MAX378/MAX379
High-Voltage, Fault-Protected
Analog Multiplexers
_______________________________________________________________________________________ 7
Input switching, however, must be done with a fault-
protected MAX378 multiplexer, to provide the level of
protection and isolation required with most data acqui-
sition inputs. Since external signal sources may contin-
ue to supply voltage when the multiplexer and system
power are turned off, non-fault-protected multiplexers,
or even first-generation fault-protected devices, will
allow many milliamps of fault current to flow from out-
side sources into the multiplexer. This could result in
damage to either the sensors or the multiplexer. A non-
fault-protected multiplexer will also allow input overvolt-
ages to appear at its output, perhaps damaging
Sample/Holds or A/Ds. Such input overdrives may also
cause input-to-input shorts, allowing the high current
output of one sensor to possibly damage another.
The MAX378 eliminates all of the above problems. It
not only limits its output voltage to safe levels, with or
without power applied (V+ and V-), but also turns all
channels off when power is removed. This allows it to
draw only sub-microamp fault currents from the inputs,
and maintain isolation between inputs for continuous
input levels up to ±75V with power supplies off.
_______________Detailed Description
Fault Protection Circuitry
The MAX378/MAX379 are fully fault protected for contin-
uous input voltages up to ±60V, whether or not the V+
and V- power supplies are present. These devices use
a “series FET” switching scheme which not only pro-
tects the multiplexer output from overvoltage, but also
limits the input current to sub-microamp levels.
Figures 7 and 8 show how the series FET circuit pro-
tects against overvoltage conditions. When power is
off, the gates of all three FETs are at ground. With a -60V
input, N-channel FET Q1 is turned on by the +60V gate-
G
D
Q
1
S
-60V
-60V
OVERVOLTAGE
N-CHANNEL MOSFET
IS TURNED ON
BECAUSE V
GS
= +60V
P-CHANNEL
MOSFET IS OFF
G
D
Q
2
S
G
D
Q
3
S
Figure 7. -60V Overvoltage with Multiplexer Power OFF
Q
1
V
TN = +1.5V
-15V +15V -15V
+13.5V
+60V
OVERVOLTAGE
N-CHANNEL MOSFET
IS TURNED ON
BECAUSE V
GS
= -45V
Q
2
Q
3
N-CHANNEL
MOSFET IS ON
+13.5V
OUTPUT
+15V FROM
DRIVERS
-15V FROM
DRIVERS
Figure 10. +60V Overvoltage Input to the ON Channel
Q
1
-15V +15V -15V
-60V
-60V
OVERVOLTAGE
N-CHANNEL MOSFET
IS TURNED OFF
BECAUSE V
GS
= +45V
Q
2
Q
3
P-CHANNEL
MOSFET IS OFF
N-CHANNEL
MOSFET IS OFF
+60V FORCED
ON COMMON
OUTPUT
LINE BY
EXTERNAL
CIRCUITRY
-15V FROM
DRIVERS
+15V FROM
DRIVERS
Figure 9. -60V Overvoltage on an OFF Channel with
Multiplexer Power Supply ON
G
D
Q
1
S
+60V
OVERVOLTAGE
N-CHANNEL MOSFET
IS TURNED OFF
BECAUSE V
GS
= -60V
G
D
Q
2
S
G
D
Q
3
S
Figure 8. +60V Overvoltage with Multiplexer Power OFF
MAX378/MAX379
High-Voltage, Fault-Protected
Analog Multiplexers
8 _______________________________________________________________________________________
to-source voltage. The P-channel device (Q2), howev-
er, has +60V V
GS
and is turned off, thereby preventing
the input signal from reaching the output. If the input
voltage is +60V, Q1 has a negative V
GS
, which turns it
off. Similarly, only sub-microamp leakage currents can
flow from the output back to the input, since any volt-
age will turn off either Q1 or Q2.
Figure 9 shows the condition of an OFF channel with
V+ and V- present. As with Figures 7 and 8, either an
N-channel or a P-channel device will be off for any
input voltage from -60V to +60V. The leakage current
with negative overvoltages will immediately drop to a
few nanoamps at +25°C. For positive overvoltages,
that fault current will initially be 40µA or 50µA, decaying
over a few seconds to the nanoamp level. The time
constant of this decay is caused by the discharge of
stored charge from internal nodes, and does not com-
promise the fault-protection scheme.
