Bridge Rectifier Issues / Wall Outlet

[mar2020]

I am posting a thread similar to an old thread that Crutschow chimed in on with good advice; our application, however, *slightly* differs from the OP.

We ran into a similar "spark" issue when we energized, via a standard 120VAC wall outlet (service line at building) an AC/DC power supply /bridge rectifier that is installed on a piece of trailered equipment for off-grid use. We noticed that the trailer equipment assemblers jumpered N and G within the box on the trailer, so basically another jumper downstream of the neutral-to-earth bond at the breaker at our building.

We are going a bit back and forth about what to do. Here are some of the known facts:

- When we remove the jumper from N and G in the trailer equipment box that is downstream of the breaker box in the building, but upstream of the AC input to the bridge rectified power supply, we do not have issues with the power supply performance, or any sparking.

- When the jumper is used, it sparks.

- We use similar AC/DC power supplies on other equipment, where G from the AC input and V- from the isolated DC output are chassis grounded. I have not had issues with these, presumably because I do not jumper N and G.

My questions are:

What is the right approach here? Do we continue to not jumper N and G? As far as I understand (and I say that loosely) NEC prohibits secondary neutral-to-earth bonds anywhere other than the service panel / origin except in special circumstances with secondary equipment.

Or do we jumper N and G (seemingly a needless echo that they are ultimately jumpered somewhere else) and just place a diode between the chassis ground of the trailer and the ground of the wall outlet? (See pic) I have never seen this approach done on any oil&gas / industrial installation I've worked on.

Thanks

 

[gophert]

A single earth ground insures the current into L is equal to the current out at N - and no current travelleing through earth ground to supply current to your devices. A difference between L and N will trip ground fault interrupters and leaves no way to identify ground faults.

Do you know anything about the spark? Is there a Dc bias causing the spark when secondary ground cable is not used?
Is the grounding cable adiquately bonded at each junction? What else can you tell us about the size or sound of the spark (a small static spark vs a loud snap).

 

[unclejed613]

ground should go to the chassis/frame. neutral is connected at the box. exactly what is running off the line? if you run a bridge rectifier directly off the line, and have something on the other side of the bridge rectifier connected to chassis ground, you're going to have trouble. what is usually done, is the bridge rectifier is part of a power supply (usually a SMPS ). the bridge rectifier and associated hot side circuits are isolated from the chassis ground. on the cold (secondary) side the power supply ground is bonded to the chassis.

 

[mar2020]

Gophert: According to the person who plugged it in at the extension cord run from the wall outlet, it was a "fireball." As far as I know, it is adequately bonded. We use this source, outlet and cord very frequently.

UncleJed: The line from the 120V source is powering two bridge rectified 120VAC/12VDC power supplies, both DIN rail mount, both sharing DC-. The equipment cases of these supplies, like most others, are grounded to the G terminal and of course through the rail since we bolt a chassis ground to the box, which the rails are mounted on. We have always wired these types of projects like automotive applications for the DC side. I haven't seen any warnings on the manufacturer documentation of the supplies saying not to chassis ground the DC side. There are many of the trailers out now running normally (as far as I can tell) with N and G not jumpered in the box.

What are we missing here? Why do any of the power supplies work as common chassis ground on both sides if this isn't due to jumping N and G? I am aware N and G are typically bonded at breakers. I have also always seen NEC mention G is not meant to be used as a current carrying conductor. Jumping N and G, as Gophert mentioned, would also trip GFIs routinely.

 

[unclejed613]

if i got your drawing interpreted properly, you have a diode between L1 and B-(which is tied to ground)... that alone is enough to create a "fireball". L1 and N should go to the AC input terminals on the power supplies. if there's a frame ground for the supply it should go to the panel ground. the AC ground from the wall should also be tied to the panel ground. don't assume N is N at the wall outlet, as it's a common error made in outlet wiring to have N and L reversed, which is one reason NEC says not to tie N to chassis in the equipment. you can verify proper wiring with one of those outlet testers from home depot, it will tell you if N and L are reversed and whether or not G is grounded.

