Hi everyone, I started this post to share about how to build bullet proof circuits capable of supporting misconnections, overvoltages and such other stresses that customers can do…
My last one: I protect my circuit against polarity inversion using the usual diode +PTC fuse, I protect each analog IO against overvoltage using TVS clamp and PTC fuse, I protect COM port using opto transistor… but this is not enough, a customer short circuits a 10V voltage source with the metallic enclosure of our device. This is not normally a problem since the enclosure is not electrically connected to the board, but I forgot that this customer has a remote control option with an antenna fixed on the enclosure (to enhance range due to ground plane effect)… and patatra the board has smoked (a net wire that is supposed to serves as ground signal for voltage reference), the G400D is not responsive anymore, and the XBEE chip is bricked…
How you manage such case (adding more protections), do not let customers open the enclosure to diagnose a problem?
Isolating the antenna will decrease the ground plane effect in our next device we move to BLE technologie it does not need external antenna so this will be solved for the antenna.
How do you protect your board against voltage applied on ground?
You can only do so much. They will still find a way to destroy it. My design was protected by a PTC fuse and a reverse protection diode and designed for +24dc so when the client doesn’t read the instructions and inputs 120Vac all the protection didn’t work.
of course when you say “some”, the answer will be yes.
if your device can be left sitting on a table where the user could spill water on it, then a conformal coating may help. From what little I know it’s more used in an environmental protection scenario, simple moisture protection or vibration protection (depending on whether it’s a spray coat or it’s fully potted changes how much protection it’s given). That might help in dusty or vibration prone situations (a CNC for example ). But you still have to expose the + and - wire, and there’s still a chance someone uses the wrong plug pack or connects it to AC, so you can’t protect against the common issues that way.
When I worked in the ROV industry, all of the subsea boards where conformally coated and rework was a nightmare. In saying that, I was working on some of them a few weeks back and they are still working even when the metal chassis itself was showing signs of mild corrosion from moisture in the pods.
I tough to use MOSFET on minus power supply but this would involve a voltage drop as PTC and would complicated analog voltage measurement. According to my boss our product is not as reliable as what we can found in end user products or in industry… Saying that he asked me to reverse engineering state of the art devices in marine equipment and I didn’t find nothing more that what we use already for protection unless having a pcb muc bigger than what we have now, and with an increase of the parts bill. Am I wrong?
As an exemple I consider that anaog input is referenced to the main minus power supply; ideally we should not make this assumption but this woould involved separating each ground (8 analog inputs) plus adding operationnal amplifier on each (how to support negative voltage on analog…) this is a nightmare to manage just because they want to plug before reading specs… :wall:
). But you still have to expose the + and - wire, and there’s still a chance someone uses the wrong plug pack or connects it to AC, so you can’t protect against the common issues that way.