• Have you tried out dark mode?! Scroll to the bottom of any page to find a sun or moon icon to turn dark mode on or off!

diy solar

diy solar

Stand Alone Solar Arc-fault protection devices.

That link has a lot of good PV fire pictures.

They seem to be pushing "low voltage DC" very hard.
They said AFCI won't protect against arc to ground. I think it will if fault only a 1 place, no second path to ground available.
A fault + to - would keep burning. So situations can develop you can't stop.
As it said, micro inverters run on voltage of single panel, and optimizers can drop the voltage significantly.

400bird said:
Informative, but pushing for micro inverters.
Click to expand...

Yup.


The commercial scale PV world uses high voltage strings of panels. You can get a 4 MW inverter from SMA, feed it a huge fused PV array.
Arc-fault can be a problem. WalMart had 7 fires from Solar City installations, and some went up even with operation of the system shut off.
I don't know how to make that failsafe except with electronics (or electro-mechanical) devices per panel, and I hate to add the complexity.
 
Something like that, module-level RSD.
Coupled with AFCI control to activate RSD.

These are usually switch-mode power supplies, take in PV panel voltage and perform either optimizing, or drop to a low (e.g. 1.0V) output in shutdown.
One issue is if their own playing MPPT and outputting a different power/voltage curve messes with the head of string level MPPT.
Another would be interference (that could be mistaken for arc-fault), which this one claims to avoid:

"STABLE
  • Eliminates Unwanted AFCI Nuisance Tripping
  • RSD-S-PLC Noise Spectrum Density is far away (isolated) from AFCI noise"
 
I’d like to go roof top. But I’m just not going to do it without tier 1 panel level shutdown capability and all fault detection.

My roof will handle about 30 panels; so it’s another roughly $1,000 for my off-grid cabin. Worth it.
 
scottvanv said:
I'm pretty frustrated by this ... even the "built in" AFCI in the inverters don't do enough ... here's a rather relevant article out of Australia (where they use "rooves" which made me chuckle): https://www.acsolarwarehouse.com/news/solar-fires-dc-arc-faults-on-solar-systems/
Click to expand...
mrzed001 said:
Wow ... this is a very good article.
Click to expand...
Just read it - fantastic article. Clear and explains/shows the issue.

This is one of the reasons I soldered my wire -> MC4 connectors instead of just crimping. Hoping my Midnite Classic Arc Fault detection will cover one of the cases / better than nothing. :)
 
OffGridInTheCity said:
Just read it - fantastic article. Clear and explains/shows the issue. This is one of the reasons I soldered my wire -> MC4 connectors instead of just crimping. I have a cement tile roof and I'm hoping not to have to fool with panels up there for the rest of my life :)
Click to expand...

Series arc-fault was already handled by AFCI.
If a rodent chews through your wires and you get a parallel fault, it will arc.
Only fix we've identified is to plug those MC connectors into module-level shutdown. But your fireproof roof is a plus.

Remember pay phones? They had a metal loom around the handset cable. that should be fairly rodent-proof. Flexible metallic conduit?

CALPIPE S61000CTFX 1" STAINLESS STEEL FLEX CONDUIT 25' COIL | Gordon Electric Supply, Inc.

CALBRITE S61000CTFX 1FT SS316 STAINLESS STEEL BARE FLEX CONDUIT
www.gordonelectricsupply.com www.gordonelectricsupply.com
 
Hedges said:
Series arc-fault was already handled by AFCI.
If a rodent chews through your wires and you get a parallel fault, it will arc.
Only fix we've identified is to plug those MC connectors into module-level shutdown. But your fireproof roof is a plus.

Remember pay phones? They had a metal loom around the handset cable. that should be fairly rodent-proof. Flexible metallic conduit?

CALPIPE S61000CTFX 1" STAINLESS STEEL FLEX CONDUIT 25' COIL | Gordon Electric Supply, Inc.

CALBRITE S61000CTFX 1FT SS316 STAINLESS STEEL BARE FLEX CONDUIT
www.gordonelectricsupply.com www.gordonelectricsupply.com
Click to expand...
Yes sir. In my case I have a cement tile 'shingle' roof under my panels on the roof, so that's some comfort in terms of fire catching easily and I don't have to replace the roof anytime soon.
1645035066154.png
 
Hedges said:
1) You would think. But even after they shut off systems on surviving store roofs, more caught fire. With zero MPPT current drawn from them.
Something about faulty or incorrect/incompatible MC connectors. But that doesn't compute.

2) Only counting the ones up right now. Actually, I think that's 138 up. I'm in the process of swapping around; upgrading, hope to deploy more.
Not sure what I'll use all the power for. Any ideas?
Click to expand...
Have you considered pouring anvils to put your excess power to work? I myself would probably run grow lights. Pound for pound - weed is worth quite a bit more than steel anvils. I'm still searching for stand alone afci.
 
