
SBAS309 MAY 2004
www.ti.com
10
DEFINITION OF SPECIFICATIONS
Analog Bandwidth
The analog input frequency at which the spectral power of
the fundamental frequency (as determined by FFT
analysis) is reduced by 3dB.
Aperture Delay
The delay in time between the falling edge of the input
sampling clock and the actual time at which the sampling
occurs.
Aperture Uncertainty (Jitter)
The sample-to-sample variation in aperture delay.
Clock Pulse Width/Duty Cycle
A perfect differential sine wave clock results in a 50% clock
duty cycle on the internal conversion clock. Pulse width
high is the minimum amount of time that the ENCODE
pulse should be left in logic ‘1’ state to achieve rated
performance. Pulse width low is the minimum time that the
ENCODE pulse should be left in a low state (logic ‘0’). At
a given clock rate, these specifications define an
acceptable clock duty cycle.
Differential Nonlinearity (DNL)
An ideal ADC exhibits code transitions that are exactly 1
LSB apart. DNL is the deviation of any single LSB
transition at the digital output from an ideal 1 LSB step at
the analog input. If a device claims to have no missing
codes, it means that all possible codes (for a 14-bit
converter, 16384 codes) are present over the full operating
range.
Effective Number of Bits (ENOB)
The effective number of bits for a sine wave input at a given
input frequency can be calculated directly from its
measured SINAD using the following formula:
ENOB
SINAD
1.76
6.02
If SINAD is not known, SNR can be used exceptionally to
calculate ENOB (ENOB
SNR
).
Effective Resolution Bandwidth
The highest input frequency where the SNR (dB) is
dropped by 3dB for a full-scale input amplitude.
Gain Error
The amount of deviation between the ideal transfer
function and the measured transfer function (with the offset
error removed) when a full-scale analog input voltage is
applied to the ADC, resulting in all 1s in the digital code.
Gain error is usually given in LSB or as a percent of
full-scale range (%FSR).
Integral Nonlinearity (INL)
INL is the deviation of the transfer function from a
reference line measured in fractions of 1 LSB using a “best
straight line” or “best fit” determined by a least square
curve fit. INL is independent from effects of offset, gain or
quantization errors.
Maximum Conversion Rate
The encode rate at which parametric testing is performed.
This is the maximum sampling rate where certified
operation is given.
Minimum Conversion Rate
This is the minimum sampling rate where the ADC still
works.
Nyquist Sampling
When the sampled frequencies of the analog input signal
are below f
CLOCK
/2, it is called Nyquist sampling. The
Nyquist frequency is f
CLOCK
/2, which can vary depending
on the sample rate (f
CLOCK
).
Offset Error
Offset error is the deviation of output code from
mid-code when both inputs are tied to common-mode.
Propagation Delay
This is the delay between the input clock rising edge and
the time when all data bits are within valid logic levels.
Signal-to-Noise and Distortion (SINAD)
The RMS value of the sine wave f
IN
(input sine wave for an
ADC) to the RMS value of the noise of the converter from
DC to the Nyquist frequency, including harmonic content.
It is typically expressed in decibels (dB). SINAD includes
harmonics, but excludes DC.
SINAD
20Log
(10)
Input(V
S
)
Harmonics
Noise
Signal-to-Noise Ratio (without harmonics)
SNR is a measure of signal strength relative to background
noise. The ratio is usually measured in dB. If the incoming
signal strength in
μ
V is V
S
, and the noise level (also in
μ
V)
is V
N
, then the SNR in dB is given by the formula:
SNR
20Log
(10)
V
S
V
N
This is the ratio of the RMS signal amplitude, V
S
(set 1dB
below full-scale), to the RMS value of the sum of all other
spectral components, V
N
, excluding harmonics and DC.
P