Document Text (Pages 61-70) Back to Document

Integrated Circuit Interface for SAW Biosensors Applications

by Aggour, Khaled, MS


Page 61

Chapter 4: SAW Sensor Interface Circuit Design 48

4.4 Frequency to Digital Conversion

The mixer sinusoidal output signal should be converted to a square wave with the
same frequency for digital processing.

Vdd

Bandgap cell Bias cell

Vdd

Vinp

Mixer
C1 Comparator

Square wave
output

C2
Vinn

Figure 4- 13 Mixer output to square wave conversion

The circuit schematics used for this conversion is shown in figure 4-13. The
circuit utilises a comparator which produces either Vdd or ground depending on the
polarity of the difference between its positive and negative input terminals Vinp and

Vinn, respectively. The comparator is an analogue cell (acmpc02) from XFAB XC06

process (see appendix A.3.3). The comparator cell is biased using a biasing cell
(abiac01) which in turn biased by a band gap reference cell (abgpc01) (Appendix
A.3.2 and A.3.1).

The mixer output is AC coupled through the capacitor C1, and feeds one of the
comparator terminals. A delayed version of the mixer output signal is AC coupled
via capacitor C2. Figure 4-14 shows both the mixer output signal and its delayed
version. The difference signal (Vinp – Vinn) is shown also. It can be seen that the
difference signal is a sinusoidal signal with changing polarity around zero.

When the difference signal is positive, the comparator output is high (VDD).
The comparator output is low when the difference is negative. So a square wave is
generated with the same frequency as the mixer output frequency. The square wave
is now ready for digital processing. The comparator input and output signals are
shown in figure 4-15.


Page 62

Chapter 4: SAW Sensor Interface Circuit Design 49

Figure 4- 15 Comparator input signals (mixer output and its delayed version)

Figure 4- 14 Comparator input and output signals


Page 63

Chapter 4: SAW Sensor Interface Circuit Design 50

The block diagram of the square wave conversion to the digital word is shown
in figure 4-16. The square wave is applied as a clock to a digital 16-bit counter, the
counter ENABLE signal is chosen with a long period to increase the resolution
(Gardner et al., 2001). Since the required minimum frequency resolution is 100 Hz
(10 ms period), the ENABLE signal will have a pulse width of 10 ms and a period of
20ms (10ms × 2).

Figure 4- 16 Square-wave to digital word conversion

The counter output word will be the signal frequency divided by 100 Hz.

The mixer difference frequency fmixer is calculated in terms of the counter
output word Bmax as follows:


����� = 100 ×
��� eq.4.25

From equation 4.24 it can be seen that the system is capable of measuring
frequencies up to


��� = 100 × (
���) = 100 × (2
�� 1) = 6.5535 푀퐻푧 eq.4.26

fmax is the maximum measurable frequency.

Bmax is the maximum counter output word.

The clock used for the ENABLE signal is a 50Hz clock, the XFAB (arcoc03)
oscillator is used to generate the required clock. The analogue cell is an RC oscillator
that generates a clock of a frequency of 10.06 + 0.015 (due to its TCF) kHz (see
appendix A.3.4 for the cell details).

A digital counter is implemented in VHDL with its input as the 10.075 kHz
oscillator clock output (Appendix B.1). Since the required ENABLE signal is 50 Hz,
a division by 215 is required (10.075×1000/ 50). The VHDL code of the clock


Page 64

Chapter 4: SAW Sensor Interface Circuit Design 51

divider is shown in Appendix B.2. The functional diagram of the 50Hz clock
generation is shown in figure 4-17.

The counter is set to produce an output for each clock transition, so counter
output is incremented twice each clock. After the enable voltage is applied, the
counter resets and produces a logic 1 output. The counter counts till the count
reaches 215 (107.5×2). At this point the counter produces an output low signal (logic
0). The counter continues counting clock event signals from the 10.075 kHz
oscillator until the count reaches 430 (215×2). After reaching this value (f= (107.5
kHz×2) /430 =50Hz); the counter produces an output high signal (logic 1) and sends
a reset order to reset the count value to 0. The process continues as long as the
ENABLE signal is still high.