 

[gophert]

Do you really have an earth ground? A grounding rod driven into soil. Very rocky or dry sand do not count if there is no moisture to conduct with the rod and bring it from the isolation-transformer's float to earth ground.

Also, your diodes are making the pole you labeled "B+" the negative DC pole. I think you need to flip your diodes.

 

[mar2020]

Sorry for the confusion. The two diodes (one between B- and L1, one between B+ and N) were scribbled on there when we were analyzing current flow during one of the swings of the AC voltage. Those do not contain diodes. The third diode, between the Wall and Box grounds, was the proposed idea to prevent a short from L1 to N/G (N/G because it is jumpered in the current wiring which correlated to the spark.)

 

[unclejed613]

in the example shown here would be the proper way to wire a panel with two power supplies. two common methods of naming the input terminals are shown here. the incoming AC power is at the left. the incoming ground wire goes first to the panel ground, then to the terminal strip. the incoming hot and neutral go directly to the terminal strip. the AC breakouts for the power supplies go from the terminal strip to the power supplies. notice that the neutral (shown here as grey, but USA color code would be white) does not connect to the panel ground. similarly, the DC outputs have the negative terminal tied to chassis ground and power wires go to the equipment being powered. power supply AC inputs may sometimes not have a ground terminal, as it may be connected through the chassis mounting of the supply.

in industrial equipment the green incoming ground wire is a couple of inches longer than the hot and neutral. if the cord ever gets accidentally pulled loose, the ground wire remains connected after the hot and neutral break off.

 

[mar2020]

That's exactly how we wire our equipment, minus a few of the termination hardware pieces used. This configuration has traditionally worked well for us, with no disruption to the performance of the power supply. It is when we take the grey/N/L2 and green/G and jump them together in the equipment box that contains the power supplies that we see "fireballs." While I maintain it's not suitable for sourcing power from GFI equipped sources nor NEC-kosher and wouldn't recommend doing it, we are trying to understand why that's happening if N is ultimately bonded to earth ground reference at the breaker.

I feel like there's a way to describe why return current must propagate back via N and it make physical sense to isolate it at the box despite it ultimately being tied to ground at the source, but I haven't found that grace yet. i.e., it ohming out, but requiring that current flow not take that branch (G) back from the equipment enclosure.

 

[unclejed613]

sounds to me like the outlet has N and L backwards... it's simple enough to stick a voltmeter probe into a slot in the outlet, and the other probe to the outlet ground and see which one of the L or N slots is "hot".

we are trying to understand why that's happening if N is ultimately bonded to earth ground reference at the breaker.

it is bonded to G at the box, but if it's on the wrong pin at the outlet, then your N wire in your equipment is really L, and that's what's causing the fireworks display. something plugged into the outlet will still work properly because N and L are there, and stuff that runs off AC doesn't care which is which (after all, you can plug stuff in that doesn't have a ground pin and it works). it's when you try to tie G to N in your panel where having the two wires in the outlet reversed is causing problems. you should be able to tell with a voltmeter whether L or N is hot with respect to G. you can even do that test in your panel. one of the wires is going to be hot measured from ground. if it's the N wire, then the outlet wiring is reversed. from your panel ground, L should be hot, N should be 0. you will only find out for sure by measuring. DON'T try to ohm out a live AC line, ohmmeters aren't made to handle 120Vac across the probes, you will get a fireball in your meter.

 

[mar2020]

I thought about that and assumed no-dice because I feel like we've used the outlet a lot in the past with ground connection devices. I will double check.

 

[mar2020]

By the way, in your traditional rectified PS there must be further isolation in addition to just after the bridge, which is why this has been working otherwise?

 

[gophert]

By the way, in your traditional rectified PS there must be further isolation in addition to just after the bridge, which is why this has been working otherwise?

Do you have a scope? It looks like your diodes are backwards to get B+. I think that node would be negative.