LFMC instead of expansion fitting on long Rooftop EMT conduit Run

Status
Not open for further replies.
Okay, we'll how about that as a justification for why the dangers with PV circuits are higher. :lol:

Seriously, I actually thought about that. It gets into a bit of a "NEC philosophy" question. Say with correct quality installation and all other things equal, System A and B have the same reliability. Thus each connection or termination in each system has the same failure rate. Say system B requires twice the connections. Or say system B tends to be installed by less experienced workers. Do we have stricter requirements and methods for the system B's?
 

iwire

Moderator
Staff member
Location
Massachusetts
Seriously, I don't have that much of a bone to pick in this fight, but I think that there have been some articles published in PV circles recommending expansion joints because:

Seriously neither do I, I simply find it misleading to say PV circuits are somehow more dangerous than other circuits and as such need special care when installing raceways that contain them.


1) PV circuit conduits usually must be run on hot rooftops where the issue is most extreme

This in itself is nothing new at all, this has been done since there was electrical work.


2) There are unique dangers (I'll withdraw 'higher') to PV circuits which argue for above-code standards.

You keep saying vague things like this without ever specifically citing a danger that you see.


3) There has been at least one well-known incident in which the lack of expansion joints was thought to have contributed to major damage.

First there is not type of electrical installation in the world we can't find at least one indecent of failure.

And secondly when I look at figure 2 in the link you sent it says the failure occurred in an expansion joint. That hardly suggests I should start adding more of them.


This thread has gotten crazy, can we just agree that following the NEC and using good installation practices are needed for any power wiring?
 

SolarPro

Senior Member
Location
Austin, TX
DC arc-fault detection (Section 690.11) and enhanced ground-fault detection (Section 690.5) would have prevented the Bakersfield Fire.

Rapid Shutdown (Section 690.12) is geared more toward first responders, not fire prevention.
 

jaggedben

Senior Member
Location
Northern California
Occupation
Solar and Energy Storage Installer
You keep saying vague things like this without ever specifically citing a danger that you see.

No, you keep not listening to what I say. Old school PV circuits have no kind of shut down mechanism such as breakers, fuses, GFD, GFCI, etc, that most other circuits have. There you go. If you think it doesn't warrant higher standards of raceway installation, then we can agree to (not even really) disagree. But I specified the concern.
 

jaggedben

Senior Member
Location
Northern California
Occupation
Solar and Energy Storage Installer
So the use of an expansion joint caused the fire

Well shoot. I think that the caption has it incorrect. It was the lack of an expansion joint that caused the fire. (You don't see an expansion joint in the picture, do you?)
(SolarPro, can you comment? Are the captions correct or not?)

If you read the article, the use of expansion joints was one of the recommendations from the original report.

The reason I raised figure 2 (diagram) was regarding the amps of the output circuit.
I agree that the thread has gotten crazy at this point.
 

jaggedben

Senior Member
Location
Northern California
Occupation
Solar and Energy Storage Installer
Seriously, I actually thought about that. It gets into a bit of a "NEC philosophy" question. Say with correct quality installation and all other things equal, System A and B have the same reliability. Thus each connection or termination in each system has the same failure rate. Say system B requires twice the connections. Or say system B tends to be installed by less experienced workers. Do we have stricter requirements and methods for the system B's?

Right. Code doesn't care, but it's a consideration from a business standpoint.

Simply put, I design my systems and train my guys to exceed NEC requirements in a few areas precisely because I know that much of my crews aren't experienced electricians.

For example, running a wire EGC for PV circuits even though strictly speaking I could just use the EMT. Might get into some arguments with a few AHJs on details. But regardless, I'm simply not going to rely on outdoor EMT connectors alone to carry fault current, when many of them are installed in hard to reach places and exposed to the elements. I expect my crews to do their best and I still don't count on that. It costs some money to run a wire but it's justified.
 

ggunn

PE (Electrical), NABCEP certified
Location
Austin, TX, USA
Occupation
Electrical Engineer - Photovoltaic Systems
DC arc-fault detection (Section 690.11) and enhanced ground-fault detection (Section 690.5) would have prevented the Bakersfield Fire.
I don't agree. The fault was a positive to negative DC short through the conduit and the inverter was not in the current path; nothing that the inverter could have done would have quenched the arcs. Arc fault and enhanced ground fault detection would have shut the inverter down and hopefully attracted someone's attention but they could not have directly prevented the fire.
 

69gp

Senior Member
Location
MA
Getting back to the question that was originally posted and some of my views on items talked about here.

I don't think that LFMC qualifies as a substitute for an expansion fitting. The code list the use of LFMC for where you need flexibility does not say anything about expansion. From looking at the photo i would say that the application shown is a code violation as it could easily be stepped on or material could fall on the flex therefore damaging it. 350.12 (1).

I am in the Northeast it would be like saying when you come out of the ground with PVC it would be ok for me to use LFMC in place of an expansion fitting. I don't think that will fly with the AHJ.

As for the separated conduit it looks as if the conduit was never seated all the way into the coupling. It would have been nice to see the next set of couplings down to see how they lined up.

As for the use of expansion fittings on the roof the NEC sets just the minimum. If the plans call for expansion fittings installed every XXX feet of conduit run you are obligated to install them per the customer request. If going over a building expansion joint you need them. Also you need to be careful as some expansion fittings are not internally grounded and you need to use ground clamps with a jumper.
 

SolarPro

Senior Member
Location
Austin, TX
I don't agree. The fault was a positive to negative DC short through the conduit and the inverter was not in the current path; nothing that the inverter could have done would have quenched the arcs. Arc fault and enhanced ground fault detection would have shut the inverter down and hopefully attracted someone's attention but they could not have directly prevented the fire.

The enhanced ground-fault detection required in NEC 2014 would have detected the first fault. The only reason this became a conflagration is because that first fault went undetected.

Here's is a great article on precisely this topic:

Resolving Fire Hazards from the Ground-Fault Detection Blind Spot


This Table is especially compelling:
 

Attachments

  • 5_SP_8_6_pg12_QA_fig_5.jpg
    5_SP_8_6_pg12_QA_fig_5.jpg
    75.5 KB · Views: 0
Last edited:

ggunn

PE (Electrical), NABCEP certified
Location
Austin, TX, USA
Occupation
Electrical Engineer - Photovoltaic Systems
The enhanced ground-fault detection required in NEC 2014 would have detected the first fault. The only reason this became a conflagration is because that first fault went undetected.
I understand, but actually preventing the fire would have taken human intervention after the fault was detected. I have known cases where a customer's inverter shut down for some reason and it was months before anyone noticed. Rapid shutdown would have disconnected the DC conductors at the array. But of course, rapid shutdown depends on the tech you speak of in order to register the fault.

But we are splitting hairs, I guess. We can agree that there have been several developments in PV technology since Bakersfield which would have at least made that fire a whole lot less likely.
 

SolarPro

Senior Member
Location
Austin, TX
Right, someone has to go out and do something about the first fault. But if the inverter had shut down and the investors were losing money, that would have gotten fixed quick!
 
Status
Not open for further replies.
Top