ESD Protection for External Connector

[Luchik]

I would like to read opinions about using a ferrite bead in my design.
My intention is to provide ESD protection, adjoined to an external connector, in order to protect the circuit and the micro-controller.
I use USB connector but I do not use it for USB communication. I use it for both flash interface (Microchip's ICSP) and for UART at 9600 bps.
Specifications:
- Maximal signal frequency (ICSP) is about 10MHz.
- Operation voltage: 3-5V.
- This device must comply with EU standard for ESD protection: EN 301 480-3/1.



Questions:
1. I already use an ESD protection component (AZ23C3V3-7-F). Will the ferrite bead be an unnecessary redundancy?
2. If needed, which ferrite bead should i use?

[plazma]

I would connect the protection diodes before the ferrites. I have used both TVS ESD diodes and ferrite beads on some designs with difficult signals.

[janoc]

I use USB connector but I do not use it for USB communication. I use it for both flash interface (Microchip's ICSP) and for UART at 9600 bps.

Please, don't. You can be sure that someone will stick a regular USB device in it sooner or later and cause a failure - either of your product or the USB gizmo. And you will get (rightfully) blamed. Reusing an ubiquitous connector for a completely unrelated function is a terrible idea, IMO.

There are plenty of cheap 4 pin connectors on the market, just pick one that cannot be so easily confused with something else.

[Kilrah]

Yep... being concerned about a certification standard while being perfectly OK with doing a terribly bad thing is kinda ironic.

[Luchik]

I use USB connector but I do not use it for USB communication. I use it for both flash interface (Microchip's ICSP) and for UART at 9600 bps.

Please, don't. You can be sure that someone will stick a regular USB device in it sooner or later and cause a failure - either of your product or the USB gizmo. And you will get (rightfully) blamed. Reusing an ubiquitous connector for a completely unrelated function is a terrible idea, IMO.

There are plenty of cheap 4 pin connectors on the market, just pick one that cannot be so easily confused with something else.

I did not want to elaborate too much about it because it doesn't concern my main question, but the circuit IS intended for someone to connect a USB charger (or a "master" device that can deliver 5V) into it. This is the idea behind using USB connection and this method is very common in consumer product design. A mini or micro USB is usually used for charging.
So over-voltage or EMI are not a concern and they cannot harm the circuit, BUT I do want to have ESD protection on the two data lines as in the original question.

Yep... being concerned about a certification standard while being perfectly OK with doing a terribly bad thing is kinda ironic.

Please explain what is that bad thing that you see in the scheme. I would greatly appreciate your sincere response over a sarcastic one.

[janoc]

I did not want to elaborate too much about it because it doesn't concern my main question, but the circuit IS intended for someone to connect a USB charger (or a "master" device that can deliver 5V) into it. This is the idea behind using USB connection and this method is very common in consumer product design. A mini or micro USB is usually used for charging.
So over-voltage or EMI are not a concern and they cannot harm the circuit, BUT I do want to have ESD protection on the two data lines as in the original question.

I wouldn't be worried too much about overvoltage, but you said that you are using it for ICSP and serial coms too. What would happen if a device with shorted data pins (e.g. a generic charger) got plugged into it? Would a pin driver on your micro blow?

Or in reverse (not sure whether you are using A or B type connector) - what happens if someone plugs their USB stick or phone in it? Could it damage something on that $700 smartphone when it sees unexpected signals? Do you want to be the one to find out?

Using USB connectors for charging or delivering power is common, yes, but then the pins are not re-purposed for other things - like serial line or ICSP. That is the bad idea part of it. If you want to do that and still use USB connector, switch to USB-C where you have the alternate modes. However, properly implementing that would be a lot more complex. A better idea would be to keep the USB only for power (and USB if you need it) and have a separate connector for the rest.

[ConKbot]

Mind your pin states when charging, if someone uses a quick charge capable charger, and both pins go to 3.3v, your vbus will be greeted with 20v
https://blog.oxplot.com/quickcharge/

[Kilrah]

I did not want to elaborate too much about it because it doesn't concern my main question, but the circuit IS intended for someone to connect a USB charger (or a "master" device that can deliver 5V) into it.

Please explain what is that bad thing that you see in the scheme. I would greatly appreciate your sincere response over a sarcastic one.

No problem using USB for charging, but then don't use the data lines. You should not pass anything on them if it's not USB.

[David Hess]

I question whether a ferrite bead is going to do anything for ESD protection.

As far as using a USB connector for other functions, even if you protect your own devices, you are still training people to stick non-USB devices into USB devices and will your non-USB devices refrain from damaging someone else's USB device?

[plazma]

Ferrites do help. I have solved some ESD issues by adding a ferrite. There is still a TVS esd diode on the input side of the ferrite. The ferrite prevents the esd voltage to propagate towards the IC and rather discharge through the TVS diode.