AN-847: Measuring a Grounded Impedance Profile Using the AD5933 (Rev. A) PDF

AN-847
APPLICATION NOTE
One Technology Way • P.O. Box 9106 • Norwood, MA 02062-9106, U.S.A. • Tel: 781.329.4700 • Fax: 781.461.3113 • www.analog.com
Measuring a Grounded Impedance Profile Using the AD5933
by Sean Brennan
INTRODUCTION
to the receive side transimpedance amplifier presents a virtual
earth voltage of VDD/2 to the output excitation voltage at
VOUT. Ohm’s law dictates that a signal current flows through
unknown impedance and into the feedback pin because of the
potential difference from VOUT to VIN. The current is transformed into a voltage on the output of the transimpedance
amplifier. The signal processing that follows on the output
voltage of the transimpedance amplifier allows the impedance
to be accurately calculated (see the AD5933 data sheet for
further details). The AD5933 can measure an impedance value
between 100 Ω and 10 MΩ to a system accuracy of 0.5% for
excitation frequencies between 1 kHz and 100 kHz.
This application note describes the circuit architecture and details
required to measure the impedance profile of a grounded sensor
using the AD5933 impedance-to-digital converter. The AD5933
is a high precision impedance converter system (see Figure 1)
that combines an on-board frequency generator with a 12-bit,
1 MSPS, analog-to-digital converter (ADC). The frequency
generator allows an external complex impedance to be excited
with a known frequency. The response signal from the impedance
is sampled by the on-board ADC, and a discrete Fourier
transform (DFT) is processed by an on-board DSP engine. The
DFT algorithm returns a real (R) and imaginary (I) data-word
at each output frequency. The magnitude of the impedance and
relative phase of the impedance at each frequency point along
the sweep are easily calculated using Equation 1 and Equation 2.
Magnitude =
R2 + I 2
A fundamental problem exists for certain applications where
the impedance to be analyzed must be connected between the
output (VOUT) and ground. In applications where the unknown
impedance/sensor being measured has one connection to
ground, with no signal return path, the AD5933 as shown in
Figure 1 cannot be used to analyze the impedance or sensor. This
document describes a circuit architecture using the AD5933 and
external components that allows the system designer to measure
the impedance profile of a grounded load. This application note
also describes the performance of such a system in measuring a
load that varies from 20 kΩ to 60 kΩ between 99.9 kHz and
100.1 kHz, running from a single supply of 5.0 V.
(1)
−1
Phase = Tan (I/R)
(2)
The AD5933 requires a calibration process before any valid
measurement can take place. The calibration process simply
requires that a measured impedance (for example, a precision
resistor) is connected between VIN and VOUT and a scaling
factor (gain factor) is calculated for subsequent measurements,
as described in the AD5933 data sheet. As shown in Figure 1,
the output pin (VOUT) of the transmit side is an ac excitation
voltage signal with an appropriate dc bias. The input pin (VIN)
MCLK
AVDD
DVDD
DDS
CORE
(27 BITS)
OSCILLATOR
DAC
ROUT
I2C
INTERFACE
TEMPERATURE
SENSOR
Z(ω)
AD5933
REAL
REGISTER
IMAGINARY
REGISTER
RFB
1024-POINT DFT
VIN
ADC
(12 BITS)
GAIN
LPF
VDD/2
AGND
06091-001
SCL
SDA
VOUT
DGND
Figure 1. AD5933
Rev. A | Page 1 of 12
AN-847
Application Note
TABLE OF CONTENTS
Introduction ...................................................................................... 1
Driving the AD8220 Reference Terminal ...................................8
Revision History ............................................................................... 2
Setting AD8220 Common-Mode Input Range .........................9
Overview ............................................................................................ 3
Grounded Impedance Measurement ........................................... 10
Circuit Operation Explained....................................................... 4
System Calibration ..................................................................... 10
Powering the Circuit .................................................................... 5
Grounded Impedance Calculation .......................................... 10
Choosing the Output Excitation Range..................................... 5
System Clock Settings ................................................................ 11
Selecting the Sense Resistor ........................................................ 6
Results .............................................................................................. 12
System Gain Factor ...................................................................... 7
Conclusion .................................................................................. 12
Instrumentation Amplifier Considerations ................................ 7
REVISION HISTORY
8/12—Rev. 0 to Rev. A
Changes to Figure 2 .......................................................................... 3
Changes to Figure 3 .......................................................................... 4
Changes to Selecting the Sense Resistor Section and changes
to Figure 5 .......................................................................................... 6
6/06—Revision 0: Initial Version
Rev. A | Page 2 of 12
Application Note
AN-847
OVERVIEW
The following sections detail the external circuitry surrounding
the AD5933, as shown in Figure 2, as well as the gain factor
required for the new system to successfully measure the
grounded impedance. Finally, this application note describes the
overall performance of the circuit shown in Figure 2 in measuring
a grounded impedance. The grounded complex impedance
measures between 20 kΩ and 60 kΩ over a frequency range of
99.9 kHz to 100.1 kHz. The required steps to perform a frequency
sweep are detailed in the AD5933 data sheet.
Figure 2 shows the circuit block diagram used to measure the
impedance profile of a grounded load. The circuit consists of
three major blocks.
The first block is a modified high-side current sense using a
precision single supply, rail-to-rail output, JFET instrumentation
amplifier (AD8220) to measure the current flowing through a
ense resistor connected in series with the grounded impedance.
The second block is the bias circuitry and a reference buffer to
ensure that the input sense signal and output signal do not
saturate the internal amplifiers of the AD8220 amplifier and the
receive side transimpedance amplifier of the AD5933.
Last there is the AD5933 impedance-to-digital converter.
DVDD
AVDD
MCLK
VDD
VBIAS
10µF,
0.1µF
0V
DDS
CORE
(27 BITS)
OSCILLATOR
VOUT
DAC
ROUT
SDA
I2C
INTERFACE
0.1µF
TEMPERATURE
SENSOR
1MΩ
6
AD8220
3
1
REAL
REGISTER
0.1µF
IMAGINARY
REGISTER
1MΩ
VCM
RFB
100kΩ
1024-POINT DFT
VIN
ADC
(12 BITS)
GAIN
VDD
8
4
SENSE
RESISTOR
AD5933
AD820
VCM
GROUNDED
IMPEDANCE
7
VDD/2 = VCM
2
5
RGAIN
VDD/2
0V
100kΩ
RSERIES
LPF
VDD/2
AGND
06091-002
SCL
VDD
VCM
DGND
Figure 2. Grounded Impedance Measurement Circuit
Rev. A | Page 3 of 12
AN-847
Application Note
CIRCUIT OPERATION EXPLAINED
wave output voltage signal. The AD8220 output voltage is
then connected to a series resistor, which in conjunction with
an equal value feedback resistor, RFB, results in an inverted
unity gain voltage being generated at the output of the receive
side current-to-voltage (I-to-V) amplifier within the AD5933.
Figure 3 shows a close-up of the modified high-side, current
sense used in the circuit. The AD5933 output excitation voltage
is developed across the sense resistor, which is in series with the
grounded load. The resulting current through the grounded
load is monitored by the AD8220 high precision in-amp, which
measures the difference voltage across RSENSE. The AD8220
sine wave output signal is symmetric about a set bias value of
VDD/2 and connected back into the AD5933 for digital signal
processing (DSP).
VDD
VBIAS
0V
10µF,
0.1µF
VOUT
VDD
VCM
0.1µF
1MΩ
4
1MΩ
VCM
7
3
1
RFB
The magnitude of the real and complex component given by
Equation 3 is a measure of the peak signal sampled by the
internal ADC at each frequency. To convert the peak signal to
a impedance figure, the magnitude of the real and imaginary
component is multiplied by a number called a gain factor and
subsequently inverted in a nearby processor to give the resulting
impedance value at each frequency.
VDD
6
AD8220
0.1µF
AD820
VCM
8
RSENSE
AD5933
The output voltage of the I-to-V amplifier is then low-pass
filtered and finally sampled by the 1 MSPS ADC. The ADC
output samples are windowed and presented to the internal
DSP, which performs a synchronized single-point DFT at the
same frequency as the AD5933 output excitation waveform.
The DFT output is a complex answer and is stored internally in
the form of a 16-bit, twos complement real and imaginary value
at each frequency of the sweep.
2
5
RGAIN
GROUNDED
IMPEDANCE
100kΩ
Magnitude =
R2 + I 2
(3)
100kΩ
VIN
RSERIES
06091-003
VDD/2
0V
Figure 3. Modified High-Side Current Sense Using an AD8220
Instrumentation Amplifier
Referring to Figure 3, with a system supply value (VDD) of 3.2 V,
the user can program the AD5933 output ac excitation voltage
of the AD5933 to be one of four values (2 V p-p, 1 V p-p,
400 mV p-p, or 200 mV p-p) by setting Bit D10 and Bit D9 in
the AD5933 control register (see the AD5933 data sheet for
more details). The dc bias of the output excitation in each case
is set according to the excitation voltage chosen. Both the
output ac excitation voltage and the corresponding dc bias
scales with the AD5933 supply voltage (VDD).
Therefore, the system designer can choose any one of the four
output voltages to act as the stimulus voltage signal to the
grounded load. The choice of excitation voltage depends upon
the value of the sense resistor and the span of the grounded
impedance and the common-mode input range of the AD8220.
The AD8220 instrumentation amplifier shown in Figure 3
measures the current through a sense resistor (RSENSE) which
is connected between the output pin of the AD5933 and the
remaining terminal of the grounded impedance. The current
flowing through the grounded load results in a voltage drop
across this sense resistor. This voltage is measured and
amplified by the AD8220 to provide a single-ended (with
respect to the AD8220 reference voltage, Pin 6 (REF) sine
As described in the AD5933 data sheet, the gain factor for the
AD5933 circuit (as shown in Figure 1) is a number determined
once at system calibration. The gain factor simply involves
placing a known impedance between the VIN and VOUT
pins and calculating the magnitude of the resulting code at a
midsweep frequency programmed by the user. The AD5933
system gain settings must be carefully considered so that the
optimum signal is presented to the ADC input without saturation of the ADC over the unknown impedance range being
measured. Therefore, the user must have prior knowledge of the
impedance span under test in order to calibrate the AD5933
correctly. Mathematically the gain factor is given by Equation 4.


