Connectors shielding, plastic housing ESD test
Wlada_85 , 10-02-2023, 04:29 AM
Hi,
I have a PCB project where my PCB have to be interface between 3 systems. And because cables from battery are long and full of noise, we decided to separate grounds on PCB and to take ground from S1 and ground from S2 as reference for analog voltage (isolated DAC).
For connection between systems and MY PCB we use standard M connectors with shielding.
Of course, more than one ground more problems.
Problem is with shielding and ESD (one board passed one failed ... )
PCB is in plastic housing.
In first scenario all connectors shielding are connected to corresponding ground. For example connector for system S1, shield is connected to the ground of S1. And so. But, when S1 is not connected, gnd floating. Same for S2. Please check figure:
In this first scenario when every connector has corresponding GND to shield, ESD caused failure of one PCB. And S2 wasn't attached (somehow one pin of MCU is short and that pin doesn't goes outside of PCB, we couldn't repeat this on second PCB)
Second scenario is that all shields of connectors are connected to minus line of battery.
Third scenario is that all shields are not connected.
I saw a video: https://www.youtube.com/watch?v=F9tB6JtG7oQ
But what when housing is plastic?
On the o side of systems S1 and S2, shield is not connected. Battery voltage is 24V
What would be the best way to connect shielding ? Or to put zero ohm bridge and to test
?
TY
I have a PCB project where my PCB have to be interface between 3 systems. And because cables from battery are long and full of noise, we decided to separate grounds on PCB and to take ground from S1 and ground from S2 as reference for analog voltage (isolated DAC).
For connection between systems and MY PCB we use standard M connectors with shielding.
Of course, more than one ground more problems.
Problem is with shielding and ESD (one board passed one failed ... )
PCB is in plastic housing.
In first scenario all connectors shielding are connected to corresponding ground. For example connector for system S1, shield is connected to the ground of S1. And so. But, when S1 is not connected, gnd floating. Same for S2. Please check figure:
In this first scenario when every connector has corresponding GND to shield, ESD caused failure of one PCB. And S2 wasn't attached (somehow one pin of MCU is short and that pin doesn't goes outside of PCB, we couldn't repeat this on second PCB)
Second scenario is that all shields of connectors are connected to minus line of battery.
Third scenario is that all shields are not connected.
I saw a video: https://www.youtube.com/watch?v=F9tB6JtG7oQ
But what when housing is plastic?
On the o side of systems S1 and S2, shield is not connected. Battery voltage is 24V
What would be the best way to connect shielding ? Or to put zero ohm bridge and to test
?TY
qdrives , 10-02-2023, 02:53 PM
In my opinion shielding is more complicated than even Rick Hartley states.
You pointed out correctly that your enclosure is plastic. So there is no Faraday cage. Is shielding in such cases useless -- no, it definitely is good when done correctly.
To start of: "long cables" -- but how long?
For analog signals I would suggest using 4...20mA.
If everything is powered from a single battery, there is no need to isolate Gnd. What you do need is filters.
I would personally go for a single Gnd and connected all the shields also to that Gnd.
How fast are those digital signals?
Do add AC termination to them.
How much current are the boards/systems drawing from the battery?
You pointed out correctly that your enclosure is plastic. So there is no Faraday cage. Is shielding in such cases useless -- no, it definitely is good when done correctly.
To start of: "long cables" -- but how long?
For analog signals I would suggest using 4...20mA.
If everything is powered from a single battery, there is no need to isolate Gnd. What you do need is filters.
I would personally go for a single Gnd and connected all the shields also to that Gnd.
How fast are those digital signals?
Do add AC termination to them.
How much current are the boards/systems drawing from the battery?
Wlada_85 , 10-03-2023, 02:30 AM
Yes it is.
I agree, that is the reason why I asked for help. Maybe someone has similar situation ...
Cables from battery to system S3 are 6 - 10 meters. System S3 can draw current up to 20A or more. and it is not constant current, there is a lot of spikes ...
