SE95_7
?NXP B.V. 2009. All rights reserved.
Product data sheet
Rev. 07 2 September 2009
6 of 27
NXP Semiconductors
SE95
Ultra high accuracy digital temperature sensor and thermal watchdog
7.2 OS output and polarity
The OS output is an open-drain output and its state represents results of the device
watchdog operation as described in
Section
7.1. In order to observe this output state, an
external pull-up resistor is needed. The resistor should be as large as possible, up to
200 k&, to minimize the Temp reading error due to internal heating by the high OS sinking
current.
The OS output active state can be selected as HIGH or LOW by programming bit
OS_POL of register Conf: setting bit OS_POL to logic 1 selects OS active HIGH and
setting to logic 0 sets OS active LOW. At power-up, bit OS_POL is equal to logic 0 and the
OS active state is LOW.
7.3 OS comparator and interrupt modes
As described in
Section
7.1, the OS output responds to the result of the comparison
between register Temp data and the programmed limits, in registers Tos and Thyst, in
different ways depending on the selected OS mode: OS comparator or OS interrupt. The
OS mode is selected by programming bit OS_COMP_INT of register Conf: setting bit
OS_COMP_INT to logic 1 selects the OS interrupt mode, and setting to logic 0 selects the
OS comparator mode. At power-up, bit OS_COMP_INT is equal to logic 0 and the OS
comparator is selected.
The main difference between the two modes is that in OS comparator mode, the OS
output becomes active when Temp has exceeded T
os
and reset when Temp has dropped
below T
hyst
, reading a register or putting the device into shutdown mode does not change
the state of the OS output; while in OS interrupt mode, once it has been activated either
by exceeding T
os
or dropping below T
hyst
, the OS output will remain active indenitely until
reading a register or putting the device into shutdown mode occurs, then the OS output is
reset.
Temperature limits T
os
and T
hyst
must be selected so that T
os
> T
hyst
. Otherwise, the OS
output state will be undened.
7.4 OS fault queue
Fault queue is dened as the number of faults that must occur consecutively to activate
the OS output. It is provided to avoid false tripping due to noise. Because faults are
determined at the end of data conversions, fault queue is also dened as the number of
consecutive conversions returning a temperature trip. The value of fault queue is
selectable by programming the two bits OS_F_QUE[1:0] in register Conf. Notice that the
programmed data and the fault queue value are not the same.
T
ab
le
3 shows the
one-to-one relationship between them. At power-up, fault queue data = 00 and fault queue
value = 1.
Table 3.
Fault queue table
Fault queue data
Fault queue value
OS_F_QUE[1]
OS_F_QUE[0]
Decimal
0
0
1
0
1
2
1
0
4
1
1
6