[go: up one dir, main page]
Scribd
Upload
53 views17 pages

Errors in Adc

The document discusses the significance of accuracy in Analog-to-Digital Converters (ADCs) and the various types of errors that can occur during the conversion process, including offset error, gain error, differential non-linearity, integral non-linearity, quantization error, noise, and jitter. Each error type is explained in terms of its causes, impact on ADC output, and methods for measurement. Understanding and mitigating these errors is crucial for maintaining the integrity of digital data in applications such as healthcare and audio processing.

Uploaded by

shanefrancisjavier21
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
53 views17 pages

Errors in Adc

The document discusses the significance of accuracy in Analog-to-Digital Converters (ADCs) and the various types of errors that can occur during the conversion process, including offset error, gain error, differential non-linearity, integral non-linearity, quantization error, noise, and jitter. Each error type is explained in terms of its causes, impact on ADC output, and methods for measurement. Understanding and mitigating these errors is crucial for maintaining the integrity of digital data in applications such as healthcare and audio processing.

Uploaded by

shanefrancisjavier21
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 17

Errors in

Analog to Digital
Conversion
Importance of Accuracy
The precision of Analog-to-Digital Converters
(ADCs) holds immense significance, as it
establishes the proximity of the digital output
from an ADC to the actual value of the input
analog signal. This holds immense weight in
domains like healthcare apparatus, exacting
measuring equipment, and audio manipulation.
Even a minor discrepancy in the conversion
process can carry substantial repercussions.
To illustrate, in medical imaging, an inaccuracy
might trigger an incorrect assessment of a
patient's condition. Similarly, when dealing with
audio manipulation, it has the potential to
induce distortion or compromise the overall
sound quality. Hence, a comprehensive grasp
of error factors within ADCs and their
mitigation becomes pivotal in upholding the
credibility of the transformed data.
Types of Errors

1. Offset Error
2. Gain Error
3. Linearity Errors
a) Integral Linearity Error
b) Differential Linearity Error
4. Quantization Error
5. Noise
6. Jitter
Offset Error Types of Errors

When the input is at zero scales or the minimum, offset error is the difference between the
ADC's output and the ideal output. The output values shift across the full input range as a
result of the addition of a constant bias to every conversion. The offset mistake creates a
change in the code transition points by moving the transfer function along the horizontal axis.

Figure 1: A graph showing the +1 LSB offset, the Figure 2: A graph showing the -1.5 LSB offset, the
actual response, and the ideal response actual response, and the ideal response
Offset Error Types of Errors

Causes

The primary causes of offset errors include:

Component Mismatches: Flaws and irregularities arising during the fabrication of


internal ADC elements, like resistors and capacitors, can induce discrepancies that
manifest as offset errors.

Temperature Drift: Shifts in ambient temperature have the capacity to modify the
attributes of electronic components, provoking alterations in their values, and thereby
yielding offset errors.

Supply Voltage Variations: Oscillations in the supply voltage have the potential to impact
the reference voltage and the scope of analog input, ultimately leading to offset errors.
Offset Error Types of Errors

Impact on ADC Output

The ADC output values move


uniformly when there is an offset error.
This may be seen as shifting the
ADC's entire transfer characteristic
upwards or downwards. The
implication is that the ADC output will
show a non-zero number even when
the input is zero or at its minimum. In
applications that call for precise DC
measurements or in differential
measurement systems where small
voltage variations are crucial, this can
be particularly challenging.
Offset Error Types of Errors

Methods for Measuring Offset Error

Direct Measurement: Initiate the ADC with the minimum input (commonly zero voltage)
and record the resultant digital output. The disparity between this reading and the ideal
output (typically zero) signifies the offset error.

Best Fit Line Method: Employ an array of recognized input voltages for the ADC and
construct a graph of the corresponding output values. By fitting a line through these points
using a linear regression algorithm, the divergence of this line from the anticipated zero
input-output characterizes the offset error.

Histogram Method: Implement a known constant voltage in proximity to the midpoint of


the ADC's input range and gather a substantial quantity of samples. Crafting a histogram
from this dataset unveils the mean value. Comparing this mean value with the input value
facilitates the revelation of the offset error.

It is crucial to remember that if the offset error is calculated, it is frequently possible to


adjust and compensate for it in software by adding a correction factor to the ADC's output
during the processing phase.
Gain Error Types of Errors

Gain error, also known as full-scale error, is the difference between the actual and ideal
output slopes of an ADC. It denotes a scaling mistake, where the output of the ADC is either
compressed or stretched relative to the ideal linear response, to put it simply. The gain
mistake is depicted in Figure 3 for clarity.