I'm still pondering ARC fault.

Here's the code related to arc fault.

Screenshot_20220424-124150~2.png

How do you interpret the exception?

"Not on buildings, or direct buried, or metallic raceways, or metallic trays"
Meaning that if I put my DC conductors in conduit (emt) for the full run, I could skip the arc fault detection?

Or

Does the exception only apply to PV systems not on buildings?
 
400bird said:
I'm still pondering ARC fault.

Here's the code related to arc fault.

View attachment 92456

How do you interpret the exception?

"Not on buildings, or direct buried, or metallic raceways, or metallic trays"
Meaning that if I put my DC conductors in conduit (emt) for the full run, I could skip the arc fault detection?

Or

Does the exception only apply to PV systems not on buildings?
Click to expand...
Excellent question. The wording kinda implies 'OR' but it is not very clear.

Since the PV arc-fault requirements are all about fire prevention (not shock prevention), reading it as 'OR' kinda makes sense. (If the wire is enclosed in metal, an arc would have difficulty starting a fire)

BTW: The Solar Charge Controller is a "DC-DC converter" Consequently, it's output has to be in metal to meet the exception. Where does the metal have to stop? at a common bus-bar? at the battery?


Another interesting point is that the exception only talks about PV output circuits and not source circuits.

PV Output Circuit. The dc circuit conductors from two or more connected PV source circuits to their point of termination.

PV Source Circuit. The dc circuit conductors between modules and from modules to dc combiners, electronic power converters, or a dc PV system disconnecting means.
Click to expand...

This implies that you can have standard MC4 cables to the combiner box, but it must be in metal conduit (or buried) all of the way after that.
 
FilterGuy said:
Excellent question. The wording kinda implies 'OR' but it is not very clear.

Since the PV arc-fault requirements are all about fire prevention (not shock prevention), reading it as 'OR' kinda makes sense. (If the wire is enclosed in metal, an arc would have difficulty starting a fire)

BTW: The Solar Charge Controller is a "DC-DC converter" Consequently, it's output has to be in metal to meet the exception. Where does the metal have to stop? at a common bus-bar? at the battery?
Click to expand...
Hmmm
But, with out the PV you don't need arc fault from the inverter to battery.
(Edit: wtf does that mean? I wish I proofread this stuff)
Maybe the break point is where you step down below 80 volts? So, at the charge controller?

FilterGuy said:
Another interesting point is that the exception only talks about PV output circuits and not source circuits.

This implies that you can have standard MC4 cables to the combiner box, but it must be in metal conduit (or buried) all of the way after that.
Click to expand...
What if you don't have a combiner box? Would it then be a "PV source circuit" and the arc fault detection exception is not applicable? Arc fault would be required.
That doesn't make sense! Why would combining more circuits together make it ok to go without arc fault?
 
Last edited:
400bird said:
Maybe the break point is where you step down below 80 volts? So, at the charge controller?
Click to expand...
Sounds reasonable to me...... but I am just another guy trying to figure out the code.

400bird said:
Hmmm
But, with out the PV you don't need arc fault from the inverter to battery.
(Edit: wtf does that mean? I wish I proofread this stuff)
Maybe the break point is where you step down below 80 volts? So, at the charge controller?


What if you don't have a combiner box? Would it then be a "PV source circuit" and the arc fault detection exception is not applicable? Arc fault would be required.
That doesn't make sense! Why would combining more circuits together make it ok to go without arc fault?
Click to expand...
There have been complaints about the poor definitions in the NEC about PV source and output circuits. The general consensus is that it is a source circuit till you combine two or more in parallel and then it becomes an output circuit.

So.... if you have two strings that tie together with one of those MC4 Y connectors, the strings are source circuits and the outputs of the Y connectors are output circuits.


Now add to all this confusion the fact that the local code inspector may interpret it all differently...... and it is anyone's guess what he will pass.
 
FilterGuy said:
BTW: The Solar Charge Controller is a "DC-DC converter" Consequently, it's output has to be in metal to meet the exception. Where does the metal have to stop? at a common bus-bar? at the battery?
Click to expand...

At 80V? (As the Bird said)

FilterGuy said:
This implies that you can have standard MC4 cables to the combiner box, but it must be in metal conduit (or buried) all of the way after that.
Click to expand...

Which wouldn't protect against arcs and fires from all those connections. Isn't that what burned the Walmarts? (even when system switched off, which means arc-fault wouldn't stop it but RSD might.)
 
Hedges said:
At 80V? (As the Bird said)



Which wouldn't protect against arcs and fires from all those connections. Isn't that what burned the Walmarts? (even when system switched off, which means arc-fault wouldn't stop it but RSD might.)
Click to expand...
I can barely understand what the exception is trying to say..... I certainly can't defend it.

