I asked about this. And this seems like a good answer:
Also the bit about lightening strikes.
Since the JCI stuff is made to be isolated, it can survive without a ground, unlike Alerton and others.grounding is a MUST on those, but not a must universally.
Add to that various unusual but possibly significant issues that can arise. I really like the isolated setup. Not only does it seem more robust and reliable, it can easily confuse the less savvy so I can get a foot into the building. I will still ground every Alerton, Distech, and ALC system I meet. But it is good to have an eye out for other systems that do not need it.
Very interesting conversation.
Good day numbawunfela,
As mentioned, there is no "right" or "wrong" way of powering gear...there are pros and cons to both approaches in terms of having a power supply input grounded or isolated. In my case, I have never had a specific need and felt that the benefits car out weighed the negatives of having an isolated power input. I have being using this approach for over 30 years in all of my designs and I still have some of these old devices in the field... and are still working. One in particular is part of an commercial tower's freight elevator and has been running 24/7 since 1996 with zero failures... not even a reboot has been required. So, any reliability concerns are out the window. That said, there are systems out there with a ground referenced power supply and these can work a long time too. It all depends upon the application, etc.
Relying on a common ground can be very problematic, as people assume that the ground potential is the same everywhere... this is not the case and can vary under a variety of conditions where a Lightning strike, as I mentioned, is the most common. However, even local changes can occur from a static discharge. It is much harder to control the current/energy flow from a static zap in a ground referenced device. If one does not "control" or direct the static zap's current, then the device (and/or downstream connected devices) can become damaged or cause them to reboot/restart. Isolated powered devices are not immune from this issue, but it is a bit easier to control the static zap energy to a location that has a dedicated or isolated ground point... now the static discharge bypasses the internal components (hopefully) and so minimizes damage or a triggered reboot ... this is all part of EMC compatibility and/or approvals that may be required for the devices.
The other issue that can occur is differential ground currents. Remember that if there is a voltage differential between two points, a current will flow. So... if the "ground" at one place is different than the "ground" at another, then you will get a differential current and these currents can be significant. This is why the RS485 interfaces typically have a 100 ohm or so resistor between the "ref" terminal and the device's local reference "ground". If not, then a differential current can occur and create all sorts of issues including a potential safety hazard.
Anyway... there is a lot to all of this and my intent was to remind others to be mindful of these details and understand the "why" things are done... and not simply do things because it was done before, or because this is required for some Manufacturers, etc. Failing to understand the why can be lead to all sorts of issues.
Cheers,
Sam
