I’m having problems passing ESD immunity on a device I’m designing. It’s a portable USB device in an aluminum enclosure with a metal USB connector that is in contact with the enclosure. When the device is hit by the ESD gun, the discharge ultimately finds its way onto the USB connector through the shield and into the ground plane where it causes havoc. I’ve accepted that there’s no way to stop the discharge from getting onto the USB connector shield, and so I want to try to attenuate and disperse it there. Can anyone recommend a filter circuit for the USB shield pins?
\$\begingroup\$
\$\endgroup\$
1
-
1\$\begingroup\$ It would help to see how the schematics how the USB interface is currently designed, to assess what might be the cause and how to fix it. \$\endgroup\$– JustmeCommented Nov 3, 2020 at 8:31
Add a comment
|
1 Answer
\$\begingroup\$
\$\endgroup\$
6
Usually the ESD test will find it's way to the USB shield and the signal lines are usually also protected.
Shield should be connected to GND through a resistor (1M) and a parallel capacitor (100n + 1n). You can add other connector shield or the case to device GND using the same connection point. Some people will disagree, but this solution has been tested many times functionally and in EMC/ESD tests and it's also recommended by USB chip designers.
You should also have TVS diode (datasheets will state if the diode is good for your USB protocol) on the signal and V_BUS -lines and series resistors, maybe even a common mode choke on the signal lines.
-
\$\begingroup\$ I do have a USB TVS diode on the board protecting 5V and the data lines. Would a ferrite on the shield help in there as well? Thanks, BTW \$\endgroup\$ Commented Nov 3, 2020 at 9:17
-
\$\begingroup\$ Good, I also like to put the series resistors in, in case the USB chip internal protection kicks in first and it also cuts radiated emissions. Ferrite won't help with ESD, but it will help with radiated emissions and immunity. Have a pack of clampable ferrites, ferrite sheet and copper tape with you at the test lab :) \$\endgroup\$– RalphCommented Nov 3, 2020 at 9:50
-
\$\begingroup\$ My understanding was that a ferrite in series with the shield connection would turn some of the high frequency energy into heat. Is that not the case? I basically want to turn that USB shield connection into a sort of ESD lightening rod. \$\endgroup\$ Commented Nov 4, 2020 at 8:12
-
\$\begingroup\$ That's also correct, but if you want to pass all test cases (and they try to simulate the true use cases), then you have multiple problems to solve and my example has been tested in practice multiple times and is therefore recommended by USB chip designers and other professionals. The very optimized solution depends on your application, but you would only want to optimize it if you run into problems in tests or real application. \$\endgroup\$– RalphCommented Nov 5, 2020 at 10:00
-
\$\begingroup\$ Thank-you again for your continued feedback. I put together a couple different versions and test them both. \$\endgroup\$ Commented Nov 5, 2020 at 12:35