I was very surprised when the battery voltage dropped below 2.7V (exactly
2.62V) and ADuM1401 begun to generate
sporadically 800ns long pulses at the output VOD (pin no. 6).
The other side is supplied with 5V and works correctly.
After an IC change, study of the data sheet and finally a battery replacement
the problem is solved.
I would like to ask you a question.
Is this behaviour normal for all circuits from the ADuM family?
Has anybody mentioned this behaviour to you?
The ADuM series of parts should not miss-operate at 2.62V. From my work with
the parts, the actual threshold of operation at room temperature is right
around 2V. At that low voltage all sorts of odd behavior can happen. Near the
threshold encoded pulses are near the detection threshold and can be missed so
random things can start happening. I would check the integrity of the power
right at the Vdd pins. Be sure that there is a good bypass and maybe a
stiffening cap on the Vdd pins.
The refresh period is close to 1uS, probably shorter at low voltage. The
signals you are seeing is likely refresh pulses that are being partially missed
by the output decoder due to low levels, it sets the wrong state and then is
corrected at the next refresh. Again, this should not be happening until Vdd
is close to 2.0V. Are the 880nS pulses highs or lows? It could also be
related to the default high output state.
You can look at the ADuM3401, it has an under voltage lockout feature that
should preclude strange behavior if the power supply is weak. It is pin
compatible and has only a small price difference.