are dual comparator really independant?

[Mark]

Is a dual/quad compartor in the same DIP totally indepentant of each other
or is there some interaction between the outputs?

 

[John Larkin]

Is a dual/quad compartor in the same DIP totally indepentant of each other
or is there some interaction between the outputs?

There may be some small crosstalk, in that outputs that cross close in
time may influence each other a bit.

Sometimes DC biasing of one stage can affect another. LM339 types can
interact two ways: if one channel has both inputs very high or open, a
shared current source can partially collapse and degrade the others;
and if you pull any input below ground, the entire chip can go nuts.

Try it!

John

 

[John Popelish]

Is a dual/quad compartor in the same DIP totally indepentant of each other
or is there some interaction between the outputs?

There are no two anything in the universe that are truly independent
of each other. The two amplifiers in a dual comparator share power
supply pins and usually some bias circuits, so there are several sneak
paths between them, besides the field paths that connect them since
they are so close ot each other. If you are really pushing the
isolation envelope with your circuit, single units may be helpful.

 

[Winfield Hill]

Mark wrote...

Is a dual/quad compartor in the same DIP totally indepentant
of each other or is there some interaction between the outputs?

NSC doesn't have a spec for their quad comparator IC, but they
do for ICs designed at the same time with similar technology.
The LM392 (dual-function opamp + comparator) has a spec called
"Amplifier-to-Amplifier Coupling," (f = 1 kHz to 20 kHz), which
is −100 dB typical (input referred). And their LM324 quad opamp
data sheet spec is −120dB typ. Both specs have this same note,
"Due to proximity of external components, insure that coupling
is not originating via stray capacitance between these external
parts. This typically can be detected as this type of capacitance
increases at higher frequencies."

Assuming a 12V output swing, a -100dB spec would imply a 120uV
input feedthrough effect. In addition to external-component
stray capacitance, one should watch out for power-supply-rail
feedthrough. The LM324 opamps have a 100dB typical power-supply
rejection ration spec. The opamp's input circuit is similar to
the LM339 comparator and we can expect it has a similar weakness.
So care should be taken that current switched by the comparator's
output doesn't cause too much pin 12 supply-rail noise.

Thanks,
- Win

whill_at_picovolt-dot-com

 

[Winfield Hill]

Winfield Hill wrote...

The LM392 (dual-function opamp + comparator) has a spec called
"Amplifier-to-Amplifier Coupling," (f = 1 kHz to 20 kHz), which
is −100 dB typical (input referred). And their LM324 quad
opamp data sheet spec is −120dB typ.

Sheesh, a 16-bit unicode - symbol? The specs were supposed to be
-100dB and -120dB.

Thanks,
- Win

whill_at_picovolt-dot-com

 

[John Larkin]

Assuming a 12V output swing, a -100dB spec would imply a 120uV
input feedthrough effect. In addition to external-component
stray capacitance, one should watch out for power-supply-rail
feedthrough. The LM324 opamps have a 100dB typical power-supply
rejection ration spec. The opamp's input circuit is similar to
the LM339 comparator and we can expect it has a similar weakness.
So care should be taken that current switched by the comparator's
output doesn't cause too much pin 12 supply-rail noise.

If you drive one LM324 section open-loop rail-to-rail, you'll see
*big* spikes on the other opamp section outputs. Cheap, and worth it.

John

 

[JeffM]

Is a dual/quad compartor in the same DIP totally indepentant of each other

or is there some interaction between the outputs?
Mark

Look at the Power Supply Rejection Ratio (PSRR) on your datasheet.
There's your interaction.