ECN Forum Forums Code Related Discussion NEC & other Code issues Pashchen's Law

Anyone who has dine much welding know striking and maintaining the arc is the tricky part until you get a feel for it but bad luck seems to have a way of accomplishing hard things.
I imagine the real reason natural AC arcs self extinguish so fast is that they can't advance the electrode at the speed it is being burned away.
When we want to put the fire out on an AC contact, we will bridge that gap with a resistor/capacitor network. I understand that makes the "zero crossing" a little longer by slowing down the rise, giving the arc a chance to go out.
I know the voltages involved are far beyond the circuit voltage if you have an inductive load.

For ease of the readership, here is what Dr Engle published in his paper (i can't seem to post the graph, sorry)

~S~


Test 12: Operation Inhibit Apparently UL determined that the arc detection schemes of both Products 1 and 2 utilized the high frequency content of the current waveform.

A continuous low current series arc was created using the “arc simulators”. These simulators use
two opposing electrodes, like those of a carbon arc lamp.


Rather than use two carbon electrodes, Simulator 1 used one carbon (graphite) electrode and one phosphor-bronze electrode.

Simulator 2 used one carbon and one copper
electrode. Both simulators created a continuous low current arcing condition. The current, after each zero crossing, is zero for a few milliseconds until an arc re-strike occurs.

This re-strike transient produces a high frequency “noise” component in the arcing current waveshape.

UL evidently theorized that Products 1 and 2 series arc detection algorithms could be masked by normal continuous series arcing, such as from the brushes of an electric drill.

Also a normal EMI filter used in power strips etc. could filter the high frequency current component, so the AFCI circuit breaker would not see the arcing event.

UL was correct in their assumptions. Products 1 and 2 were tested and failed both masking tests that indicate their technology was based on looking for high frequency noise.


Why didn’t UL use copper-copper electrodes, instead of
the odd combinations of carbon and phosphor-bronze and
carbon-copper?

Unfortunately UL didn’t address this important question, and thus the validity of the use of these “arc-simulators” is questionable.

Further the author believes that UL, by
introducing their use, inadvertently gave credibility to AFCI manufacturers’ claim that their product will respond to a series arcing event.

The use of strange materials like phosphor-bronze and carbon, conductive materials not used in house wiring, might be explained but not justified, as follows:

 Copper-Copper: This combination and copper-steel are the only valid electrode choices. If UL wanted to demonstrate “real world” series arc detection they could have used copper.

UL may have tried copper but found they only got
sparks, not the continuous low current arcing that Products 1 and 2 probably needed to trip.

The reason for a simple spark is explained by a century old law of physics. A person named F. Pashchen in 1889 published a law which sets out what has become known as Paschen's Law.

He determined the relationship between breakdown voltage, the gap between two metal plates, and the pressure.


With air as the gas, the minimum voltage is 327V, as shown in Fig. 5. The peak of a 120VAC sine wave is only 170V, and thus continuous low current arcing is, by a law of physics, not possible with copper-copper.


Thus claims that a Combination AFCI will respond to arcing at a break in a
conductor or a loose connection flies in the face of a law of physics.


Fig. 5 Paschen's Law


Copper and Phosphor-bronze and Carbon-
Copper: UL did not justify the use of such strange electrode pairs.

Practically, the use of carbon made it easy to produce
continuous low current series arcing to test the claims of Product 1 and 2.

Paschen's Law applies only to metal-to-
metal arcing. Unfortunately, it also would seem to invalidate the use of any test results as part of a home electrical fire study.

I agree the choice of electrodes is questionable. If they are really testing this device they should use zip cord conductors where the fine strands have a better chance of rising to the ignition temperature of the dust bunnies under the bed (the original justification for these things).

I do not have a dog in this fight, I am just playing the devil's advocate in the arguments.

Use of carbon arc rods for arc testing. Back in the late '80s, I used to operate carbon arc spotlights for a theater. The cabon rods were plated with copper and when they operated they cast a greenish colored light. The presence of the green light indicates to me that copper was vaporizing in the arc.

True, this was not a 100% copper electrode, but does this rule out that the electrodes the UL used were wrong?

The phosphor bronze / carbon & copper(?) combination sounds alot like the same materials used in the push and pray connections on the back of cheap switches and outlets.

The addition of carbon MIGHT be justified by assuming the presence of carbon from previous arcing events.

Larry the troublemaker.

Larry, the composition of the electrodes is what I thought of as the 'next step' in our peer review of the article.

After all, if his claim regarding the 'laws of physics' is incorrect, the rest of the paper is seriously weakened.

If he's correct - that it's just not possible for a sustained arc to form between copper electrodes at household voltages - then no manner of AFCI is possible.

Dissecting the test method only counts if the theory itself is sound.

OK ... let's see if this link works:

http://printfu.org/read/electrical-...Ll2ubbyuOhrrG5mKihlLm0w7jTrbi0oNnS2oiw6g

It should take you to a paper called "Electrical Breakdown Limits for MEMS," at a site called Printfu . "MEMS" stands for 'Micro Electro-Mechanical Systems." The paper seems to be part of an electrical engineering course. Engineers: Were you sleeping through this class?

The paper opens with the graph that Mr. Engle SHOULD have had in his article.

It seems that "Paschen's Law" is at the root of some of the problems in developing micro-technology. Simply put, there aren't enough ait molecules to let an arc form at distances much less than a human hair's thickness.