The response for an ADC with -1 LSB gain error


Figure 3: A graph showing the concept of gain error
Gain Error Types of Errors

Causes
The main causes of gain errors include:

Component Tolerances: Tolerance variations in capacitors and resistors within the ADC
may have an impact on the gain. For instance, differences in resistor values can change
the scaling factor in resistor ladder networks.

Reference Voltage Inaccuracies: The ADC's input range is mostly determined by its
reference voltage. An inaccuracy in gain can be introduced by any change in this
reference voltage.

Temperature Variations: Just like offset mistakes, variations in temperature can have an
impact on a component's properties and, as a result, the ADC's gain.
Gain Error Types of Errors

Impact on ADC Output

The ADC's output scaling is impacted


by gain error across the whole input
range. Gain error modifies the
relationship between input and output
proportionality as opposed to offset
error, which is a continuous shift. This
is particularly significant in
applications where relative
measurements or accurate amplitude
representation are crucial and can
result in inaccurate representations of
input amplitudes, especially at higher
input levels.
Gain Error Types of Errors

Methods for Measuring Gain Error

End-point Method: This approach entails introducing two recognized voltages to the
ADC, one at the lower limit of the input range (usually zero) and another at the upper
boundary. By gauging the output codes and juxtaposing them with the anticipated values,
the gain error can be deduced as the contrast between the factual and anticipated incline
of the transfer function.

Best Fit Line Method: Analogous to its application in measuring offset error, the best-fit
line method proves valuable in determining gain error as well. By introducing a sequence
of known input voltages and configuring a line that best aligns with the output data points,
the gradient of this line can be compared with the ideal slope to calculate the gain error.

Code Density Test: In this methodology, a noisy signal is directed to the ADC, and a
substantial volume of samples is recorded. The dissemination of output codes can unveil
non-linear tendencies, facilitating the calculation of gain error.
Differential Non-linearity (DNL) Error Types of Errors

The ideal response exhibits a


uniform staircase input-output
characteristic, meaning that each
transition occurs at 1 LSB (least
significant bit) from the previous
transition. In practice, the step
width might differ from the ideal
value (1 LSB). The purple curve
above shows the response of a
hypothetical ADC where the steps
are not uniform. In this example,
the width of code 010 is 1.25 LSB,
while the next code exhibits a
smaller width of 0.54 LSB. The
DNL specification characterizes
how the ADC steps deviate from
An example showing the ideal transfer function for a
the ideal value. 3-bit unipolar ADC.
Differential Non-linearity (DNL) Error Types of Errors

For an ADC, the DNL of the k-th


code is defined by the following
equation:

Where W(k) and Wideal denote


the width of the k-th code and the
ideal step size, respectively. As an
example, for code 1 (or 001) in the
above figure, we have:

An example showing the ideal transfer function for a


3-bit unipolar ADC.
Integral Non-linearity (DNL) Error Types of Errors

The deviation of a step width from


the ideal value is characterized by
the differential nonlinearity (DNL)
specification. However, the DNL
error cannot completely describe
the transfer function deviation
from the ideal response because
the response we obtain depends
on how the positive and negative
DNL errors are spread across
different codes. The INL
specification allows us to
characterize the deviation of code
transitions from their ideal values.

Example transfer curve for an ADC.


Integral Non-linearity (DNL) Error Types of Errors

To calculate the INL of code k, we


can use the following equation:

Where Ta(k) and Tideal(k)


respectively denote the actual and
the ideal transition from code k-1
to k; and the “ideal step size” is
the LSB of the ADC. For the
above example, the actual
transition from code 1 (001) to
code 2 (010) occurs at 0.125 LSB
above the ideal transition.
Therefore, the INL of code 2 is Example transfer curve for an ADC.
INL(2) = +0.125 LSB.
Jitter Types of Errors

Jitter is the deviation from the


ideal timing of an event to the
actual timing of an event. To
understand what this means,
imagine you are sending a digital
sine wave and plotting it on graph
paper. Each square corresponds
to a clock pulse; because the
vertical lines are equidistant apart,
you end up with a perfectly
periodic clock signal. At each
clock pulse, you receive three bits
and plot that point on your graph
paper. Because of the periodic
nature, it ends up as a nice sine
wave. A sample clock that is periodic allows a digital
system to communicate correctly and accurately.
Jitter Types of Errors

Now, imagine that those lines


aren’t equidistant apart. This
would make your clock signal less
periodic. When you plot your data,
it isn’t at the same intervals and,
thus, doesn’t look correct.
In Figure 5, you can see that the
distance between the transitions
in the clock signal is uneven; this
is jitter in the clock. Although the
above figure has an exaggerated
amount of jitter, it does show how
a jittery clock can cause samples
to be triggered at uneven
intervals. This unevenness
introduces distortion into the
waveform you are trying to record If a clock signal has jitter, it results in distortion of
the digital waveform.
and reproduce.

You might also like