Data Sheet AD7877
Rev. D | Page 17 of 44
It is only during the sample or acquisition phase of the ADC
operation of the AD7877 that noise from the LCD screen has an
effect on the ADC measurements. During the hold or
conversion phase, the noise has no effect, because the voltage at
the input of the ADC has already been acquired. Therefore, to
minimize the effect of noise on the touch screen measurements,
the ADC acquisition phase should be halted.
The LCD control signal should be applied to the STOPACQ pin.
To ensure that acquisition never occurs during the noisy period
when the LCD signal is active, the AD7877 monitors this signal.
No acquisitions take place when the control signal is active. Any
acquisition that is in progress when the signal becomes active is
aborted and restarts when the signal becomes inactive again.
To accommodate signals of different polarities on the
STOPACQ pin, a user-programmable register bit is used to
indicate whether the signal is active high or low. The POL bit
is Bit 3 in Control Register 2, Address 0x02. Setting POL to 1
indicates that the signal on STOPACQ is active high; setting
POL to 0 indicates that it is active low. POL defaults to 0 on
power-up. To disable monitoring of STOPACQ, the pin should
be tied low if POL = 1, or tied high if POL = 0. Under no
circumstances should the pin be left floating.
The signal on STOPACQ has no effect while the ADC is in
conversion mode, or during the first conversion delay time.
(See the Control Registers section for details on the first
conversion delay.)
When enabled, the STOPACQ monitoring function is imple-
mented on all input channels to the ADC: AUX1, AUX2, BAT1,
BAT2, TEMP1, and TEMP2, as well as on the touch screen
input channels.
TEMPERATURE MEASUREMENT
Two temperature measurement options are available on the
AD7877: the single conversion method and the differential
conversion method. The single conversion method requires
only a single measurement on ADC Channel 1000b. Whereas
differential conversion requires two measurements, one on
ADC Channel 1000b and a second on ADC Channel 1001b.
The results are stored in the results registers with Address 11000b
(TEMP1) and Address 11001b (TEMP2). The AD7877 does not
provide an explicit output of the temperature reading; the
system must perform some external calculations. Both methods
are based on an on-chip diode measurement.
Single Conversion Method
The single conversion method makes use of the fact that the
temperature coefficient of a silicon diode is approximately
−2.1 mV/°C. However, this small change is superimposed on
the diode forward voltage, which can have a wide tolerance. It
is, therefore, necessary to calibrate by measuring the diode
voltage at a known temperature to provide a baseline from
which the change in forward voltage with temperature can be
measured. This method provides a resolution of approximately
0.3°C and a predicted accuracy of ±2.5°C.
The temperature limit comparison is performed on the result in
the TEMP1 results register, which is simply the measurement of
the diode forward voltage. The values programmed into the high
and low limits should be referenced to the calibrated diode for-
ward voltage to make accurate limit comparisons. An example
is shown in the Limit Comparison section.
Differential Conversion Method
The differential conversion method is a 2-point measurement.
The first measurement is performed with a fixed bias current
into a diode (when the TEMP1 channel is selected), and the
second measurement is performed with a fixed multiple of the
bias current into the same diode (when the TEMP2 channel is
selected). The voltage difference in the diode readings is pro-
portional to absolute temperature and is given by the following
formula:
∆VBE = (kT/q) × (ln N) (4)
where:
VBE represents the diode voltage.
N is the bias current multiple (typical value for AD7877 = 120).
k is Boltzmann’s constant.
q is the electron charge.
This method provides a resolution of approximately 1.6°C, and
a guaranteed accuracy of ±4°C without calibration. Determina-
tion of the N value on a part-by-part basis improves accuracy.
Assuming a current multiple of 120, which is a typical value
for the AD7877, taking Boltzmann’s constant, k = 1.38054 ×
10−23 electrons V/°K, the electron charge q = 1.602189 × 10−19,
then T, the ambient temperature in Kelvin, would be calculated
as follows:
∆VBE = (kT/q) × (ln N)
T°k = (∆VBE × q)/(k × ln N)
= (∆VBE × 1.602189 × 10−19)/(1.38054 × 10−23 × 4.65)
T°C = 2.49 × 120 × ∆VBE − 273
∆VBE is calculated from the difference in readings from the first
conversion to the second conversion. The user must perform
the calculations to get ∆VBE, and then calculate the temperature
value in degrees. Figure 33 shows a block diagram of the
temperature measurement circuit.
03796-011
TEMP1 TEMP2
MUX ADC
I
V
BE
I20 × I
Figure 33. Block Diagram of Temperature Measurement Circuit