OK ... so the reference to Paschen's law was relevant, and correct. Where does that leave us?

It seems to leave us with the strange situation that a sustained electrical arc - as opposed to the occasional spark - cannot cannot form between two ROUNDED COPPER electrodes in ordinary air at household voltages. We can discount the high impulse voltage that is created when the circuit opens since this is not a sustained voltage. Sparks, yes; arc, no.

For an arc to form, something else must happen. Jagged contacts. Materials other than just plain copper. An atmosphere besides air.

Oddly enough, this conforms to a sample I had, where a nicked wire was able to very neatly arc-cut through EMT for a length of about 1/8".

That still leaves us with any 'arc detecting' device being dependent on something else also happening. Sounds pretty iffy to me.

So, I'll grant Mr. Engle his point. Arc detection is pretty much a solution in search of a problem.

NEC Panels: Time to dump the AFCI. Period. Completely. It's pure snake oil. Absence of any actual arc testing in the UL standard ought to make that plain. (Perhaps that's why the standard is priced at approximately it's weight in gold).

Where do we go from here? Well, now it's the 'glowing contact.' I'm not so sure about Mr. Engle's opinions there, and that's a topic for another thread.

While Dr Engle could have expanded on Pashen's law, vaious atmospheres & metals, etc, the fundamental concept of a 'sustained arc' vs. a 'spark'remains available.



http://en.wikipedia.org/wiki/Spark_gap

A spark gap consists of an arrangement of two conducting electrodes separated by a gap usually filled with a gas such as air, designed to allow an electric spark to pass between the conductors. When the voltage difference between the conductors exceeds the gap's breakdown voltage, a spark forms, ionizing the gas and drastically reducing its electrical resistance. An electric current then flows until the path of ionized gas is broken or the current reduces below a minimum value called the 'holding current'. This usually happens when the voltage drops, but in some cases occurs when the heated gas rises, stretching out and then breaking the filament of ionized gas


We're all familiar with Jacob's ladder, note the volage values>>

http://www.repairfaq.org/sam/jacobs.htm#jlhdj


Perhaps even more everyday is our trade being in the presence of a 'spark' , i.e.- old switches , small to large disconnects

i would venture commutator motors provide a series of sparks, instead of a sustianed arc , even though we all can note the nusiance trips from them

in any version of Pashchen's law , we do not see a sustained arc event @ nominal household voltages


part of the problem Dr Engle laments is the definitional difference bettwen these phenomenon >>


The proposed changes to UL 1699 should have been
acceptable to all STP members. It involved no core changes;

it simply attempted to bring the Standard’s language in line with other UL standards. In particular, while the term “arcing” is defined in UL 1699, the UL terms “series arcing” and “parallel
arcing” are not defined.

Yet the manufacturers and UL on their
web sites claim that the mandated Combination, unlike the Branch/feeder, provides “series arcing” protection of cords.





Further, if UL had not taken it upon themselves to create a standard, instead of testing something to a standard, we wouldn't be discussing this issue at all>>>>




As will be explained in the section discussing UL1699, the fire curve later became the UL1699 40.4 Carbonized path arc clearing time test. This test is the only arc performance test difference between the Branch/feeder and the Combination
AFCI.
After accepting the test, the Task Force admitted its failure to develop a Standard, and turned the task over to UL.
This is not a normal UL responsibility; manufacturers develop Standards, and UL is paid to test and list products to these Standards.






The unfortunate result being>>>>




“Carbonized path arc clearing time test”
This test was discussed earlier, it makes no sense. It was thought that if the technical issues were honestly presented and discussed at this STP meeting, the test would be removed from UL 1699.

Once removed, the Branch/feeder and Combination requirements would be the same (see Fig. 16),
so there would be only a Branch/feeder AFCI. The mandate of NEC 2005 would be moot, as there would be no Combination AFCI.

The author was permitted to speak to the group for about half an hour.

The test was carefully described including the
important fact that it represented nothing more than a carefully prepared parallel fault in a VERY long extension cord.

Available short-circuit fault currents as low as 5A were used, even though UL had earlier determined that the lowest available current in a home was 75A (see Fig. 2).
It was not a series arcing fault and had nothing to do with home electrical safety.

The vote was along “party lines”. NEMA manufacturers, with a few exceptions, and UL voted against removing test.

This block of votes exceeded the 1/3 required to defeat theproposal, so it failed. The “Carbonized path arc clearing time test” would remain in UL 1699.


~S~

I think the whole AFCI thing is snake oil but if it is really a valid fix for something, we should establish what it fixes and demonstrate that this technology does the job.
The step that they all skipped was creating a real world arc that starts a real fire. Then detect that.
They have created a straw man and knocked it down.

Greg

my opinion is that if a bluecollar redneck such as i can be educated to the specific physics involved here, we all can

it's just a metter of perseverance and diligence amonst those who truly care about the trade

~S~

After review of most (not all) of the information within the two threads regarding AFCI, I made a choice on the further use of the video info I use in my classes.

Within the NEC classes, until the powers that be remove the combo designated requirements, I have to reference the Code requirements. However, I will add, as my commentary, that this technology may be flawed, and reference back to some of the subject matter in this thread.

Within my Basic class, some of the subject matter that has been included within this thread will be offered, along with the fact that this is still a NEC requirement.

That said, as AHJ, I have no recourse at the present time but to require these devices.