Ground Plane interference

[player80]

Hi,

I'm doing some Audio application, now I discovered with a cheap DVD Player when I connect the Line-Out to the Line-In Jack of my application the audio chip might reset.
The problem is that the chip requires 3.3v on an EN pin if it drops below it might trigger a reset.
Now I checked the voltage on the shielding of the DVD Player, the scope says around 130 Volt (using the ground from the target application). As far as I read up on the internet this is a common case and very likely caused due induction on the AC adapter of the DVD player (I asked my friends to check their consumer electronic players they also confirmed that the voltage is usually >100V on the shielding, the brand does not matter Sony, Panasonic, etc. they all seem to have a similar behaviour).
If everything is already connected before booting up before initializing the audio chip everything is fine.

Does anyone have an idea what can be done here?

An audio transformer is a possible solution, but I wonder if there's any way to protect my circuit from such inrush shock shock situation?
It doesn't really matter where I touch the shieldings of the 2 devices, it happens easier if I do it at the connector, but also happens if I just touch the grounding of the dvd player on the grounding of the AC adapter of my application.

According to the scope, the Pk-Pk with the audio transformer is 106V (RMS 27V), without audio transformer 280V (RMS 91V)
The current must be small otherwise I wouldn't be able to write this!

 

[player80]

This is what the scope shows up on the 3.3v supply when grounding of my device touches the grounding of the line in shielding

 

[player80]

According to the documentation the reset pin can be left floating, however if I leave it floating ESD impact becomes worse.
What I did so far is I added a 0.1uF capacitor to the reset pin and pulled it high (3.3v), adding a capacitor only without pulling it high improved the situation but did not solve it completely. I'm still testing.
If someone has another idea about that I'd be glad to hear about it.

 

[FvM]

Apart from the special problem of connecting equipment suplied by an unearthed power supply, your circuit can be expected to show similar effects with regular static discharge. So it's more a problem of unsuitable design, e.g. insufficient ground, missing bypass capacitors, missing overvoltage protection or filtering of external cable connection.

A real product would most likely fail respective EMC tests and has to be redesigned.

 

[player80]

Apart from the special problem of connecting equipment suplied by an unearthed power supply, your circuit can be expected to show similar effects with regular static discharge. So it's more a problem of unsuitable design, e.g. insufficient ground, missing bypass capacitors, missing overvoltage protection or filtering of external cable connection.

A real product would most likely fail respective EMC tests and has to be redesigned.

I think there should be more than enough bypass capacitors, however I did not protect the i2c and reset lines. 0.1uF cap to GND and 4.7r resistor to 3v3 what do you think about that (that's what I added to the reset pins so far).
Do you know some good ways for testing compliance in a home environment? The DVD player test seems to be good but is limited of course.

I think I should get along with a few modifications, the mentioned 4.7r and 0.1uF cap already help a lot.
After testing for 1 hour it seems like I cannot reproduce any hardware crash anymore. I'll try an EMC laboratory within the next days to get some more information.

 

[FvM]

I don't know the design details, so I'm just guessing. It might be that the problem is brought up by additional connections in the development setup (debug and programming cables, probes, etc.)

Consider a board with a continuous groundplane where all external cables (power supply, signals) are connected in a way that they can't inject transient voltages into the circuit Then plugging and unplugging a cable can't disturb the system. It can only cause common mode noise but can't couple to the inner circuit.

Without a groundplane or at least a tightly meshed ground net, noise current injected into the ground at one point will be converted to a differential mode interference and reach all parts of the circuit.