GFP/2 questions

Discussion about the Power System DC Disconnect (PSDC)

Moderators: OutBack Moderator Team, OutBack

Post Reply
Alex
Forum Member
Posts: 8
Joined: Thu Feb 19, 2004 11:01 pm
Location: Orange County CA

GFP/2 questions

Post by Alex » Sat Jul 31, 2004 5:49 pm

Question 1: the wiring diagrams of the input side of MX60s consistenly show the DC breaker upstream of the GFP/2. Is there a reason/code for this arrangement? If not is there any obstacle in reverting the order of these 2 devices?


Question 2: can (2) or more GFP/2s work when tied together to the same shunt?

From the installation doc, the GFP/2 unit has one sensing connection to the equipment ground and another to the shunt. I understand the unit compares the PV array current to the current going thru the shunt (how it compensates for the current tapped from the array to the inverter is unclear to me since this current is not "visible" at the shunt).

Anyhow, it seems to me that several GFP/2 using the same shunt as reference can't work properly. The current measured at the shunt will be the sum of each MX60 contribution minus current going to the inverters; this figure will be substantially different from the current measured by each GFP/2...


Thank you for helping me understand these issues,

Alex

Kent Osterberg
Forum Emperor
Posts: 1523
Joined: Mon Jun 09, 2003 11:01 pm
My RE system: SMA Sunny Boy 2500 with eight Mobil Solar Ra 180 modules
Location: Cove, Oregon
Contact:

Re: GFP/2 questions

Post by Kent Osterberg » Sun Aug 01, 2004 11:19 am

Question 1: the wiring diagrams of the input side of MX60s consistenly show the DC breaker upstream of the GFP/2. Is there a reason/code for this arrangement? If not is there any obstacle in reverting the order of these 2 devices?
Operation of the GFP/2 will not be compromised by changing the order of the PV disconnect breakers and the GFP/2. In fact, NEC 690.13 seems to allow the use devices such as the GFP/2 as a pv array disconnect. However, NEC 110.3(B) requires installation per the listing or labeling. One reason for having the GVP/2 and PV array disconnects in the order shown is so that the PV array may be disconnected where it first enters the enclosure; this minimizes exposure to live terminals during maintenance. Second, the GFP/2 should not normally be used as a PV array disconnect device because it also replaces the normal system ground with a high impedance; effectively ungrounding the system.
Question 2: can (2) or more GFP/2s work when tied together to the same shunt?
Two GFP/2 devices cannot be used in one system. Ground fault protection and array disabling for more than two circuits would only be required when you have more PV array current than two MX-60s can handle and the PV arrays are roof mounted. This would require a GFP device that is capable of handling three or more circuits.
From the installation doc, the GFP/2 unit has one sensing connection to the equipment ground and another to the shunt. I understand the unit compares the PV array current to the current going thru the shunt (how it compensates for the current tapped from the array to the inverter is unclear to me since this current is not "visible" at the shunt).
The GFP/2 doesn't compare the plus and minus array currents in the manner that at GFCI compares hot and neutral currents in ac circuits. Ground fault protection is provided by the GFP/2's 1-amp circuit breaker. Normally there is no current (or virtually none) in the system grounding connection. The GFP/2 doesn't rely on the shunt to observe the ground current.

Alex
Forum Member
Posts: 8
Joined: Thu Feb 19, 2004 11:01 pm
Location: Orange County CA

Post by Alex » Sun Aug 01, 2004 2:06 pm

Kent,

Thank you very much for your prompt reply, and on a weekend day on top of that.

I understand now the rationale of having the DC breakers upstream of the GFP/2 as to provide means of disconnection.

However, I am a bit at a loss regarding the use of multiple GFP/2s in a system. If 2 or more GFP/2s can't be used, how can I implement ground fault protection in a system using more than 6kW of PV power (each MX60 handles about 3kW of PV power and each GFP/2 can protect 2 arrays) ?

Outback's product catalog (3rd edition, page 17) shows a table labeled "small breaker/ GFP combinations" that seems to imply that up to 2 GFP/2s can be installed in a PSDC enclosure. How is this made?

Sorry to come back with more questions...

Alex

Kent Osterberg
Forum Emperor
Posts: 1523
Joined: Mon Jun 09, 2003 11:01 pm
My RE system: SMA Sunny Boy 2500 with eight Mobil Solar Ra 180 modules
Location: Cove, Oregon
Contact:

Re GFP/2 questions

Post by Kent Osterberg » Sun Aug 01, 2004 5:51 pm

Outback's product catalog (3rd edition, page 17) shows a table labeled "small breaker/ GFP combinations" that seems to imply that up to 2 GFP/2s can be installed in a PSDC enclosure. How is this made?
Alex, you got me on that one. In fact on page 18 the catalog says "up to three DC-GFP/2 units can be installed in a single PSDC enclosure."

Putting in two or three GFP/2s with the ground current sensing breakers in series would keep the threshold at 1 amp. In this configuration, only one GFP/2 would trip. Unfortunately, not necssarily the correct one to disable the array with the ground fault. Series doesn't do the whole job that the GFP/2 is required to do.

Putting two or three GFP/2s with the ground current sensing breakers in parallel (all of them installed per the instructions for installing one) would increase the ground current required to trip the GFP/2s to two or three amps. When the first one trips, the remaining ones will trip too and all of the arrays will be opened. But increasing the detection threshold to 2 or three amps is probably not acceptable.

rdmcghie
Forum Whiz
Posts: 44
Joined: Mon Mar 25, 2002 11:01 pm
Location: Chico, California

Post by rdmcghie » Sun Aug 01, 2004 8:59 pm

OutBack's website http://www.outbackpower.com/OBDCGFP.htm list the trip current as 1/2A. Thus, any kind of direct short should trip any reasonable number of GFP/2s. The problem would be with high resistance shorts. At 48V nominal it appears that two GFP/2s could allow a 50W short without tripping. If this power was dissipated over a large enough area it should not be a problem but in a small confined area it might present a problem. However, it would appear that a large enough area may not be that large. At 1000W per square meter (full sunlight) 50W would equal the energy falling on a 9 inch square.

Robert

Christopher

Re: Re GFP/2 questions

Post by Christopher » Wed Aug 04, 2004 8:47 am

Wow - great discussion - I approve of Kent's screen name...

The OutBack GFP/2 does have a trip level of 1/2 amp instead of 1 amp like the "other" similar GFP offering by a former company. This lower current level was done specifically to allow paralleling of multiple GFPs in the same system - doing with 1 amp devices results in a GFP level which is too high in my opinion.

As far as the detection level goes - SMA uses a 1 amp trip level (a fuse actually) in their GFP system built into the grid-tie inverter. At their voltages this results in a several hundred watt fault level required... so 50 watts seems tolerable.

The order of the breakers is done to make things easiler to understand and less hazardous - having a power source go directly to the disconnect makes the most sense to me.

Using the GFP as the disconnect is a gray area of the code and UL standards - the AEI (now Beacon) MM-5 has done this and still does - but I don't think its the best idea. We are going to do it that way on the upcoming PS1 grid-tie package we are working on - but that is a completely listed prewired system specifically for grid-tie applications with a single MX60.

Thanks for all helping out on the forum! :D

Christopher
8)

Post Reply