And now MIT adds their two cents to the AFCI debacle

Status
Not open for further replies.
hbiss said:
Damaged romex can be solved through a number of simple solutions, from requiring insulated staples that have a plastic liner to the outright ban of romex. Wire residential with MC and steel boxes, cheaper than a panel full if GFCIs.
Click to expand...


Ingenieur said:
a gfci is more for human protection than a hi z conductor fault
Click to expand...

Meant to say AFCIs. :slaphead:

-Hal
 
mbrooke said:
There is the hardware and software which correctly does it, with 100% accuracy. In addition to copious evidence showing interruption of dangerous faults and none tripping on benign signatures.


https://cdn.selinc.com/assets/Liter...lyers/Arc-Sense_PF00160.pdf?v=20161031-073656
Click to expand...

Fair enough, but this is for high voltage distribution. It is pretty rare to plug a vacuum cleaner with a universal motor into a 7200V line directly :)

I would also dispute the 100% number as being physically impossible, but for the limited application of distribution protection I can see approaching it. You have fewer lines to sense, you can spend more on the protection device, and you probably have fewer 'normal operations that look like arcs' to contend with.

In a residential application you want an inexpensive device to reliably differentiate between arcs that indicate faults and loads which may have internal arcing.

-Jon
 
winnie said:
Fair enough, but this is for high voltage distribution. It is pretty rare to plug a vacuum cleaner with a universal motor into a 7200V line directly :)
Click to expand...

Voltage is irrelevant. The computer is looking at the secondary CT outputs, or more precisely current and its waveform.



I would also dispute the 100% number as being physically impossible, but for the limited application of distribution protection I can see approaching it.
Click to expand...

I wouldn't say its that limited- though granted you do have a point based on how you look at it in terms of the number of devices vs standard protection relays as a whole. However, there is extremely intensive logic and extensive research into creating it spanning several decades. These devices are not toys, and manufacturers like SEL make relays knowing that not only is the whole world watching, but also the continuity of a nation depends on them. No POCO will tolerate nuisance tripping or a device that will not make good on its promise- especially when the waveform of each trip can be brought up and analyzed. In fact when an inadvertent trip takes place the first thing thats called into question is the field programming (settings) of the device rather than the logic itself built into the device.

You got me on the 100%, it might be possible (impossible) from a probability and physics standpoint now that I think about- however still take my word on the accuracy claim. ;)


You have fewer lines to sense, you can spend more on the protection device, and you probably have fewer 'normal operations that look like arcs' to contend with.
Click to expand...


Technically you could get away with one device in residential. CTs from each 120 volt circuit feeding into a central processor. The processor then decides which breaker to trip- similar to the Bender insulation and power quality monitoring scheme. That was one of the original ideas when AFCIs were being developed.







In a residential application you want an inexpensive device to reliably differentiate between arcs that indicate faults and loads which may have internal arcing.

-Jon
Click to expand...

But there in lies the issue. Inexpensive and reliable just don't go together at this point in time. You need computing power, and lots of it.
 
Last edited:
gadfly56 said:
Yeeeeeeeah, I don't think that's going to fly for your average residential application. :roll:
Click to expand...


If you could cram that type of processing power backed up by quality research and development into a 1 inch breaker, then it would fly high. The only issue being that breakers would start at $250. A better idea would be a central processor like the Bender approach, but still, serious AFCIs require serious $$$$.



K8MHZ said:
Especially with the re-closers and all....
Click to expand...


Reclosers and feeder breakers "unlatch" through a solenoid- as do AFCI breakers. Its either on or off output logic. CTs for current sensing, as does a resi AFCI.
 
mbrooke said:
...
Technically you could get away with one device in residential. CTs from each 120 volt circuit feeding into a central processor. The processor then decides which breaker to trip- similar to the Bender insulation and power quality monitoring scheme. That was one of the original ideas when AFCIs were being developed...
Click to expand...

I defy you to describe how any processor of any sort can look at the aggregate wave form at the service entrance and decide which of 40 breakers needs to be tripped. Did you mean to say a central processor with 40 CT's (for example)?
 
Adamjamma said:
Ran into a product over here in Britain yesterday for installation in certain circumstances called a firefly that has me wondering from my own experience in the USA wether it might be usable to stop some of the problems, but I may not be thinking straight. We use staples to hold the wire to the studs so it does not flop around inside the wall cavity and also to help reduce possible strain on the wire, if I recall what I was taught in shop class in 1979.
So, these firefly clamps, while designed for inside the surface mount systems to keep the wire from pulling off the wall in a fire, would also be able to hold the wire to studs, as the wire would simply slide under the clip, if I am thinking right.
because you would apply the clips before the wire, there should be no damage to the Romex, thus no more overstapling.
But, it would not be code right now, so no good, unless I am missing something??
however, I think the inspector in Jamaica would use my logic, for the few places we would use it in Jamaica since most buildings are not built with studs...lol
Click to expand...

How about gorilla tape, or even better have an adhesive side to NM:D... darn that wouldn't be much fun pulling through studs:(.
 
The difference between utility distribution and residential wiring is not simply voltage.

As you note you are measuring _current_ and comparing current waveforms.

The issue is the difference between normal current waveforms and fault condition waveforms.

If the range of normal waveforms overlaps with the range of fault condition waveforms, then it doesn't matter how much computation you throw at the problem; you will have some situations where you cannot differentiate between normal fault operation.

