[sdiy] NJM2068 gone bad

[Frédéric (Opensource)]

Hi Tom, hi Jay,

first, thank you for your answers.

For the ESD protection I meant TVS diodes, not regular
bipolar diodes so the protection is for both pulse polarity.

For current injection, there is a 20k input impedance that
limits the risk in case of overvoltage below the threshold
of the TVS (which is about 16V, this parameter has a big
tolerance).

 > was the system and/or part known good when it shipped
 > out?

Yes, it seems to have correctly passed the factory test bench.
The system was used by a tester who did not noticed any weird
behavior and then shipped over to an other tester who noticed
the distortion immediately.

Today, I made a series of tests going above all ratings on power
rails input levels and everything I could think of and could not
reproduce the issue, damaging the circuit. Also the device has
successfully passed all ESD tests when we went for the CE / FCC
compliance.

So maybe a problem of soldering or a defective chip.

Thanks again for your time and answers,
Frédéric


Le 25/06/2020 à 00:23, Tom Corbitt a écrit :
> Just to clarify, was the system and/or part known good when it shipped 
> out?
>
> Assuming it was, you mention the protection ESD diode and reverse 
> protection but I wonder does the circuit have any other protection for 
> voltages below ESD levels but over the safe maximum for the chip?
>
>  I see +/- 18V max listed on the datasheet I looked at.
>
> That's just my first instinct, that someone plugged in something with 
> too great of a potential and it damaged one of the two channels.
>
> It's easy enough to test and there are very simple fixes if it is an 
> issue (36V zener with a current limiting resistor in parallel with 
> input is a very basic old school way to do it, plenty of smarter 
> folks here on the list with better ideas and ways I'm sure)
>
> Tom
>
>
> On Wed, Jun 24, 2020 at 1:23 PM Frédéric (Opensource) 
> <marzacdev at gmail.com <mailto:marzacdev at gmail.com>> wrote:
>
>     Hi to everyone,
>
>     In a client audio design, I used a NJM2068 dual op-amp
>     as the input signal conditioner before the ADC circuit.
>
>     The op-amp is powered with symmetrical supply rails (+12V, -12V)
>     provided by the power bus of an eurorack system.
>     Rails are protected by series shottky diodes therefore the
>     case of a polarity reversal can be discarded I presume.
>
>     Audio input has an ESD protection diode and the impedance
>     between the + input of the amp and the input jack tip is about
>     20k (resistor). The input is specified for "eurorack level" signals.
>
>     Unfortunately, after the field tests, we got back one module
>     whose input op-amp had a defective channel. Input signal would
>     be saturated while being far away from the rails.
>
>     Had any of you a bad experience with this component?
>
>     I am also open for all suggestions about what wrong could
>     have happened?
>
>     Best from Bonn,
>     Frédéric
>
>     _______________________________________________
>     Synth-diy mailing list
>     Synth-diy at synth-diy.org <mailto:Synth-diy at synth-diy.org>
>     http://synth-diy.org/mailman/listinfo/synth-diy
>