1


Impedance
 Admittance  

Gain Factor = 
=
Code
Magnitude


(4)
Consider the following example of a gain factor calculation with
these assumptions:
Output excitation voltage = 2 V p-p
Calibration impedance value, ZCALIBRATION, = 200 kΩ
PGA gain = ×1
Current to voltage amplifier gain resistor = 200 kΩ
Calibration frequency = 30 kHz
The typical contents of the real and imaginary register after a
frequency point conversion would be:
Real register: = F064 hex = −3996 decimal
Imaginary register: = 227E hex = 8830 decimal
Rev. A | Page 4 of 12
Application Note
AN-847
Magnitude = (−39962 + (8830)2 = 9692.106
1




200
kΩ
 = 515.819E − 12
Gain Factor = 
 9692.106 




1
16
2
15
3
14
4
13
5
Because the circuit shown in Figure 2 is different from the
circuit shown in Figure 1, a new gain factor must used to
transform the AD5933 output codes into an impedance figure.
The new gain factor accounts for the fact that the circuit shown
in Figure 2 has a sense resistor that remains constant for the rest
of the measurement and the fact that the impedance under test
is no longer directly connected to the AD5933 input terminal
(VIN). The new gain factor is clearly described in the System
Gain Factor section.
Once the system designer accounts for the new gain factor, the
same calibration procedure remains: the system is calibrated at
a specific frequency, using a measured grounded load and sense
resistor the gain factor is calculated, and so any change in the
grounded load can be measured by the AD5933 output code.
The impedance profile can be determined in a nearby processor
either at a single frequency or over a range of frequencies.
POWERING THE CIRCUIT
The circuit in Figure 2 operates on a single supply of 5.0 V. The
AD5933 and AD8220 have a high PSRR specification. However,
for optimum performance, a stable dc voltage should be used
to power both chips because noise on the supply lines can
adversely affect circuit performance. The power supply to
both parts is decoupled using standard surface-mount, 0.1 μF
ceramic chip capacitors and 10 μF electrolytic tantalum
capacitors. The separate analog and digital grounds of the
circuit are connected at one point only. It is recommended to tie
all three supply pins of the AD5933 together and run from the
single supply (VDD = 5.0 V in this system). Similarly, it is
recommended to connect the three ground pins of the AD5933
together, as shown in Figure 2. The AD5933 and the AD8220
run from the same positive and negative supply, as shown in
Figure 4, so the system remains ratio-metric.
The AN-202 and AN-581 Application Notes provide more information on grounding and decoupling single supply amplifiers.
AD5933
GND
C1
0.1µF, 10µF
12
6
11
7
10
8
9
C1
0.1µF, 10µF
VDD = 5.0V
C1
1
8
2
7
4
AD8220
6
C1
5
0.1µF, 10µF
GND
06091-004
3
0.1µF, 10µF
Figure 4. AD5933/AD8220 Single Supply Configuration
CHOOSING THE OUTPUT EXCITATION RANGE
The transmit side of the AD5933 generates a sinusoidal output
voltage at VOUT. The AD5933 performs a linear frequency
sweep. The frequency sweep is defined by the contents of three
registers (start frequency, frequency increment, and number of
increments) preprogrammed by the user (see the AD5933 data
sheet). The user has the added flexibility to select one of four
possible output excitation peak-to-peak ranges by setting Bit D8
and Bit D9 in the control register. At a supply voltage of 3.2 V,
the AD5933 was designed to provide the four selectable
excitation ranges outlined in Table 1. The corresponding dc bias
of each excitation range is also outlined.
Table 1. AD5933 Output Excitation Ranges at VOUT
(VDD = 3.2 V)
Output Excitation
Voltage Amplitude
Range 1: 2 V p-p
Range 2: 1 V p-p
Range 3: 400 mV p-p
Range 4: 200 mV p-p
Output DC Bias Level
1. 6 V (VDD/2)
0. 8 V (VDD/4)
0.320 V (VDD/10)
0.160 V (VDD/20)
The receive side pin (VIN) of the AD5933 is an ac virtual earth
to the output excitation voltage at VOUT. Therefore, the input
(VIN) presents a constant hard-biased voltage of VDD/2 to any
of the chosen excitation signals given in Table 1. The reference
to the ADC is also approximately the supply range (0 V to
VDD), making the AD5933 a truly ratio-metric system. As the
supply voltage varies between 2.7 V and 5.5 V, the peak-to-peak
of the ac output excitation signal and the corresponding dc bias
increases/decreases in conjunction with the ADC reference, so
the output code is not affected by any supply variations. This
provides added immunity to any power supply drift over time.
Rev. A | Page 5 of 12
AN-847
Application Note
The peak-to-peak value (of Range 2) of the ac output excitation
voltage is scaled with the supply voltage (VDD) according to
Equation 5:
1. 0
Output Excitation Voltage ( p − p) ≡
× VDD
(5)
3. 2
Correspondingly, the dc bias value of the output excitation voltage is
also scaled with the supply voltage (VDD) according to Equation 6:
800 mV
(6)
Output Excitation Voltage ( p − p ) ≡
× VDD
3.2
The circuit shown in Figure 2 was designed to operate from a
supply of 5.0 V. Therefore, the peak-to-peak voltage of Range 1
at VDD = 5.0 V is given by Equation 7:
1. 0
Output Excitation Voltage ( p − p) ≡ 
× 5.0  = 1.5625 V p − p (7)
 3. 2