Count=107.5

Output High

10kHz RC oscillator
cell
Clk_in Counter

Output Low

Count=215

Reset

Figure 4- 17 Functional diagram of 50Hz clock generation

4.5 Temperature Control

SAW device sensitivity to temperature varies with the material. The 'surrounding
temperature' of the SAW resonator will change; therefore there is a need for a
temperature control mechanism to keep the SAW resonator at a constant
temperature.

4.5.1 Temperature effect and control circuit

The gold Temperature Coefficient of Resistance (TCR) depends on its
thickness and ranges between 1.5 × 10-3 °
-1C and 7 × 10-3 °-1C as shown by Nayak et
al. (1993). The experimental measurement of the gold heater indicated 6 × 10-3 °-1C
TCR for the gold heater. Figure 4-18 shows the SAW resonator and heater chip.

The SAW heater is a heating element that is used to control the temperature.
The temperature control block diagram is shown in figure 4-19.


Page 65

Chapter 4: SAW Sensor Interface Circuit Design 52

IDTs

Heater

Figure 4- 18 SAW resonator and heater chip photo

Figure 4- 19 Temperature control circuit

The main concept of the circuit is to keep the SAW heater resistance Rh
constant. This is done by changing its heating power via controlling the NMOS
transistor gate voltage.

Since the SAW heater resistance may vary with the process, and the required
temperature also varies with the application; it will be beneficial to provide a flexible


Page 66

Chapter 4: SAW Sensor Interface Circuit Design 53

temperature control mechanism capable of controlling a range of heater resistances
and a span of temperatures.

To sweep the temperature in the required range 200C to 500C; an experiment
was carried out to measure the necessary heating power and measure also the heater
temperature coefficient of resistance accurately. In the experiment, a variable DC
voltage was applied across heater terminals and its resistance and temperature were
measured.

Resistance
(Ω)
450

445

440

435

430

425

420

415

410

Heater Resistance vs. Temperature

16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

Temperature (°C)

Figure 4- 20 Experimental results for Resistance vs. Temperature

Figure 4-20 indicates the resistance change with temperature. The resistance
changes with temperature according to the following formula (assuming only first
order effect):
() =
(1+ (푇 − 푇
)) eq. 4.27

Where α is the temperature coefficient of resistance of the material.

T is the temperature in Celsius.

A typical TCR of gold is about 3.4 x 10-3 C-1 (Chopra, 1969 cited Barwinski
1990, p.167). From the experiment results in figure 4-20 it can be found that the
temperature coefficient of resistance αh of the gold heater is equal to 6 x 10-3 °C-1.
The experiments were carried out with LiTaO3 SAW resonators working at 60MHz
resonant frequency.

A poly silicon resistor (poly 1 from Appendix A.2.2) is placed in series as a
constant resistance. The poly silicon resistance TCR αp is 1.1 × 10-3 C-1 (less than


Page 67

Chapter 4: SAW Sensor Interface Circuit Design 54

fifth of the heater) to enable the calculation of the heater current by calculating the
voltage across the poly silicon resistance. The resistance was chosen to be the
process minimum (2Ω) in order to avoid unnecessary power dissipation.

The heater resistance is calculated from the following formula:

=


=


����



=

����

eq. 4.28

Vh and Vp are the voltages across the heater and poly resistor respectively and

Ih is the current flowing in the circuit.

70

60

Temperature vs. heating power

50

40

30

20

10

0

0

20

40

60

80

100

120

140

Temperature
(°C)

160

180

200

220

240

260

280

300

320

Heating Power (mW)

Figure 4- 21 Experimental results for Temperature vs. Heating Power

Figure 4-21 shows the temperature changes according to the heating power. It
can be seen that 220mW heating power is capable of raising the temperature by
thirty degrees.