Systems S2 and S3 are close to MyPCB.
I agree for 4-20mA but I can't change systems S1 and S2. They accept 0-3V3 only. And that voltage need to be stable. And with this isolation of grounds system works well, only issue is shielding :/
It is not possible, I wanted that also, but because of resistance and inductance in 6 meters of cable I can't have stable GND. Digital signal are 100 - 500 KHz.
System S1 is around 1A and S2 maybe 100mA ... And S1 and S2 are not close connected to the battery.
My thinking in the first was that every connector have to have shielding connected to corresponding GND. But when there is no system attached it can cause floating shield. My second thought is that every connectors shielding can be connected to battery minus. Or to be left floating. But not sure what would be pros and cons.
It would be nice to have some practical examples, or some experience...
What I also notice during ESD tests is that 4kV contact discharge can't shoot more than 2-3 times (by standard it need 10 times) but after 2-3 shots system is full with energy. There is no connection to earth and no discharge so we needed to wait.
TY for your reply
BR,
I agree, that is the reason why I asked for help. Maybe someone has similar situation ...
Cables from battery to system S3 are 6 - 10 meters. System S3 can draw current up to 20A or more. and it is not constant current, there is a lot of spikes ...
Systems S2 and S3 are close to MyPCB.
I agree for 4-20mA but I can't change systems S1 and S2. They accept 0-3V3 only. And that voltage need to be stable. And with this isolation of grounds system works well, only issue is shielding :/
It is not possible, I wanted that also, but because of resistance and inductance in 6 meters of cable I can't have stable GND. Digital signal are 100 - 500 KHz.
System S1 is around 1A and S2 maybe 100mA ... And S1 and S2 are not close connected to the battery.
My thinking in the first was that every connector have to have shielding connected to corresponding GND. But when there is no system attached it can cause floating shield. My second thought is that every connectors shielding can be connected to battery minus. Or to be left floating. But not sure what would be pros and cons.
It would be nice to have some practical examples, or some experience...
What I also notice during ESD tests is that 4kV contact discharge can't shoot more than 2-3 times (by standard it need 10 times) but after 2-3 shots system is full with energy. There is no connection to earth and no discharge so we needed to wait.
TY for your reply
BR,
qdrives , 10-03-2023, 02:51 PM
0.0175 [ohm * mm² / m] or 17.5e-9 [ohm * m] and U=I * R
If possible:
- Place battery close to S3 as it draws much current
- Battery supply and Gnd go from S3 to your (central) board and from there to S1 and S2 (a bit like a star topology)
You could use a differential pair for the digital signals. A simple RS422 driver could be used for this.
ESD testing is about damage (or at least not getting damaged). However, I have never done a battery powered system. I wouldn't be surprised if a drain wire may be attached to speed up the testing. But I am not a NoBo (Notified Body) nor did I read the standards (IEC 61000-4-2).
For ESD you want the Gnds to be hard connected. Shielding has low impedance. Alternatively, add high power braided ground straps (https://www.digikey.nl/en/products/f...aid-straps/494)
If possible:
- Place battery close to S3 as it draws much current
- Battery supply and Gnd go from S3 to your (central) board and from there to S1 and S2 (a bit like a star topology)
You could use a differential pair for the digital signals. A simple RS422 driver could be used for this.
ESD testing is about damage (or at least not getting damaged). However, I have never done a battery powered system. I wouldn't be surprised if a drain wire may be attached to speed up the testing. But I am not a NoBo (Notified Body) nor did I read the standards (IEC 61000-4-2).
For ESD you want the Gnds to be hard connected. Shielding has low impedance. Alternatively, add high power braided ground straps (https://www.digikey.nl/en/products/f...aid-straps/494)
Wlada_85 , 10-16-2023, 08:48 AM
TY for answer. Most probably PCB died because minus line of battery wasn't soldered properly and falloff later ... On repeated tests boards were fine.
Use our interactive Discord forum to reply or ask new questions.