In addition, the bias voltage is given by Equation 8:
800 mV
Output Excitation Voltage (p − p) ≡
× 5.0 V = 1.25 V (8)
3.2
SELECTING THE SENSE RESISTOR
Once the AD5933 output excitation range for the circuit is selected,
choose a suitable sense resistor (RSENSE) that allows the AD8220 to
accurately measure the range of grounded impedance (20 k Ω to
60 kΩ). The voltage developed across the sense resistor is multiplied by the internal voltage gain of the AD8220 (set by the resistor
value seen between Pin 2 and Pin 3 of the AD8220) to set the peakto-peak ac output voltage. The AD8220 voltage gain helps the
system designer choose an appropriate RSENSE value.
The first requirement at this stage is to select a AD8220 gain
resistor (which determines the system bandwidth). The second
requirement is to select a sense resistor (RSENSE) that provides a
suitable AD8220 output voltage which is linear and symmetric
(about a bias point of VDD/2) over the entire unknown
impedance range, while preventing saturation of the internal
circuitry of the AD8220. Once these requirements are achieved,
the AD8220 output signal can be further amplified by selecting a
suitable combination of RFB/RSERIES and PGA setting such that the
dynamic range of the AD5933 internal ADC is used.
Figure 5 shows how to choose a suitable sense resistor (RSENSE)
to measure the grounded complex impedance. Having selected
Range 1 as the excitation range (a 1 V p-p sine wave biased at
VDD/4) and knowing that the grounded impedance range is
between 20 kΩ and 60 kΩ, the gain resistor of the AD8220 is
selected to set the required bandwidth and common-mode
rejection ratio (see the section on Instrumentation Amplifier
Considerations). The AD8220 voltage gain was set to 11 by
connecting a 4.67 kΩ resistor between Pin 2 and Pin 3, which
corresponds to a 3 dB point of 400 kHz. Therefore, when the
AD5933 output excitation frequency is sweeping between
99.9 kHz and 100.1 kHz, the AD8220 operates in the flat midband gain region with a good common-mode rejection of
almost 80 dB.
The circuit in Figure 5 is calibrated at the midpoint of the
impedance range (≈40 kΩ), to ensure that the peak-to-peak
AD8220 output voltage is linear and symmetric about this
impedance point. With this knowledge, the goal is to choose the
sense resistor to generate a 1 V p-p AD8220 output voltage for a
calibration grounded impedance of 40 kΩ and AD8220 gain of
11, as shown in Figure 5. Working through the circuit analysis,
this requires that the voltage (∆V) developed across the sense
resistor RSENSE and measured by the AD8220 be given by
∆V =
1.5625V p-p
= 90 mV p − p
(9)
DC =
AC =
39.9 kΩ
39.9 kΩ + 2.67 kΩ
× 1.25 V ≡ 1.1715 V
39.9 kΩ
× 1.5625 V p − p ≡ 1.4644 V
39.9 kΩ + 2.67 kΩ
(10)
(11)
VDD
VCM
1.25V
AD820
Thus, the total common-mode input voltage seen by the AD8220
input is a 1.46 V p-p sine wave biased on a 1.17 V dc level.
VDD
0V
VOUT
8
4
1V
6
AD8220
3
1
2
7
VDD/2
0V
REQUIRED OUTPUT
SIGNAL
5
RGAIN
06091-005
RSENSE
GROUNDED
IMPEDANCE
= 40kΩ
11
Knowing the value of the calibration grounded impedance
(40 kΩ, measured 39.9 kΩ), a simple resistor divider analysis
shows that an RSENSE value of 2.67 kΩ satisfies the requirement
to generate a 90 mV p-p signal across RSENSE, resulting in a
1 V p-p signal. The common-mode input of the AD8220
contains both an ac and dc signal due to the dc bias (1.17 V) of
the AD5933 ac excitation signal (1.46 V) and is given in by the
restive divider law shown in Equation 10 and Equation 11.
VDD
10µF, 0.1µF
1 V p−p
The voltage developed on the AD8220 output is therefore
(1.56 V − 1.46 V) × 11 ≈ 1.1 V p-p, which is close to the
required 1 V p-p output. The output ac signal is biased on a dc
voltage of (1.25 V − 1.1715 V) × 11 = 863 mV.
The output dc bias of 863 mV must be level shifted to VDD/2
by applying a buffered voltage at Pin 5 (VREF) of the AD8220 to
prevent internal saturation of the AD5933 impedance (see the
Instrumentation Amplifier Considerations section).
Figure 5. Choosing the Correct Sense Resistor (RSENSE) for the Calibration
Grounded Impedance
Rev. A | Page 6 of 12
Application Note
AN-847
SYSTEM GAIN FACTOR
INSTRUMENTATION AMPLIFIER CONSIDERATIONS
As explained in the Circuit Operation Explained section, the
gain factor for the proposed circuit in Figure 2 is different
from the AD5933 standalone configuration shown in Figure 1
(as shown in the data sheet). The new gain factor can be
determined by examining the output code of the AD5933 in
Figure 2. The AD5933 output code is proportional to voltage
presented to the input of the internal ADC (that is, the voltage
gain through the system), and is given by Equation 12:
The AD8220 voltage gain is determined by placing a resistor
across the RGAIN terminals, or more precisely, by whatever
impedance appears between Pin 1 and Pin 8. The AD8220
offers accurate gains using 0.1% to 1% tolerance resistors.
Table 2 shows required values of RGAIN for various gains. Note
that for G = 1, the RGAIN terminals are unconnected (RGAIN = ∞).
For any arbitrary gain, RGAIN can be calculated by using
Equation 15:
RFB
Output Code ∝ I × ∆V × In - Amp Gain ×
× PGA (12)
R
SERIES


