How AFCI's Work

Status
Not open for further replies.
mshields

mshields

Senior Member
Location
Boston, MA
AFCI's detect 30mA of arcing current and are designed to protect faults between phase and neutral caused by compromised insulation. i.e. either frayed insulation or insulation damaged by a staple gun etc. Right?

But how do they work. e.g. they're not supposed to trip on a switching arc or when a receptacle is pulled out of the wall. How do they know the difference? Is it just a duration thing?


Mike
 
D

dnem

Senior Member
Location
Ohio
You may likely get an answer from someone in the industry who can really do this question justice. . Alot of people of various backgrounds post here. . I remember the first explanation I recieved of AFCI operation was that it operated by smoke and mirrors.

I believe AFCIs operate by monitoring rate of current rise. . The voltage is supplied by the utility at a very specific sine wave pattern. . You get 60 cycles a second, 120 half cycles a second, which means your first voltage peek hits at 1/240 of a second. . I believe the AFCI looks at the first half of that curve to the first peek, which would be the first 1/480 of a second. . While the voltage has a set curve from the utility, the current can spike quickly when there's a arc. . A circuit that contains an arc doesn't have a normal rate of rise within that first 1/480 of a second. . The sine wave curve of the current is different and can be detected by the AFCI.

Hopefully someone from the manufacturing end will jump in and explain it. . What exact difference it's seeing in the the rate of rise.

David
 
W

winnie

Senior Member
Location
Springfield, MA, USA
Occupation
Electric motor research
You guys are describing how AFCIs were _supposed_ to work when the concept of requiring them was sold to the CMP. AFCIs were supposed to detect series arcs, meaning arcing connections in the circuit where the current is limited by the load.

I do not know the details of how arcing current is different from normal current flow; it simply is (somewhat) different, and with suitable signal processing you can detect the difference from normal current flow. However the problem of _reliably_ detecting this difference in a mass produced product with minimal error has not been solved. AFCIs simply cannot reliably detect series arcs and differentiate them from normal loads with intentional arcing (eg motors, switches, etc).

Instead they look for 'parallel arcs', in other words, arcing short circuits. Here the current is not limited by the load, and is only limited by circuit impedance. Parallel arcs _are_ a significant danger, and on this point AFCIs are a real improvement, because they can detect parallel arcing conditions that will not trip a normal breaker. We are talking instantaneous current levels that are well in excess of the nominal trip rating of the breaker, but which might have average current levels low enough that the breaker would take quite a while to trip. If the breaker sees current that exceeds some rather high threshold, _and_ it looks like an arc, then it opens.

Additionally, AFCIs have built in ground fault detection. This is not GFCI grade 5mA detection, but instead 50-70mA 'ground fault protection for equipment'. The idea is that any arc or fault to ground will cause the breaker to open, even with current flow far to low to cause an overcurrent condition.

Net result:
no detection of series arcs; no protection from 'glowing splices'
enhanced detection of parallel arcs; once things have failed and you do have a short circuit you get improved response
enhanced detection of short circuits to bonded metal; very low ground fault current will open the breaker.

See:
http://www.mikeholt.com/mojonewsarchive/AFCI-HTML/HTML/AFCITesterStory~20020826.htm

http://www.mikeholt.com/mojonewsarchive/AFCI-HTML/HTML/Arc_Fault_Protection~20020124.htm

-Jon
 
N

NoVA Comms Power

Senior Member
Location
Alexandria, VA
Last edited:
C

coulter

Senior Member
winnie said:
... AFCIs simply cannot reliably detect series arcs and differentiate them from normal loads with intentional arcing (eg motors, switches, etc). ...
Click to expand...
Jon -

Thank You. Far better explanation than I got a couple of years ago.

carl
 
D

dnem

Senior Member
Location
Ohio
In my post I expressed my understanding that the voltage is a set function of the frequency set by the utility, but the link from NoVA Comms Power shows that the voltage sine wave gets screwed up in an arc more than the current sine wave. . That was surprising to me. . I guess it shouldn't have surprised me because inductor coils and capacitors cause leading and lagging between the voltage and current also.

I'm also surprised that there's no smoke and teeny tiny mirrors inside the AFCI. . I really thought there was some. :rolleyes:

David
 
N

NoVA Comms Power

Senior Member
Location
Alexandria, VA
dnem said:
... I'm also surprised that there's no smoke and teeny tiny mirrors inside the AFCI ...
Click to expand...
LOL!

TI's DLP chip has already grabbed the "teeny tiny mirror" patent! :smile:

While the AFCI detection theory is indeed interesting, I'm still somewhat skeptical that there can be a reliable "one size fits all" detection device given the wide variety of loads out there.
 
Last edited:
mshields

mshields

Senior Member
Location
Boston, MA
voltage waveform

voltage waveform

that's interesting info on the voltage waveform. As for the reliability of these devices and their ability to differentiate between a fault and all other non fault arc's, it does make one wonder. I know that with the GE version of a branch/feeder device, they actually look at 8 sample waveforms to make sure the device is looking at a fault and not just a power tool or a switch or an incandescent lamp etc. I don't know if it also looks at the voltage waveform. Would be interesting to know.

Anyway, whether they work or not, they're comming in a big way in the 2008 NEC.

Mike
 
mshields

mshields

Senior Member
Location
Boston, MA
Zlan website

Zlan website

That's an interesting site. Thanks for the tip. Did these guys develope technology that one or more of the major breaker vendors utilize?
 
Status
Not open for further replies.
Top