Figure 10 shows the condition of the ON channel with
V+ and V- present. With input voltages less than ±10V,
all three FETs are on and the input signal appears at the
output. If the input voltage exceeds V+ minus the N-
channel threshold voltage (V
TN
), then the N-channel
FET will turn off. For voltages more negative than V-
minus the P-channel threshold (V
TP
), the P-channel
device will turn off. Since V
TN
is typically 1.5V and V
TP
is typically 3V, the multiplexer’s output swing is limited
to about -12V to +13.5V with ±15V supplies.
The
Typical Operating Characteristics
graphs show typi-
cal leakage vs. input voltage curves. Although the max-
imum rated input of these devices is ±65V, the
MAX378/MAX379 typically have excellent performance
up to ±75V, providing additional margin for the unknown
transients that exist in the real world. In summary, the
MAX378/MAX379 provide superior protection from all
fault conditions while using a standard, readily pro-
duced junction-isolated CMOS process.
Switching Characteristics
and Charge Injection
Table 1 shows typical charge-injection levels vs.
power-supply voltages and analog input voltage. Note
that since the channels are well matched, the differen-
tial charge injection for the MAX379 is typically less
than 5pC. The charge injection that occurs during
switching creates a voltage transient whose magnitude
is inversely proportional to the capacitance on the mul-
tiplexer output.
The channel-to-channel switching time is typically 600ns,
with about 200ns of break-before-make delay. This 200ns
break-before-make delay prevents the input-to-input short
that would occur if two input channels were simultaneous-
ly connected to the output. In a typical data acquisition
system, such as in Figure 6, the dominant delay is not the
switching time of the MAX378 multiplexer, but is the set-
tling time of the following amplifiers and S/H. Another limit-
ing factor is the RC time constant of the multiplexer
R
DS(ON)
plus the signal source impedance multiplied by
the load capacitance on the output of the multiplexer.
Even with low signal source impedances, 100pF of capac-
itance on the multiplexer output will approximately double
the settling time to 0.01% accuracy.
Operation with Supply Voltage
Other than ±15V
The main effect of supply voltages other than ±15V is
the reduction in output signal range. The MAX378 limits
the output voltage to about 1.5V below V+ and about 3V
above V-. In other words, the output swing is limited to
+3.5V to -2V when operating from ±5V. The
Typical
Operating Characteristics
graphs show typical R
DS(ON)
,
for ±15V, ±10V, and ±5V power supplies. Maxim tests
and guarantees the MAX378/MAX379 for operation from
±4.5V to ±18V supplies. The switching delays are
increased by about a factor of 2 at ±5V, but break-
before-make action is preserved.
The MAX378/MAX379 can be operated with a single +9V
to +22V supply, as well as asymmetrical power supplies
such as +15V and -5V. The digital threshold will remain
approximately 1.6V above GND and the analog character-
istics such as R
DS(ON)
are determined by the total voltage
difference between V+ and V-. Connect V- to 0V when
operating with a +9V to +22V single supply.
This means that the MAX378/MAX379 will operate with
standard TTL-logic levels, even with ±5V power sup-
plies. In all cases, the threshold of the EN pin is the
same as the other logic inputs.
Table 1a. MAX378 Charge Injection
Test Conditions: C
L
= 1000pF on multiplexer output; the tabu-
lated analog input level is applied to channel 1; channels 2
through 8 are open circuited. EN = +5V, A1 = A2 = 0V, A0 is
toggled at 2kHz rate between 0V and 3V. +100pC of charge
creates a +100mV step when injected into a 1000pF load
capacitance.
Supply Voltage Analog Input Level Injected Charge
±5V
+1.7V
0V
-1.7V
+100pC
+70pC
+45pC
±10V
+5V
0V
-5V
+200pC
+130pC
+60pC
±15V
+10V
0V
-10V
+500pC
+180pC
+50pC
MAX378/MAX379
High-Voltage, Fault-Protected
Analog Multiplexers
_______________________________________________________________________________________ 9
Digital Interface Levels
The typical digital threshold of both the address lines
and the EN pin is 1.6V, with a temperature coefficient of
about -3mV/°C. This ensures compatibility with 0.8V to
2.4V TTL-logic swings over the entire temperature
range. The digital threshold is relatively independent of
the supply voltages, moving from 1.6V typical to 1.5V
typical as the power supplies are reduced from ±15V to
±5V. In all cases, the digital threshold is referenced to
GND.