where:
I is the current flowing from the AD5933 output pin (VOUT)
through the grounded load.
∆V is the differential voltage seen by the AD8220 across RSENSE.
PGA is the gain setting of the internal AD5933 programmable
gain amplifier.
RFB/RSERIES is the ratio of the AD5933 feedback resistor to
AD8220 output load resistor.
The AD5933 output code can now be written as follows:

100 kΩ 
1
× (V + − V − ) × 11 ×
× 1
Output Code ∝ VOUT ×
(Z + RSENSE )
100 kΩ 

(13)
The ratio RFB/RSERIES, PGA setting, and AD8220 gain settings
all remain constant once the system shown in Figure 2 has
been calibrated for the unknown impedance range under test.
RGAIN = 49.4 kΩ/(G − 1)
Table 2. AD8220 Gain Setting Resistor Values
1% Standard Value of RGAIN
49.9 kΩ
12.4 kΩ
4.7 kΩ1
2.61 kΩ
1.0 kΩ
499 Ω
249 Ω
100 Ω
49.9 Ω
Chosen resistor for circuit shown in Figure 2.
The AD8220 defaults to a gain of G = 1 when no gain resistor is
used (RGAIN = ∞). Gain accuracy of the AD8220 is determined by
the absolute tolerance of RGAIN. The temperature coefficient of the
external gain resistor increases the gain drift of the AD8220 output.
70
60
40
GAIN = +100
(14)
30
(15)
where:
Code is the magnitude of the real and imaginary component
stored given by Equation 1 at each subsequent sweep frequency.
RSENSE is the chosen sense resistor for the unknown impedance
range under test.
GAIN = +10
10
GAIN = +1
–10
–20
06091-006
Finally, the measured impedance of this new architecture
must account for the sense resistor, which is in series with
the grounded load (RSENSE remains constant throughout the
measurement) and needs to be subtracted in order to measure
the true grounded impedance at each frequency. This is given
by Equation 15:
20
0
where:
R and I are the real and imaginary components returned by the
AD5933 at the calibration frequency.


1
 − RSENSE
Z( f ) = 
 Gain Factor × Code 


GAIN = +1000
50
GAIN (dB)






Calculated Gain
1.990
4.984
10.51
19.93
50.40
100.0
199.4
495.0
991.0
1
The gain factor used in the circuit is given by the Equation 14:
1


