16 ATS's all 3 pole and the matter of Ground Fault

Status
Not open for further replies.
mshields

mshields

Senior Member
Location
Boston, MA
I have been tasked with replacing two aged generators, combining them into two paralleled units. The load on these plus the new load we'll have means that we are going to be looking at about 1500kW worth of generation. It had been 1000kW until recently. And when it was, I was gravitating toward the Generac Gemini whose breakers on the output of each 500kW generator is 900A. I like this, because, it meant that I could feed the Emergency Generator Switchboard with two 900A breakers with all feeder breakers less than that.

I liked it because that meant I didn't need to have ground fault and that, in turn meant, that I wouldn't have to replace 16, 3 pole ATS's with 4 Pole ATS since ground fault protection at the generator won't work with 3 pole ATS's.

What I want to know is whether or not I am either over thinking this or thinking about incorrectly. I would welcome your input.

Thanks,

Mike
 
I never understood (perhaps I didn't try hard enough) why the ground fault protection does not work with a 3-pole ATS. Is there an easy explanation? When you say "ground fault," does planet Earth come into the current path? I ask that because I have seen that phrase to refer to a connection between an ungrounded conductor and the enclosure of a component.
 
charlie b said:
I never understood (perhaps I didn't try hard enough) why the ground fault protection does not work with a 3-pole ATS. Is there an easy explanation? When you say "ground fault," does planet Earth come into the current path? I ask that because I have seen that phrase to refer to a connection between an ungrounded conductor and the enclosure of a component.
Click to expand...
Charlie,
It usually has to do with the parallel paths caused by having multiple ground to neutral bonding points associated with multiple separately derived sources.
 
Usually if you have that large of a single generator, it will have ground fault protection, same as the service, the generator would need to be a sds at this point, in order for the ground fault switch on the generator to operate correctly. The neutral would need to be bonded on the line side of each gf breaker to work correctly. If a three pole switch is used, that would defeat the ground fault control on both gf mains because of parallel paths. That’s why a four pole switch is needed if the service is gf. But if the generator is small enough, and does not require gf protection, then a three pole switch, along with the generator being non sds is acceptable. In the OP’s case, I don’t know if it can be considered 900 amps if the breakers are paralleled. Should be a total of 1800 amps.
 
The multiple ground paths desensitize the ground fault relay. There is a good explanation in the PIECE green book and an evenbetter one at the Cummings generator website.
 
There are ground fault schemes, like Modified Differential, that can be applied an systems with multiple source N-G bonding.
 
This one of the many reasons why I advocate everything being connected phase-phase.

What code article mandates GF protection on a gen?
 
It can also cause nuisance tripping due to some current bypassing the gf ct. I was troubleshooting a new installation where the complained every time a large air compressor started, the gf main would trip. This particular gear had a large window CT that all the cables passed thru. Turns out, they pulled one wire a little short, and went under the window, instead of through it. The start up current was enough to trip 100 amp setting on the breaker. Lowered the CT to encompass all the wires, no more problem.
 
hillbilly1 said:
Usually if you have that large of a single generator, it will have ground fault protection, same as the service, the generator would need to be a sds at this point, in order for the ground fault switch on the generator to operate correctly. The neutral would need to be bonded on the line side of each gf breaker to work correctly. If a three pole switch is used, that would defeat the ground fault control on both gf mains because of parallel paths. That’s why a four pole switch is needed if the service is gf. But if the generator is small enough, and does not require gf protection, then a three pole switch, along with the generator being non sds is acceptable. In the OP’s case, I don’t know if it can be considered 900 amps if the breakers are paralleled. Should be a total of 1800 amps.
Click to expand...
Right. To amplify on that, an individial external bond between any two main neutrals or main and generator neutral and a common GES will allow some service A's neutral current to travel over the panel A bond to the panel B bond to the panel B neutral, causing a false trip of a panel A main GFCI.
The situation is similar when operating even one panel from the generator because some generator neutral current will come back to it through the EGC or ground and the generator bond, tripping the generator and potentially panel A GFCI breakers.

Sent from my Pixel 4a using Tapatalk
 
mbrooke said:
This one of the many reasons why I advocate everything being connected phase-phase.

What code article mandates GF protection on a gen?
Click to expand...
With the dual source bus, the only way I can see, or have seen done, is to have all line to line connected loads and absolutely zero for any line to neutral connected loads.

Three pole switch is possible, solidly connect the neutrals. Stop carrying the neutral at the first point code allows. The system should be clearly marked as three phase, three wire, (four or five conductors, separate insulated N and G). I would be sure in a large older dual source bus installation, the haphazard existing N to G connections could be in lots of places, and then when you find some and measure the (neutral current on the system bonding jumper).

If you keep the all thee pole switches, I would bet on a load survey of the existing loads, specifically only for neutral connected loads. And then move those loads downstream of a line to line connected transformer.

Lighting load in particular is what they would have wanted to put on the neutral. Guys did not know that, that system does not have a neutral. It's three phase three wire. I would likely assume, with the multiple N to G connections, I would not use that neutral for loads. solidly connected neutral with no load on it, is the more reliable system.
 
__dan said:
With the dual source bus, the only way I can see, or have seen done, is to have all line to line connected loads and absolutely zero for any line to neutral connected loads.

Three pole switch is possible, solidly connect the neutrals. Stop carrying the neutral at the first point code allows. The system should be clearly marked as three phase, three wire, (four or five conductors, separate insulated N and G). I would be sure in a large older dual source bus installation, the haphazard existing N to G connections could be in lots of places, and then when you find some and measure the (neutral current on the system bonding jumper).

If you keep the all thee pole switches, I would bet on a load survey of the existing loads, specifically only for neutral connected loads. And then move those loads downstream of a line to line connected transformer.

Lighting load in particular is what they would have wanted to put on the neutral. Guys did not know that, that system does not have a neutral. It's three phase three wire. I would likely assume, with the multiple N to G connections, I would not use that neutral for loads. solidly connected neutral with no load on it, is the more reliable system.
Click to expand...

Agree, but it seems like code doesn't allow 480 volt lighting.
 
Status
Not open for further replies.
Top