314-14 communicates with microcontrollers, DSPs
and other external circuitry via a 3-wire interface. A falling
CS edge starts a conversion and frames the serial data
transfer. SCK provides the conversion clock for the current
sample and controls the data readout on the SDO pin of
the previous sample. CS transitioning low clocks out the
first leading zero and subsequent SCK falling edges clock
out the remaining data as shown in Figures 5, 6 and 7 for
three different timing schemes. Data is serially output MSB
first through LSB last, followed by trailing zeros if further
SCK falling edges are applied. Figure 5 illustrates that dur-
ing the case where SCK is held low during the acquisition
phase, only one leading zero is output. Figures 6 and 7
illustrate that for the SCK held high during acquisition or
continuous clocking mode two leading zeros are output.
Leading zeros allow the 14-bit data result to be framed
with both leading and trailing zeros for timing and data
verification. Since the rising edge of SCK will be coincident
with the falling edge of CS, delay t2 is the delay to the first
falling edge of SCK, which is simply 0.5 • tSCK. Delays t2
(CS falling edge to SCK leading edge) and t10 (16th falling
SCK edge to CS rising edge) must be observed for Figures
5, 6 and 7 and any timing implementation in order for the
conversion process and data readout to occur correctly.
The user can bring CS high after the 16th falling SCK edge
provided that timing delay t10 is observed. Prematurely
terminating the conversion by bringing CS high before the
16th falling SCK edge plus delay t10 will cause a loss of
conversion data for that sample. The sample-and-hold is
placed in sample mode when CS is brought high. As shown
in Figure 6, a sample rate of 4.5Msps can be achieved on
the LTC2314-14 by using an 87.5MHz SCK data clock
and a minimum acquisition time of 40ns which results in
the minimum throughput time (tTHROUGHPUT) of 222ns.
Note that the maximum throughput of 4.5Msps can only
be achieved with the timing implementation of SCK held
high during acquisition as shown in Figure 6.
The LTC2314-14 also supports a continuous data clock
as shown in Figure 7. With a continuous data clock the
acquisition time period and conversion time period must
be designed as an exact integer number of data clock
periods. Because the minimum acquisition time is not an
exact multiple of the minimum SCK period, the maximum
sample rate for the continuous SCK timing is less than
4.5Msps. For example, a 4.375Msps throughput is achieved
using exactly 20 data clock periods with the maximum
data clock frequency of 87.5MHz. For this particular case,
the acquisition time period and conversion clock period
are designed as 4 data clock periods (TACQ = 45.7ns) and
16data clock periods (TCONV = 182.9ns) respectively,
yielding a throughput time of 228.6ns.
The following table illustrates the maximum throughput
achievable for each of the three timing patterns. Note
that in order to achieve the maximum throughput rate of