Z
R
+
SENSE
Gain Factor = 
2
 R + I2


(15)
–30
–40
100
1k
10k
100k
1M
10M
FREQUENCY (Hz)
Figure 6. AD8220 Gain vs. Frequency
The requirement of this experiment is to measure the grounded
complex impedance between 99.9 kHz and 100.1 kHz. The gain
of the AD8220 determines the bandwidth of the output signal,
as shown in Figure 6. The AD8220 gain was set to ≈ ×11(that is,
≈ 20 dB) using a RGAIN value of 4.7 kΩ in order to operate within
the correct frequency range of the AD8220. Figure 7 shows that
the CMRR of the AD8220 is ≈75 dB in this region.
Rev. A | Page 7 of 12
AN-847
Application Note
DRIVING THE AD8220 REFERENCE TERMINAL
GAIN = +1000
GAIN = +100
CMRR (dB)
120
GAIN = +10
BANDWIDTH
LIMITED
100
GAIN = +1
80
06091-007
60
40
10
100
1k
10k
100k
FREQUENCY (Hz)
Figure 7. AD8220 CMRR vs. Frequency
As with all instrumentation amplifiers, when the AD8220 is
used with a low single supply voltage, high amplifier gains can
render the in-amp circuit inoperative. Figure 8 shows this point.
It most commonly occurs when in-amps are operating at high
gains, such as 1000. Under these circumstances, a 10 mV p-p
input times a gain of 1000 creates a 10 V p-p signal between the
outputs of A1 and A2. When using dual ±15 V supplies, this
situation can occur. However, the circuit shown in Figure 2 is
restricted to using a 5 V single supply, so the circuit becomes
dysfunctional. Because users of monolithic ICs do not have
access to the buffer outputs of the AD8220, A1 and A2, they
only see what is happening at the final output—the output of A3.
The reference terminal potential of the AD8220 defines the zero
output voltage and is especially useful when the load does not
share a precise ground with the rest of the system. It provides a
direct means of injecting a precise offset to the output. The
reference terminal is also useful when bipolar signals are being
amplified, as it can be used to provide a virtual ground voltage.
The voltage on the reference terminal can be varied from
ground to +VS. A common application problem that must be
avoided in this experiment is when system designers try to
drive the reference pin of an AD8220 in-amp with a high
impedance source. Typical values for the impedance of the
reference input in many popular in-amps are 20 kΩ to 125 kΩ.
If a low impedance source, such as an op amp, is directly
driving the reference, there is no problem, but often, an unwary
designer tries to use a resistive voltage divider as a low cost,
ratio-metric reference and ends up introducing serious errors.
IN-AMP
SUBTRACTOR
VIN–
A1
A2
VIN+
RREF1
5V + CMV1
20kΩ
VOUT
A3
R2
20kΩ
A1
24.7kΩ
INVERTING
INPUT
SIGNAL VOLTAGE
10mV p-p
NOINVERTING INPUT
COMMON-MODE
ERROR VOLTAGE
50Ω
R1 EXTERNAL
REFERENCE
VOLTAGE
Figure 9. Using a Buffer or a Reference ADR42x
A3
VOUT
24.7kΩ
A2
20kΩ
EXTERNAL
VOLTAGE
DIVIDER
RREF2
20kΩ
1WITH 10mV p-p INPUT SIGNAL APPLIED, OUTPUT FROM A1 AND A2
IS 5mV × 1000 = 5V + THE COMMON-MODE VOLTAGE.
06091-008
–5V + CMV
Figure 8. Simplified AD8220 Internal Block Diagram
Another common application problem stems from operating
standard, non-rail-to-rail devices, from low single-supply
voltages. A high quality rail-to-rail in-amp, such as the AD8220,
with a gain of 1 can swing its output to only within 0.15 V of the
positive supply line and down to 0.2 V above ground. Its inputvoltage range can only swing to within 2 V of the positive
supply. With higher gains, the amplifier’s output swing does not
reach the supply voltage. Therefore, when using a 5 V single
supply, the amplifier has approximately a 3 V output swing.
Unfortunately, some system designers forget about amplifier
headroom and use standard, non-rail-to-rail products in these
applications. Even a good dual-supply in-amp has an output
swing within only about 2 V of either rail.
The reference input in the AD8220 is part of the outputsubtracter circuit in a typical three-op-amp instrumentation
amplifier. As such, it has a finite input resistance, approximately
equal to RREF1 plus RREF2, typically. Adding external resistor
R2, between the reference terminal and common ground,
as shown in Figure 9, unbalances the output (A3 ) subtracter
circuit, introducing a common-mode error. An obvious way
to minimize this problem is to reduce the value of R2 to
approximately 0.1% of RREF1 plus RREF2 (for 72 −dB CMR).
However, with 20 kΩ values of RREF1 and RREF2 (40,000 total
input Z), R2 must be 40 Ω. This value, in turn, unnecessarily
burns large amounts of supply current in the voltage-divider
network. There is also the issue of RREF1 and RREF2 shunting R2
and causing a reference-voltage error. The solution is to use a
simple buffer like the AD820 and drive it with a potentiometer
run from the 5.0 V supply, as shown in Figure 10.
Rev. A | Page 8 of 12
REFERENCE
INPUT
RREF
IN-AMP
R1
OP AMP
BUFFER
AD820
R2
VDD OR
EXTERNAL
REFERENCE
VOLTAGE
EXTERNAL
VOLTAGE
DIVIDER
Figure 10. Buffer the Reference Input
06091-010
140
06091-009
160
Application Note
AN-847
SETTING AD8220 COMMON-MODE INPUT RANGE
+5V SINGLE SUPPLY,
VREF = +2.5V
+3V
3
2
+0.1V, +1.7V
+4.9V, +1.7V
1
0
+0.1V, –0.5V
–0.3V
+4.9V, –0.5V
06091-0 11
