Corner Grounded Delta Question.

[4x4dually]

We have a test setup in our lab that simulates shipboard, corner grounded delta power. We are required to power some of the distribution racks this way because the Eaton power distribution units require the main and secondary inputs to be 120 degrees apart. If we try to simply power each input with different phases of the facility power, the units will ding with errors.

We are now planning on moving our lab equipment across the street. That brought me back to looking at this circuit again and now I'm questioning something. I've done a little research online and haven't exactly found the question I have answered specifically.

The question I have is do we tie the ECGs and the GEC together like we have it now in the first diagram below? This is how it's been for a while now and I am kind of scared of it honestly. Or do we corner ground the delta with the GEC which is eventually bonded to the EGCs at the first means of disconnect in the facility as shown in the second diagram below? It seems unnatural to me to tie the neutrals and the grounds together at basically the unit under test plug. Naturally I'm inclined to keep them separate but I'm also hesitant to change something we've tested product with and delivered to the customer as "passed."

Any advice on the CGD setup would be great. Thanks in advance.

 

[winnie]

IMHO this is a bit of plug in apparatus, so the NEC doesn't directly apply. Rather than debating NEC rules, I'm just debating the physics of the situation.

1) You have a newly derived system at the transformer secondary coils. Since the transformer secondary coils are galvanically isolated from the primary coils, you can connect them in the new desired arrangement (120V delta).

2) Since the 120V delta is galvanically isolated from the primary wye, you can _choose_ to ground any single terminal of the secondary to create a corner grounded delta system.

3) In diagram 1, you ground the delta by connecting to the incoming EGC. In diagram 2 you ground the delta by connecting to the incoming white wire.

4) IMHO diagram 2 is a more dangerous installation because the bonding between your 120V delta and the local chassis is via the circuitous route of the white wire back to the main service bond and then back out through the EGC.

5) IMHO diagram 1 provides a good simulation of a 120V corner grounded delta system with proper bonding of the delta.

6) IMHO diagram 1 has the problem that the supply EGC is being used as the GEC, but via a plug/receptacle. As I said, I don't think the NEC applies, so the NEC rules for GECs may not apply. IMHO a possible improvement of the design would be to provide a local 'GEC' to whatever chassis/enclosure/work surface that is supposed to simulate the ship 'hull'.

 

[4x4dually]

Good points, Winnie, and thank you for that information. We've had this system mounted to a wall next to the electrical panel for a few years. My plan is to mount it to a cart so we can wheel it back and forth between the two buildings depending on where the testing is required in the future. This as opposed to buying an entire second set of transformers and supporting devices to have it in both locations. We don't have a simulated hull to ground to per say. It will be a cart plugged into the wall and a rack sitting on a wooden pallet plugged into the cart.

I'll probably keep it as noted in the first diagram since it seems to be the safer bet. My biggest concern is if we somehow lose the EGC path, then it becomes dangerous. And as you stated that path depends on a plug/socket connection. Maybe it would be safer to mount some sort of lamp or signal from load to the EGC such that if the path is lost, the "ground present" lamp goes out. At least we'd have a visual indication of the open ground path.

 

[winnie]

If you lose the EGC, then you still have the primary circuit insulated from the equipment chassis. The secondary delta is connected to equipment chassis, but this is just like _any_ bonded generator.

You need _two_ faults for this to be an issue; loss of EGC _and_ a fault to the equipment frame, and this would then energize the equipment frame relative to other grounded material.

What about adding a separate connection (say a camlock) to the building grounding electrode system, or otherwise bonding the equipment frame to the local grounded materials?

-Jonathan

 

[4x4dually]

Concentrating on your "You need _two_ faults for this to be an issue; loss of EGC _and_ a fault to the equipment frame, and this would then energize the equipment frame relative to other grounded material" statement....if the EGC is lost, what drives how high of a potential the chassis cases rises to? Would it not rise to 120VAC with respect to ground? There is something I'm not seeing. My apologies for being a 100% visual learner and not following.

 

[winnie]

I can try to draw a picture later; working on a deadline now and I have to stop getting distracted by fun things like this forum

But why don't you try to answer: if you lose the EGC connection to the chassis, what circuit is available to raise the voltage of the chassis to 120V relative to ground? I _think_ you are concerned that the corner grounded delta might somehow pump the chassis to 120V, but this can't do that because in the scenario we are considering (EGC lost) there is _no_ connection between the transformer delta secondary and earth.

-Jonathan

 

[wwhitney]

If you lose the EGC, then you still have the primary circuit insulated from the equipment chassis. The secondary delta is connected to equipment chassis, but this is just like _any_ bonded generator.

Doesn't this latter statement depend on where you interrupt the EGC in the first diagram?

4) IMHO diagram 2 is a more dangerous installation because the bonding between your 120V delta and the local chassis is via the circuitous route of the white wire back to the main service bond and then back out through the EGC.

Without commenting on whether I agree or not, would say the difference is significant? Similar logic would imply that any longer feeder/branch circuit is more dangerous than a shorter feeder/branch circuit.

Cheers, Wayne

 

[4x4dually]

Got all the equipment pulled from the old integration lab yesterday. Had a ice/snow day with very few people here. Took advantage of the quiet time. As soon as the new cart gets here, I'll get it all mounted up and rewired so it is portable. We will need to be able to move it back and forth between here and the building across the street. Hopefully the plastic cart I got is strong enough. It's rated for 600#. These 5K xfrms are 100# each.

 

[gene6]

We have a test setup in our lab that simulates shipboard, corner grounded delta power. We are required to power some of the distribution racks this way because the Eaton power distribution units require the main and secondary inputs to be 120 degrees apart. If we try to simply power each input with different phases of the facility power, the units will ding with errors.

We are now planning on moving our lab equipment across the street. That brought me back to looking at this circuit again and now I'm questioning something. I've done a little research online and haven't exactly found the question I have answered specifically.

The question I have is do we tie the ECGs and the GEC together like we have it now in the first diagram below? This is how it's been for a while now and I am kind of scared of it honestly.

That looks correct to me, at a facility where I worked we had 'mini power centers' mounted on skids and carts, even containers with a transformer and a loadcenter and the Equipment ground in the cord had to be sized to also be a GEC, I may be misremembering but the ground pin in the pin and sleeve was actually larger.
We also had some sort of setup like you have but I never knew what it was for.

Or do we corner ground the delta with the GEC which is eventually bonded to the EGCs at the first means of disconnect in the facility as shown in the second diagram below?

In the second diagram it appears your using the primary neutral not the GEC.

IMHO this is a bit of plug in apparatus, so the NEC doesn't directly apply

I agree winnie's assessment of diagram 2, I have to respectfully disagree here, NEC definitely does not stop at outlets, construction inspectors may stop at outlets, but the NEC does not and that can be a very very expensive misunderstanding.
My understanding from our inspectors is the NEC stops where the listing begins.
And even if your cart is evaluated by a 3rd party its NEC and the 3rd party just uses the NEC on your cart.
A facility I worked had tons of equipment carts inspected (by a 3rd party) after one a facility got 100k in OSHA fines related to cord and plug stuff all of it was under the 'NEC'.