The digital inputs can also be driven with CMOS-logic
levels swinging from either V+ to V- or from V+ to GND.
The digital input current is just a few nanoamps of leak-
age at all input voltage levels, with a guaranteed maxi-
mum of 1µA. The digital inputs are protected from ESD
by a 30V zener diode between the input and V+, and
can be driven ±4V beyond the supplies without drawing
excessive current.
Operation as a Demultiplexer
The MAX378/MAX379 will function as a demultiplexer,
where the input is applied to the OUT pin, and the input
pins are used as outputs. The MAX378/MAX379 pro-
vide both break-before-make action and full fault protec-
tion when operated as a demultiplexer, unlike earlier
generations of fault-protected multiplexers.
Channel-to-Channel Crosstalk,
Off Isolation, and Digital Feedthrough
At DC and low frequencies, channel-to-channel
crosstalk is caused by variations in output leakage cur-
rents as the off-channel input voltages are varied. The
MAX378 output leakage varies only a few picoamps as
all seven off inputs are toggled from -10V to +10V. The
output voltage change depends on the impedance level
at the MAX378 output, which is R
DS(ON)
plus the input
signal source resistance in most cases, since the load
driven by the MAX378 is usually a high impedance. For
a signal source impedance of 10kor lower, the DC
crosstalk exceeds 120dB.
Table 2 shows typical AC crosstalk and off-isolation per-
formance. Digital feedthrough is masked by the analog
charge injection when the output is enabled. When the
output is disabled, the digital feedthrough is virtually
unmeasurable, since the digital pins are physically iso-
lated from the analog section by the GND and V- pins.
The ground plane formed by these lines is continued
onto the MAX378/MAX379 die to provide over 100dB
isolation between the digital and analog sections.
Table 1b. MAX379 Charge Injection
+1.7V
0V
-1.7V
+105pC
+73pC
+48pC
±10V
+5V
0V
-5V
+215pC
+135pC
+62pC
±15V
+10V
0V
-10V
+525pC
+180pC
+55pC
±5V
Test Conditions: C
L
= 1000pF on Out A and Out B; the tabulat-
ed analog input level is applied to inputs 1A and 1B; channels
2 through 4 are open circuited. EN = +5V, A1 = 0V, A0 is tog-
gled from 0V to 3V at a 2kHz rate.
+107pC
+74pC
+50pC
+220pC
+139pC
+63pC
+530pC
+185pC
+55pC
Out A Out B
Injected Charge
-2pC
-1pC
-2pC
-5pC
-4pC
-1pC
-5pC
-5pC
0pC
Differential
A-B
Supply
Voltage
Analog
Input Level
Table 2a. Typical Off-Isolation
Rejection Ratio
Test Conditions: V
IN
= 20V
P-P
at the tabulated frequency,
R
L
= 1.5kbetween OUT and GND, EN = 0V.
20V
P-P
OIRR = 20 Log ____________
V
OUT (P-P)
Frequency 100kHz 500kHz 1MHz
One Channel Driven
74dB 72dB 66dB
All Channels Driven
64dB 48dB 44dB
Table 2b. Typical Crosstalk
Rejection Ratio
Test Conditions: Specified R
L
connected from OUT to GND,
EN = +5V, A0 = A1 = A2 = +5V (Channel 1 selected). 20V
P-P
at the tabulated frequency is applied to Channel 2. All other
channels are open circuited. Similar crosstalk rejection can be
observed between any two channels.
Frequency
100kHz 500kHz 1MHz
F
L
= 1.5k
70dB 68dB 64dB
R
L
= 10k
62dB 46dB 42dB

MAX378CPE+

Mfr. #:
Manufacturer:
Maxim Integrated
Description:
Multiplexer Switch ICs 8:1 Fault-Protected Analog MUX
Lifecycle:
New from this manufacturer.
Delivery:
DHL FedEx Ups TNT EMS
Payment:
T/T Paypal Visa MoneyGram Western Union