INPUT COMMON-MODE VOLTAGE (V)
4
–1
–1
0
1
2
3
4
REQUIRED OUTPUT VOLTAGE (V)
5
6
Figure 11. AD8220 Required Output Voltage vs. Common-Mode
Input Voltage
A common, incorrect procedure is to connect a capacitor in
series with each in-amp input terminal. Again, because a
monolithic in-amp is a complete package, system designers
often fail to realize what is inside the IC. In-amps like the
AD8220 connected with these floating inputs do not have a dc
reference. The input bias currents charge up the ac coupling
capacitors until they exceed the input common-mode voltage.
In other words, the capacitors charge up to the supply line or
down to ground, depending on the direction of the input bias
currents. With a FET input device and high value capacitors,
it could take several minutes before the in-amp becomes
inoperative. Therefore, a casual lab test might not detect this
problem, thus it is important to avoid it altogether.
C1
–IN
R1
C2
VOUT
IN-AMP
VCM
R2
REFERENCE
INPUT
+IN
V
VREF (TYPICALLY, DD )
2
06091-012
Limits on AD8220 common-mode range, coupled with the fact
that in this application the common-mode input signal to the
AD8220 changes depending on the value of the grounded
impedance, forces the user to ac-couple the inputs of the inamp circuit in this application. In this experiment, the AD8220
is powered by a single 5 V supply, so ac coupling is more
complicated and requires applying a dc common-mode voltage,
VCM, to both inputs (Figure 12). This step is necessary because
the in amp output cannot swing below the negative supply—in
this case, ground—as shown in Figure 11. Here, if the in-amp
output voltage tries to swing more than a few millivolts
negative, the signal is clipped.
VDD
Figure 12. AC Coupling the AD8220 Inputs
Choosing appropriate voltages for VCM and VREF is the next
important design consideration, especially in low supply voltage
applications. In general, set VCM to the middle of the expected
input dynamic range and center VREF on the expected output
dynamic range. Output centering is similar: estimate the
amount and direction of the in-amp output swing—in most
cases, ∆V× AD8220 gain +VCM—and then apply a reference
voltage at VREF that is in the center of that range. Look at the
voltage swing at the impedance range limits. The reference
applied to Pin 5 of the AD8220 is 2.5 V for a supply voltage of
5 V. The choice of dc-return-resistor value for ac-coupled
circuits is a trade-off between offset errors and the physical and
electrical size of the input coupling capacitors. The larger
the value of the input resistor, the smaller the required input
coupling capacitor. This approach saves both money and PCB
space. However, the trade-off is that high value input resistors
increase the offset voltage error due to input offset currents.
Offset voltage drift and resistor noise also increase. With lower
resistor values, higher value input capacitors for C1 and C2 are
necessary to provide the same –3 dB corner frequency, as
shown in Equation 16.
F − 3 dB = (1/(2πR1C1))
where R1 = R2 and C1 = C2.
Rev. A | Page 9 of 12
(16)
AN-847
Application Note
GROUNDED IMPEDANCE MEASUREMENT
The circuit shown in Figure 2 was developed and used in the
measurement of grounded complex impedance. The transmit
side of the AD5933, used as a voltage excitation signal, drives the
sense resistor and the grounded impedance. As explained in the
Introduction, the AD5933 integrated circuit impedance converter
combines an internal DDS frequency generator and ADC to form
a self-contained impedance measurement system. A frequency
sweep is performed by the AD5933 to typically gather magnitude
and phase data at frequencies defined by the user. In Figure 2, the
impedance to be analyzed is placed between the sense resistor
and ground. The AD8220 is used to measure the current in the
sense resistor. The AD8220 output signal is then passed through a
programmable gain amplifier (PGA), filtered and then delivered
to the 12-bit ADC. After the receive signal is digitized by the
ADC, a DFT is performed on the data.
A nearby microcontroller communicates to the AD5933 via the
I2C interface, allowing the user to program the AD5933 sweep
parameters (start frequency, frequency step size, and number
of points) configure the control register, adjust the excitation
amplitude and PGA setting, as well as read back the measured
data from the AD5933 for the final impedance calculation.
Once the AD5933 is correctly programmed, only a single bit
in the status register at each point in the user defined sweep
must be polled to see if valid data is available to read from the
AD5933 (see the AD5933 data sheet for more details).
SYSTEM CALIBRATION
However, prior to a valid impedance measurement, the AD5933
system shown in Figure 2 must undergo a calibration process.
The calibration process requires that a known precision impedance be substituted for the subsequent unknown impedance to
be measured and a scaling factor (gain factor) is calculated for
subsequent measurements. As described in the System Gain
Factor section, the gain factor for this circuit is given in
Equation 17:
1


