MAX1153 Datasheet, PDF (21/29 Pages) Maxim Integrated Products – Stand-Alone, 10-Channel, 10-Bit System Monitors with Internal Temperature Sensor and VDD Monitor

Stand-Alone, 10-Channel, 10-Bit System Monitors
with Internal Temperature Sensor and V_DD Monitor
Table 13. Channel Configuration Register Format
B7 (MSB)
Fault B3
B6
Fault B2
B5
Fault B1
B4
Fault B0
B3
Ave B3
B2
Ave B2
B1
Ave B1
B0 (LSB)
Ave B0
Table 14. Conversion Average Encoding
CODE
0000
0001
0010
0011
0100
0101
0110
0111
1000
1001
1010
1011
1100
1101
1110
1111
N
1, no averaging
2
4
8
16
32
64
128
256
512
1024
2048
Reserved
Reserved
Reserved
Reserved
Channel Configuration Register
Each channel has a channel configuration register (Table
13) defining the number of consecutive faults to be
detected before setting the alarm bits and generating an
interrupt, as well as controlling the digital averaging func-
tion. At power-up and after a RESET command, the regis-
ter defaults to 00 hex (no averaging, alarm on first fault).
Fault Bits
The value stored in the fault bits (B7–B4) in the channel
configuration register sets the number of faults that
must occur for that channel before generating an inter-
rupt. Encoding of the fault bits is straight binary with
values 0 to 15. A fault occurs in a channel when the
value in its current data register is outside the range
defined by the channel’s upper and lower threshold
registers. For example, if the number of faults set by the
fault bits is N, an interrupt is generated when the num-
ber of consecutive faults (see following note) reach
(N + 1). The fault bits default to 0 hex at power-up.
Note: Consecutive faults are those happening in con-
secutive conversion scans for the same channel. If a
fault occurs and the next scan finds the input within the
normal range defined by the thresholds, the fault
counter resets. If the next counter finds the input signal
outside the opposite threshold, rather than the previous
one, the fault counter also resets. The fault counter
increments only when counting consecutive faults
exceeding the same threshold (Figure 4).
Averaging
The averaging calculated by the data-acquisition algo-
rithm of the MAX1153/MAX1154 improves the input sig-
nal-to-noise ratio (SNR) by reducing the signal
bandwidth digitally. The formula below describes the
filter implemented in the MAX1153/MAX1154:
current value = [(N - 1) / N] x past value +
[(present value) / N]
where N = number of samples indicated in Table 14.
The averaging bits (B3–B0) in the channel configuration
register can set the N factor to any value in Table 14.
The output of the filter-running algorithm is continuously
available in the current data register. The starting value
used by the algorithm is the initial state of the current
data register. The current data register is reset to mid-
scale (800 hex) at power-up or after a RESET com-
mand, but it can be loaded with a more appropriate
initial value to improve the filter settling time.
At power-up or after a RESET command, the B3–B0
bits of the channel configuration register are set to 0
hex, corresponding to a number of averaged N = 1, no
averaging. See Table 13 and the Write-Selected
Channel Configuration Registers section for program-
ming details. See Table 14 for N encoding.
As in all digital filters, truncation can be a cause of sig-
nificant errors. In the MAX1153/MAX1154, 24 bits of
precision are maintained in the digital averaging func-
tions, maintaining a worst-case truncation error of well
below an LSB. The worst-case truncation error in the
MAX1153/MAX1154 is given by the following:
worst- case truncation error = ^N-1 LSBs
16384
where N = number of conversions averaged.
Therefore, the worst truncation error when averaging
256 samples is 0.01557 LSBs.
______________________________________________________________________________________ 21