In general, the NFPA has good reason for everything they put in the code, but they are not perfect. The whole PV space has been growing and changing rapidly and the NFPA process for developing changes to the code runs pretty slow. Furthermore, the number of people that participate in the PV areas of the code development is a lot smaller than the rest of the code. I can imagine that this is an area where changes happened faster than the process could handle and got a bit screwed up. (Having said that.... this exception has been in for several revs of the code with very little change. )

When I search on the web for information on this part of the code, all the articles seem to concentrate on the requirement for PV arc fault protection and not on the exception. It seems that most people are just doing the protection.
 
Nothing wrong with arc-fault, so long as it doesn't nuisance trip. Some models from SMA had that problem in the past.
It is just a bit of computation, and a signal telling inverter to stop operating.

There is a problem with RSD, about 10% to 50% price adder on the PV array. It also doubles the number of MC connections and 10x the number of mixed brand connections.

Since arc-fault can't stop all arcs, arc-fault tripping RSD may be the most reliable.

RSD could be built into the junction box. As an external add-on such as presently marketed, it would require a 600V or 1000V blocking element, unless it shorts the panel and carries current. What I think they did instead is an 80V blocking element, probably a bypass diode (to avoid ever seeing 600V to 1000V), and a buck converter that delivers 1V during shutdown.

In the junction box I think it could be just a FET per diode-bypassed section disconnecting PV cells. But get thermal management of bypass diode correct this time.
 
I'd love to include AFD in my plans for a DC coupled array.

My issue is cost. I haven't found a reasonably priced stand alone arc fault detector. I would probably wire it into the RSD as you recommended, that was already on my radar.
 
I read something this morning (wish I kept the link) that stated that if the power is brought into the home, that the code would apply. My apologies if this was already covered.

As stated earlier, the Morningstar Classic seems to cover both ground and arc fault. Challenge is that I moved in the Victron direction. I saw the ground fault units for Morningstar. I saw one source that claimed a combiner box that had arc fault included. Two challenges was that it listed at $ 1900 and was out of stock. There has to be something out there to meet this need.
 
400bird said:
I'd love to include AFD in my plans for a DC coupled array.

My issue is cost. I haven't found a reasonably priced stand alone arc fault detector. I would probably wire it into the RSD as you recommended, that was already on my radar.
Click to expand...

Plans? If you don't have SCC yet, Midnight Classic has that. Not cheap, but reportedly very good. Probably costs less that whatever stand-alone Arc-fault are still offered. Also tolerates excursions above "maximum" PV input voltage so strings don't have to be shorter for extreme cold.

MidNite Solar Inc. - THE MIDNITE CLASSIC:<BR/>NORTH AMERICA'S MOST SOPHISTICATED MPPT CHARGE CONTROLLER

MidNite Solar is the industry leader and manufacture of quality Renewable Energy System electrical components and E-Panels.
www.midnitesolar.com www.midnitesolar.com


JJJJ said:
I read something this morning (wish I kept the link) that stated that if the power is brought into the home, that the code would apply. My apologies if this was already covered.
Click to expand...

I think the RSD requirements are written that way. But I doubt segmenting PV string is actually necessary for array on a remote structure, just disconnecting from the house so no voltage present inside. Exceptions for circuits inside metal conduit might apply too. For the goal of not shocking a fireman and not burning down the house, keeping PV inside rigid conduit until it reaches SCC is probably sufficient. Although, arc inside conduit could eventually burn through it.

A disconnect switch on outside wall should go a long ways.
 
I am considering blending the Morningstar scc with the Multiplus. Any reason that would not work?
 
Still frustrated by this (been a while since I've posted since I just stopped all activity on this to ... regroup). Now ready to spend whatever it takes, so long as it's stand-alone (not buying new inverters), but still can't find anything. I had found something out of China, and thought I had it bookmarked, but now can't find it. In the mean-time, while re-researching, I ran across this thread on Mike Holtz forum. If I'm reading it correctly, it says that GFCI handles any parallel faults, while the series arcs are handled in the more expensive inverters: https://forums.mikeholt.com/threads...ircuit-proection-dc-strings-3-modules.142692/
 
You must log in or register to reply here.

Similar threads

D
6000xp - arc and ground fault protection suppprt
Replies
3
Views
310
Daniel Beardsley
Daniel Beardsley
V
Multiple charge controllers and ground fault protection
Replies
0
Views
246
vtMaps
V
S
6000xp W002: AFCI high Notice
Replies
6
Views
205
SteamEngineer
S
G
ABB Inverter Arc Fault Bypass?
Replies
10
Views
744
Gwhite
G
P
Agrivoltaics gone Haywire
Replies
5
Views
317
PresterJohn
P

diy solar

diy solar
Back
Top