Z
CALIBRATION + RSENSE
Gain Factor = 

R2 + I 2






where:
R and I are the contents of the real and imaginary register
(94h to 97h) at the chosen calibration frequency point.
ZCALIBRATION is the known calibration impedance.
RSENSE is the chosen sense resistor.
The gain factor is calculated by dividing the sum of a known
precision impedance and sense resistor by the magnitude of
the real and imaginary data returned at a suitable midpoint
frequency in the sweep. Both the real and imaginary component are stored in two, 16-bit registers which must be read
after each ADC conversion, and before the next frequency point
in the sweep, where the contents of these two registers are
refreshed with new data.
The grounded complex impedance ranges between 20 kΩ and
60 kΩ. Therefore, the calibration impedance was chosen to have
a value of ≈40 kΩ and the sense resistor was 2.67 kΩ.
GROUNDED IMPEDANCE CALCULATION
Once the calibration process is finished, the unknown complex
impedance replaces the calibration impedance. After the user
issues a start frequency sweep command to the control register,
the AD5933 automatically sequences through the user-defined
frequency sweep. The frequency sweep is calculated by contents
of the three registers (start frequency, frequency step, and
number of increments register). Finally, the microprocessor
communicating to the AD5933 calculates the grounded
impedance at each frequency point by multiplying the gain
factor by the magnitude of the complex code returned at each
frequency by the AD5933 and inverting the answer. The sense
resistor is subtracted in the final calculation. The impedance at
each frequency point is given by Equation 18.
Z Complex Impedance =
1
Gain Factor ×
(R 2 + I 2 )
− RSENSE (18)
Where R and I are the contents of the real and imaginary
registers (94h to 97h) at each frequency sweep point.
(17)
Rev. A | Page 10 of 12
Application Note
AN-847
SYSTEM CLOCK SETTINGS
The frequency of the clock applied to MCLK (Pin 8) is set to
16 MHz using a stable, low jitter, quartz crystal oscillator. This
allows the AD5933 to analyze excitation frequencies from
99.9 kHz to 100.1 kHz. Table 3 lists the programmed sweep
parameters (start frequency, frequency increment, and number
of increments) used in the test to cover the 99.9 kHz to 100.1 kHz
bandwidth. The frequency step size was set to 1 Hz and the
number of sweep points was set to 200. The number of settling
time cycles register was set at 15 output cycles throughout
the experiment. The circuit in Figure 2 was calibrated at a
frequency of 100 kHz.
As described in the AD5933 data sheet register map, the start
frequency is a 24-bit word that is programmed to the on-board
RAM at Address 82h, Address 83h, and Address 84h. The
required code loaded to the start frequency register is the result of
the formula in Equation 19, based on the master clock frequency
and the required start frequency output from the DDS.
Start


