Analog to Digital Converters (ADC)
Analog to Digital Converters (ADC)
Analog to Digital Converters
Created April 2008
◊ There are several methods of converting Analog
values to digital values, including:
◊ Flash ADC
◊ Ramp ADC
◊ SAC ADC
Other Methods of ADC
◊ Up/Down Ramp
◊ Similar to the Ramp ADC but instead of starting the
count at zero for each measurement it increments the
counter up or down based on the direction of the analog
input. The time it takes for conversion is variable.
◊ Voltage-Frequency ADC
◊ This device utilizes a Voltage Controlled Oscillator (VCO).
The higher the input voltage the higher the output
frequency of the VCO. The ADC enables the output of
the VCO for a specific period of time and the output of
the oscillator is connected to a counter. The result of the
count is supplied as the digital value.
Other Methods of ADC
◊ Creates an internal bitstream based upon the difference
(Delta) between the analog input and a feedback
comparator. The running total (Sigma) is reported as
the digitized value. Operates at a high internal
for more information.
◊ Analog to Digital systems have several possible
sources of error.
◊ The errors can be the result of:
◊ Quantization errors are a normal occurrence for
ADCs. The input voltage will commonly have a
value that is between the LSB voltage step and
must be rounded up or down.
◊ For instance, if the resolution is 20mV per bit and
a value of 10 mV is applied, the ADC will interpret
this as either 0v or 20 mV, depending on the ADC
◊ There are several additional quantization issues:
◊ If the total number of bits representing the amplitude of
the signal is too small the quantization error increases.
◊ If the voltage applied to the ADC is not properly
compensated then the full range of output values is not
Quantization and Sampling
◊ Sampling-Related Quantization Errors
◊ Relatively low sampling frequencies can lead to
◊ If the sampling frequency is slightly different from the
analog signal’s frequency (or a harmonic..a factor of the
frequency), then the values may not indicate all the
peaks and valleys, leading to other harmonics (also see
Not quite perfect...lower sample rate
Above: original signal with samples
Below: imperfections in the digital output quantization
Not quite perfect...higher sample rate
◊ Noise is a factor in analog communications and it
causes problems in ADC.
◊ The sources for noise are varied and can be
◊ Sources include:
◊ External Sources
◊ Internal circuit sources
Higher quality ADC (and DAC) will contain anti-aliasing filters
to remove frequencies that are aliasing (“posing”) as
Aliasing occurs when additional frequencies are inadvertently
produced in the AD process.
For instance, if there are some frequencies present in the
original signal that are higher than the Nyquist sample rate,
a different lower frequency may appear on the conversion.
Aliasing also occurs with adequate sampling frequencies.
Aliasing with Inadequate Sampling Rate
Aliasing with an Adequate Sampling Rate
◊ Analog values are susceptible to noise (an
advantage of digital electronics!). This noise
causes problems for the AD conversion, as the
noise will be included in the conversion process.
◊ There are external and internal sources of noise.
◊ External: use filters and proper techniques to minimize
noise (such as decoupling, isolation, etc)
◊ Internal: switching noise occurs with AD converters.
Use filters, decoupling and ground isolation. Thermal
noise (aka white noise) is caused by a variety of reasons.
◊ Other solutions include processing the signal (DSP),
injecting additional white noise to help average out
the noise and ignoring the LSB.
◊ The analog input to an ADC should contain a lowpass filter to ensure high frequencies cannot pass
and create aliasing problems.
◊ Advanced filtering techniques are also employed.
These include DSP (Digital Signal Processing) to
mathematically analyze and adjust the digital
A Gain Error occurs when the ADC produces an output at a
different quantization level than desired. The output is
linear but the steps are either larger or smaller than they
should be. The resolution is poor and there is the possibility
of clipping (the binary number stops changing with a change
in analog input). The voltage values will be inaccurately
voltage to the ADC
•Improper Analog input
◊ An Offset Error occurs when the output has the
same voltage per step but the starting voltage is
•Improper ground reference
•Improper analog voltage
◊ A Monotonic Error occurs when the individual
voltage steps are non-linear.
•wrong outputs measured
•significant noise on ac input
◊ Slew represents the maximum rate of change of a
signal. Slew rate is measured in V/t.
Black: ideal binary step from ADC
Grey: binary step from ADC with slew
Analog Sample & Hold
◊ The AD conversion process may be adversely
affected when an analog signal changes voltage
◊ Analog sample & hold circuits are able to capture
the analog value and retain it, allowing the ADC to
perform the conversion without input changes.
◊ View the specification sheet for the ADC08 from
◊ Common-Mode noise is noise that occurs on both
conductors at the same time.
◊ Ratiometric means measures the ratio, or takes ratio into
account when measuring.
◊ Digital Signal Processors perform operations on
the digitized signals including:
◊ removing interference, noise and other frequency and
◊ encoding the values for transmission
◊ reconstruct signals by enhancing or decreasing specific
◊ DSPs are complex devices that rely on
programming code (algorithms) to perform
ADC-DAC with DSP
◊ Analog signals contain noise. Noise can be filtered
but sometimes the filtering process can eliminate
◊ The better means of removing noise is to take an
average, as illustrated below as the red line.
Analog Signal with Noise
DSP (a simplified representation)
©Paul R. Godin