2
\$\begingroup\$

I am working on a device that will have AC power in and out connectors. The intent is to facilitate wiring multiple units to the same power circuit without having to use a multi-outlet strip.

This box needs to pass UL safety testing. I have been researching various sources and looking at several approaches to internal power distribution and, in particular, grounding or PE (Potential Earth).

There are myriad possible configurations. I have documented three of them below.

One of the criteria (both common sense and, I believe, as a safety requirement) is that a failure of PE should not propagate through a circuit. This leads to a star configuration, where a separate wire is used to go to a PE bar with multiple grounding screws.

However, this is a typical scenario for something like the internal wiring of an automation box with DIN rails, etc. This is different, this is a stand-alone device, the size of a shoebox, with power-in and power-out. Internally it has a power supply and embedded processor PCB and a few other items.

Imagine you have 5 or 10 of these linked and plugged into a single circuit breaker for AC. I am trying to understand how UL testing will consider various circuit configurations or if they have a preferred topology.

Here are three of the many options:

Single grounding lug; Splices for L,N and PE Single grounding lug; Splices for L,N and PE The first example uses splices to connect power-in to power-out as well as the internal power supply. The PE splice contains an extra tap that then goes to the chassis ground stud via a single wire and lug, a star washer and lock nut secures it in place.

Basic DFMEA from input PE to chassis GND

  • Crimped connector pin
  • 12 AWG wire
  • Splice input
  • Splice output
  • 12 AWG wire
  • Crimped lug
  • Grounding bolt and related hardware
  • 7 potential single-point failure elements

Basic DFMEA from input PE to output PE

  • Crimped connector pin
  • 12 AWG wire
  • Splice input
  • Splice output
  • 12 AWG wire
  • Crimped connector pin
  • 6 potential single-point failure elements

A failure would, at best, cause the chassis to lose PE and, at worst, safety ground would be lost at every downstream device.

Here's an example of a splicing connector with push-in style ports. It is UL rated for 20 A, which means it is good for daisy-chaining a few of these boxes. The spring contacts are very reliable. Electricians use them. There are videos on Youtube of people torture testing these things to ridiculous extents.

Push-in splicing connector

Multiple grounding lugs on a single ground stud; Splices for L and N Multiple grounding lugs on a single ground stud; Splices for L and N This configuration wires PE from each connector and the power supply to a common chassis grounding stud. All connections terminate in a crimped circular lug. The hardware stack would include one or more external star washers as well as a nylon-insert locking nut appropriately torqued.

Basic DFMEA from input PE to chassis GND

  • Crimped connector pin
  • 12 AWG wire
  • Crimped lug
  • Grounding bolt and related hardware
  • 4 potential single-point failure elements

Basic DFMEA from input PE to output PE

  • Crimped connector pin
  • 12 AWG wire
  • Crimped lug
  • Grounding bolt and related hardware
  • Crimped lug
  • 12 AWG wire
  • Crimped connector pin
  • 7 potential single-point failure elements

It seems potential failure modes were reduced for the local device and made slightly worse for downstream devices.

Uninterrupted conductors from AC-IN to AC-OUT for L,N,PE; Insulation displacement taps to feed power supply as well as chassis ground Uninterrupted conductors from AC-IN to AC-OUT for L,N,PE; Insulation displacement taps to feed power supply as well as chassis ground

Basic DFMEA from input PE to chassis GND

  • Crimped connector pin
  • 12 AWG wire
  • Insulation displacement run-through contact
  • Insulation displacement tap contact
  • 12 AWG wire
  • Crimped lug
  • Grounding bolt and related hardware
  • 7 potential single-point failure elements

Basic DFMEA from input PE to output PE

  • Crimped connector pin
  • 12 AWG wire
  • Crimped connector pin
  • 3 potential single-point failure elements

Failure opportunities are about the same for local PE. However, it got significantly better for downstream devices.

These IDC connectors create four points of contact with the conductor. A failure would require all four to fail. While I can't put a number on it, I think this is decidedly a low probability event. Of course, proper tooling and installation procedures are critically important.

In this case the connection from in to out is solid wire from connector to connector. A failure of a pin, crimp or conductor would have to happen for downstream devices not to have a connection to PE. Insulation displacement connectors are used inside the device to tap-off of the loop conductors. There are many excellent choices for IDC connectors that can carry the required current. Here's one of them:

IDC tap connector

The question is about what UL wants. In other words, this isn't about what is "best", which can be very subjective, even for something like this.

Yes, I know the first device will see the cumulative current draw from all series-connected devices. They only draw about half an amp, so it is easy to choose connectors and wiring to support half a dozen or more of these in series.

Thanks!

EDIT: Added super-simple DFMEA analysis for each case.

Clearly there isn't a single winning option. To repeat myself, this isn't about what is best. I am trying to understand how UL look at a problem like this. If you optimize for the local device you make failure modes worse for downstream devices. Optimize for downstream devices and you won't be using the most optimal topology for the local device.

The only path that improves both is to eliminate daisy-chaining and shift the burden to the external electrical installation. In other words, a power strip of some sort or some kind of a multi-outlet installation done by an electrician.

\$\endgroup\$
1
  • \$\begingroup\$ The connector in option 3 needs exactly the correct types of wires (insulator and core) to be connected reliably. On top of that especially if the wires are run to different directions there might be stress to the connections that can even end up in cutting one of the wires. \$\endgroup\$
    – Ralph
    Feb 27, 2022 at 19:06

1 Answer 1

1
\$\begingroup\$

The last option is dangerous and that is not only a personal opinion, but will also not pass regulatory testing. A failure to install one of the connectors perfectly might end up having a loose PE-wire making contact to a live part, thus connecting the chassis of rest of the devices to dangerous voltages.

You should place a single PE-contact point within your device. Meaning a solid metallic part that has either screws to place lugs under or traditional screw connections for wires. You can put the connections under one screw or place a bus bar. You'll find suitable bus bars from component suppliers. The bus bar also needs to be a conductive enough material (plated to not oxidize if copper etc), but it can be a custom component.

The bus bar needs to be marked with a ground point symbol.

\$\endgroup\$
3
  • \$\begingroup\$ What you describe is the second option I pictured in my post. The three lugs are connected to a single stud that is press-fit into the aluminum enclosure. Single ground point with three ground wires, one from power-in, one from power-out and the third to the power supply. The power supply itself has a grounded non-anodized aluminum case which will be in contact with the product's enclosure. However, I think it would be dangerous to rely on this, or even count on it, as a legitimate power-in ground-to-chassis connection. I'm interested in understanding how UL look at this. \$\endgroup\$
    – martin's
    Feb 27, 2022 at 18:45
  • \$\begingroup\$ You are correct. I failed to mention that your examples 1 and 2 are both ok. Chassis can also be connected via a wire to the bus bar, as sometimes it's not possible to have the bus bar fixed to the chassis. Only option 3 is out of question for regulatory approval. \$\endgroup\$
    – Ralph
    Feb 27, 2022 at 19:02
  • \$\begingroup\$ I just added super-simple DFMEA for each case. Every one of these topologies is a compromise. One way I am thinking about this (which might not represent how UL views this problem) is that the failure of the local device is preferred to having ten downstream devices lose ground. If that's the criteria, then option #3 seems the clear winner. It reduces the ground failure modes for downstream devices to three from a maximum of seven. \$\endgroup\$
    – martin's
    Feb 27, 2022 at 20:42

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.