 Required Output Start
Frequency Code = 
 MCLK 





4 


Frequency



 × 227



(19)
For example, refer to the first column of Table 3. If the user
requires the sweep to begin at 99.9 kHz and has a 16 MHz clock
signal connected to MCLK, the code that must be programmed
is given by Equation 20:
Start




 99.9 kHz 
 × 227
Frequency Code = 
  16 MHz  
 
4  

≡ 332617 hexadecimal
(20)
The user programs 33 hex to Register 82 h, 26 hex to Register 83 h,
and 17 hex to Register 84 h.
Similarly, the frequency increment register is a 24-bit word
that is programmed to the on-board RAM at Address 85 h,
Address 86 h, and Address 87 h (see the AD5933 data sheet
register map). The required code loaded to the frequency
increment register is also the result of the formula shown in
Equation 21, based on the master clock frequency and the
required increment frequency output from the DDS.
Frequency Increment Code =


 Required






Frequency Increment  27
×2
 MCLK 






4 


(21)
For example, if the user requires the sweep to have a resolution
of 1 Hz and has a 16 MHz clock signal connected to MCLK, the
code that needs to be programmed is given Equation 22:
Frequency Increment Code =





 1 Hz
 ≡ 000021

  16 MHz  


 

4  

hexadecimal
(22)
The user programs 00 hex to Register 85 h, 00 hex to Register 86 h,
and 21 hex to Register 87 h.
The third parameter to define the frequency sweep is the
number of increments register. This is a 9-bit word that
represents the number of frequency points in the sweep. The
number is programmed to the on-board RAM at Address 88 h
and Address 89 h. The maximum number of points that can be
programmed is 511. For example, if the sweep needs 200 points, the
user programs 00 hex to Register 88 h and C8 hex to Register 89 h.
Table 3 shows a summary of the required sweep codes for a
16 MHz clock frequency. The settling time cycles register was
set to a value of 15 by programming 0F hex to Register 8B hex.
Table 3. AD5933 Required Sweep Codes for Frequency Range 99.9 kHz to 100.1 kHz
Programmed
Start Frequency
99.9 kHz
Required Start
Frequency Code
332617 hex
Programmed
Frequency
Increment
1 Hz
Required
Frequency
Increment Code
000021 hex
Programmed No.
of Increments
200
Rev. A | Page 11 of 12
Required No. of
Increments Code
00C8 hex
Clock Frequency
Applied to MCLK Pin
16 MHz
AN-847
Application Note
RESULTS
The system in Figure 2 was calibrated with a 35 kΩ impedance
and a sense resistor of 2.67 kΩ ( midpoint between 20 kΩ and
60 kΩ), and the gain factor was calculated at a calibration
frequency of 100 kHz ( the midpoint in the sweep) using
the sweep dial up codes outlined in Table 3. The calculated
calibration gain factor was stored in memory in a nearby
microcontroller. The calibration impedance was replaced by a
range of complex grounded impedances which varied between
20 kΩ to 60 kΩ. A frequency sweep was carried by the AD5933,
and the measured impedance was calculated at each frequency
point by multiplying the stored gain factor by code at each
frequency and inverting the answer, as shown in Equation 23
and Equation 24.
1




CALIBRATION + RSENSE
Z


(23)
Gain Factor =


R2 + I 2


Z Complex Impedance =
1
Gain Factor ×
(R 2 + I 2 )
− RSENSE (24)
Figure 13 shows that at the calibration frequency of 100 kHz,
the percentage impedance error is zero at the calibration
impedance of 35 kΩ and increases from this point. The results
of the measurement in Figure 13 show that there is less than 7%
impedance error over the entire 60 kΩ range as a result of the
single-point calibration. A reduction in error can be achieved
by calibrating over a reduced impedance range.
CONCLUSION
The AD5933 provides an accurate low cost solution to
grounded impedance measurement compared to the traditional
and expensive multichip-based solutions. Along with the
AD5933, only a few external support components are required
to incorporate the simple test circuitry to a nearby grounded
sensor with minimum board space. The impedance profile of
the grounded sensor can be evaluated upon system power-up
with minimal effort, providing a simple means of characterizing
the sensor impedance profile, examining the effects of aging
and damage, and identifying changes.
A plot of percentage impedance error vs. frequency for the
grounded impedance at the calibration frequency of 100 kHz
is shown in Figure 13.
18
FREQUENCY = 100kHz
IMPEDANCE ERROR (%)
16
14
12
10
8
6
4
06091-013
CALIBRATION IMPEDANCE = 35kΩ
2
0
60
55
50
45
40
35
30
25
20
IMPEDANCE (k Ω)
Figure 13. Grounded Impedance Measurement Results by AD5933 at 100 kHz
I2C refers to a communications protocol originally developed by Philips Semiconductors (now NXP Semiconductors).
©2006–2012 Analog Devices, Inc. All rights reserved. Trademarks and
registered trademarks are the property of their respective owners.
AN06091-0-8/12(A)
Rev. A | Page 12 of 12