I _suspect_ but do not know that the range of normal current waveforms at the distribution level is generally much smaller than the range of normal current waveforms in residential wiring. At the distribution level you have lots of loads averaged together, with transformers as well. It seems to me that if you see an arc signature at the distribution level you are pretty sure to have an arcing fault. In residential wiring an arc signature could be _normal_ operation, so then you need to differentiate between different arc signatures.

-Jon
 
winnie said:
The difference between utility distribution and residential wiring is not simply voltage.

As you note you are measuring _current_ and comparing current waveforms.

The issue is the difference between normal current waveforms and fault condition waveforms.

If the range of normal waveforms overlaps with the range of fault condition waveforms, then it doesn't matter how much computation you throw at the problem; you will have some situations where you cannot differentiate between normal fault operation.

I _suspect_ but do not know that the range of normal current waveforms at the distribution level is generally much smaller than the range of normal current waveforms in residential wiring. At the distribution level you have lots of loads averaged together, with transformers as well. It seems to me that if you see an arc signature at the distribution level you are pretty sure to have an arcing fault. In residential wiring an arc signature could be _normal_ operation, so then you need to differentiate between different arc signatures.

-Jon
Click to expand...
and how many possible arc signatures are there to keep track of and be monitoring for? 1, 10, 100, 1,000,000...?
 
gadfly56 said:
I defy you to describe how any processor of any sort can look at the aggregate wave form at the service entrance and decide which of 40 breakers needs to be tripped. Did you mean to say a central processor with 40 CT's (for example)?
Click to expand...

:roll: You guys aren't reading my posts in full :p


Thats why I mentioned Bender.


schema_rcms150.jpg



linetraxx.png



All 40 circuits have their own CT. Data from each individual CT goes to a central processor. That processor would then decided what breaker to trip.
 
winnie said:
I _suspect_ but do not know that the range of normal current waveforms at the distribution level is generally much smaller than the range of normal current waveforms in residential wiring. At the distribution level you have lots of loads averaged together, with transformers as well. It seems to me that if you see an arc signature at the distribution level you are pretty sure to have an arcing fault. In residential wiring an arc signature could be _normal_ operation, so then you need to differentiate between different arc signatures.

-Jon
Click to expand...




For the sake of the argument lets say that was entirely true for medium voltage distribution- arc signature = arc fault; perfect sine wave = no fault.


Now what happens when 1 of 12 feeders from a substation develops and arc fault? What happens when circuit 1 is mutually (capacitance) coupled to circuit 2 from sharing the same pole or being in the same right-of-way for several miles and circuit 2 develops an arc fault? What happens when the normal source 69kv sub-transmission supply to the distribution substation develops an arc fault?


In all these 3 scenarios CTs will detect an arcing signatures through out the MV system. As such it would be and is necessary to have intelligent wave form discrimination even in systems where anything but a pure sine wave indicates a problem.


As this pertains to residential, how many electricians have encountered AFCI trips caused by appliances on adjacent circuits? How many have found arc fault breakers to trip from external events caused by the utility? Its the same issue at hand, the technology is primitive and lacks intelligence to discriminate. As I said, utility AFCIs took time and mind numbing research for this reason.
 
So i still see toroidals , but now there's a better 'brain' aka-central processor

Things get digital ,and i'm lost :(

But i'll wager this processor isn't the size of a pea.....yet....

~RJ~
 
romex jockey said:
So i still see toroidals , but now there's a better 'brain' aka-central processor

Things get digital ,and i'm lost :(

But i'll wager this processor isn't the size of a pea.....yet....

~RJ~
Click to expand...

Well, the brain is simply a device between the toroid and trip solenoid...
 
mbrooke said:
As this pertains to residential, how many electricians have encountered AFCI trips caused by appliances on adjacent circuits? How many have found arc fault breakers to trip from external events caused by the utility? Its the same issue at hand, the technology is primitive and lacks intelligence to discriminate. As I said, utility AFCIs took time and mind numbing research for this reason.
Click to expand...

As I said, I like the _theory_ of AFCIs. What you seem to be saying is that utility distribution Arc Fault detection has been made reliable, and that this implies that residential Arc Fault detection could be made reliable with sufficient computational power and research.

As it pertains to residential, the question is: were residential AFCIs deployed before they were 'ready for prime time'? And if more expensive hardware is required to make AFCIs acceptably reliable, would spending that money on other forms of protection be better in terms of lives saved per dollar?

-Jon
 
winnie said:
As I said, I like the _theory_ of AFCIs. What you seem to be saying is that utility distribution Arc Fault detection has been made reliable, and that this implies that residential Arc Fault detection could be made reliable with sufficient computational power and research.
Click to expand...


Correct- I say yes to both.

However, the question which then has to be answered: is arcing even a concern at 120 volts? Sure we can partially sever cable in a lab and carbonize the gash to the point it will conduct 120 volts- but does this actually happen in the real world?


As it pertains to residential, the question is: were residential AFCIs deployed before they were 'ready for prime time'? And if more expensive hardware is required to make AFCIs acceptably reliable, would spending that money on other forms of protection be better in terms of lives saved per dollar?

-Jon
Click to expand...



Remember that when AFCIs were introduced into the code they weren't even on the market yet. Fast forward 20 years they still aren't ready for prime time.


The money electricians and contractors spend on AFCIs could go toward a technology thats far more reliable, provides far more coverage then signature analysis alone, and fully matured: GFP. Thats what the rest of the world did with their 230 volt supplies, and what the NEC did with 277/480 volt services over 999amps.

In terms of what provides most life savings for the buck, that would hands down be fire sprinklers. Fire sprinklers will do what no electrical safety technology will ever do- and pretty much what no other life saving technology could do either.
 
Status
Not open for further replies.
Top