To allow applying such high power at low DC supply (5V); which in turn will
limit the maximum heater voltage to about 2.5V; the heater resistance should be
small. The power dissipated at the heater is calculated as:


=


/eq.4.29

Substituting maximum power and voltage the required resistance is:


= ��


= .�� 30 Ω

..�� eq.4.30


Page 68

Chapter 4: SAW Sensor Interface Circuit Design 55

So 30Ω is the maximum allowed heater resistance with a nominal value of 25Ω
to be able to control resistance from (20° to 50°). This value is much smaller than
that of the actual fabricated SAW resonators. The heater resistance minimisation
should be taken into consideration in the SAW resonator design so the proposed
temperature control mechanism can work properly.

as:

The power-temperature curve can be approximated by a second order equation


() =
+
− 푇
+
− 푇
eq. 4.31

A and B are constants and Ti, Tf are the initial and final temperatures and Pi, Pf
are the corresponding heating powers at Ti and Tf respectively.

From the experiment, setting A= 17.9462, B= -0.93289 gives a good curve fit.

The power dissipated in the heater resistance Ph is calculated as follows:

=
×
= ×


eq. 4.32

The system uses the heater resistance as a measure for the temperature.

4.5.2 Temperature Control logic

The temperature control logic flow chart is shown in figure 4-22. The control logic
works as follows:

At first, an offline measurement of the SAW heater resistance is carried out at
the required temperature. An online calibration is done by activating an
asynchronous input bit labelled ENABLE CALIBRATION. When this bit is high,
the user should manually enter the required temperature and the heater resistance at
this temperature. Entering these data is performed as follows.

To avoid using many input pins; a single pin is used for the inputs and the
inputs are entered in a serial manner at the rising edge of an input clock labelled
ACTIVATE. The input is entered in Binary Code Decimal (BCD) format with the
ACTIVATE signal rising edge. After all inputs are entered, both required
temperature and corresponding heater resistance, the ENABLE CALIBRATION
signal should be turned to low. If no calibration was done, the system asynchronous
reset will set the required heater resistance to 25Ω and the required temperature to
27°C.


Page 69

Chapter 4: SAW Sensor Interface Circuit Design 56

Figure 4- 22 Temperature control logic flow chart

In order to achieve a good resolution; the heater resistance has four decimal
digits, so the actual heater value is entered multiplied by 10
4. The temperature is two
decimal digits (actual temperature multiplied by 100).

After the user puts the required Rhf and Tf; the system stores their values in a
register. Measuring the heater and poly voltages using two 10-bit XFAB (aadcc01)
Analogue to digital converter cells (appendix A.3.6).


Page 70

Chapter 4: SAW Sensor Interface Circuit Design 57

The output value of the Analogue to Digital converter VADC is calculated as:


��� = ��
eq.4.33

The current temperature Ti is calculated as shown in the flow chart after
calculation of the heater resistance Rhi.To calculate the current temperature, equation
4.26 is used. Re-arranging the equation


=
+ ���

�����


eq.4.34

The poly resistance was assumed constant at first. Using the poly resistance
temperature coefficient αp as follows:


=
��(1+ (
− 푇
)) eq.4.35

Using the new poly silicon resistance value Rp, more accurate values for the
heater resistance and temperature are calculated using equation 4.28 and 4.34,
respectively.

The next step is to calculate the heater power as shown in the flow chart Pi.
This power is used to calculate the required power value Pf from equation 4.31.

The required heater power is related to the heater voltage as shown in equation
4.28. To relate the heater voltage to the gate control voltage; a simulation was done
at different heater resistances and the result is shown in figure 4-24.

From figure 4-20, it can be noticed that the heater voltage-control voltage
relation can be approximated by a linear one, since the transistor works in the triode
region in this range. So the following equation can be used to calculate the control
voltage.


= 퐶 푉
eq.4.36

C is the proportionality constant.

After calculating the proportionality constant C, the required control voltage
needs to be calculated.

Substituting into 4.28, the required control voltage becomes:

�� = ����


eq.4.37

Substituting in 4.36 by Vh1 and VC1 and using C value in 4.36 yields:

© 2009 OpenThesis.org. All Rights Reserved.