Solid State Electronic Load Tripping AFCI

Status
Not open for further replies.
J

jpstaub

Member
Location
Portland, Maine, USA
To verify the operation of an AFCI breaker protected circuit a commodity solid state power tool battery charger with no battery installed was plugged in to the protected circuit; the battery charger was chosen for no specific reason other than it was an electrical load with a power indication LED that was close at hand. After a short delay with power indication LED on (approximately 1 second) the AFCI tripped. The troubleshooting effort was then keyed up:

1. Load removed. AFCI breaker self tested multiple times with good response. Still trips under battery charger load.

2. AFCI breaker swapped and self tested multiple times with good response. Still trips under battery charger load.

3. Circuit stripped down to a single outlet. 12/2 wiring inspected closely (no bare wire in questionable places, all mechanical connections checked tight, how precise does this whole assembly need to be it's not a rocket ship!?). Breaker self tested multiple times with good response. Still trips under battery charger load.

4. AFCI breaker swapped out with a simple breaker. Simple breaker holds battery charger load.

My primary question is:
Is there a standard that governs the way in which solid state devices transition from their powered off to their powered on state?

The follow on question:
If there is a standard that governs solid state device power up do AFCI breaker manufacturers design to this standard to prevent false positive indications?
 
jpstaub said:
To verify the operation of an AFCI breaker protected circuit a commodity solid state power tool battery charger with no battery installed was plugged in to the protected circuit; the battery charger was chosen for no specific reason other than it was an electrical load with a power indication LED that was close at hand. After a short delay with power indication LED on (approximately 1 second) the AFCI tripped. The troubleshooting effort was then keyed up:

1. Load removed. AFCI breaker self tested multiple times with good response. Still trips under battery charger load.

2. AFCI breaker swapped and self tested multiple times with good response. Still trips under battery charger load.

3. Circuit stripped down to a single outlet. 12/2 wiring inspected closely (no bare wire in questionable places, all mechanical connections checked tight, how precise does this whole assembly need to be it's not a rocket ship!?). Breaker self tested multiple times with good response. Still trips under battery charger load.

4. AFCI breaker swapped out with a simple breaker. Simple breaker holds battery charger load.

My primary question is:
Is there a standard that governs the way in which solid state devices transition from their powered off to their powered on state?

The follow on question:
If there is a standard that governs solid state device power up do AFCI breaker manufacturers design to this standard to prevent false positive indications?
Click to expand...
I think the problem is that AFCI's are programmed to recognize acceptable current waveforms, if you have a load that is not recognized it will be rejected and will trip the device. I may be only partly right or I could be totally wrong though.
 
jpstaub said:
My primary question is:
Is there a standard that governs the way in which solid state devices transition from their powered off to their powered on state?
Click to expand...
I don't know but I would think not.

The follow on question:
If there is a standard that governs solid state device power up do AFCI breaker manufacturers design to this standard to prevent false positive indications?
Click to expand...
I am sure they try. The way I understand it the AFCI manufactures have a UL standard that they must meet for detecting an arc but they can achieve it any way they want.

One thing I know for sure is I have seen AFCI breakers, the newest "combo style" to be precise, behave exactly as you described with electronic loads.
 
The arc detection is not a simple case of everything that is not permitted will cause a trip.
Instead the signature filters look for specific voltage and current profiles characteristic of arcs. But the analysis may also pick up on other similar patterns with normal sources. So the designers then tweak the filters to allow those similar signatures. When they get tripping on a new "normal" load that they have not seen before they do more tweaking.
With average luck, this will be a never ending process.

The requirement to avoid nuisance trips comes not from UL but from market forces. And with mandated use the only market forces are between brands.
 
I suspect it is the gfp in the afci breaker that is tripping. As I understand it the afci does not work unless there is 3 or 4 amps of current on the circuit.
 
Dennis Alwon said:
I suspect it is the gfp in the afci breaker that is tripping. As I understand it the afci does not work unless there is 3 or 4 amps of current on the circuit.
Click to expand...
In that case, it could well be ground referenced filter capacitors inside the power supply. These can, but should not by design, cause a GTI to sense imbalance. But the current would have to be in the 30ma or greater range.
 
GoldDigger said:
In that case, it could well be ground referenced filter capacitors inside the power supply. These can, but should not by design, cause a GTI to sense imbalance. But the current would have to be in the 30ma or greater range.
Click to expand...
Dennis said "gfp", and originally the AFCI-only breakers came with GFP whether they said so or not, so yes, it would have been at best 30mA (maybe 100?). But the new combo breakers are AFCI / GFCI, meaning Class A, so they would be 4-6mA trip, not 30mA, so it still might be that. Hard to know without the specific brand and model number. But those new combo breakers, at least the ones I have seen, have two separate indicators to tell you if it has tripped on AF or GF. I would think the OP would have differentiated if that were the case.

As far as I know, there is no "standard" for how electronic power supplies power up or even operate, at least not yet. There is some regulation coming our way (in the EU now) regarding them having power factor correction techniques and harmonic limits, and some stuff that is marketed word-wide is already arriving here with those features even though we don't yet need them, but nothing regarding how they affect AFCIs (or GFCIs). I'm not sure if those issues might have an effect, but nuisance tripping from electronics has been an issue since the day AFCIs were introduced. As Golddigger said, the issue has to do with the internal software algorithm that is looking at the current waveform for anomalies. If it sees something that its filters have never seen before, it trips to be safe, not sorry. I too think it's something we are going to live with for a while. When I was at Siemens in 2006 as they were getting ready to release their combo breakers, they had a team of a couple of dozen engineers constantly working on tweaking the filtering algorithms, because every time they tested something new, it behaved different. I remember talking to one who was testing nothing but vacuum cleaners, because those were, for some reason, notorious at that time (he suspected static build-up, and the differences in the way the mfrs dealt with it). He said it was a nightmare, because as soon as he came up with a solution for one, it re-introduced a problem he had already fixed on another one.
 
Jraef said:
Dennis said "gfp", and originally the AFCI-only breakers came with GFP whether they said so or not, so yes, it would have been at best 30mA (maybe 100?).
But the new combo breakers are AFCI / GFCI, meaning Class A, so they would be 4-6mA trip, not 30mA, so it still might be that. Hard to know without the specific brand and model number. But those new combo breakers, at least the ones I have seen, have two separate indicators to tell you if it has tripped on AF or GF. I would think the OP would have differentiated if that were the case.
.....
Click to expand...
For the sake a clarity and common electrical language those are called Dual Function or some thing close. Combo AFCI breakers are made to detect series and parallel arcs and GE brand at least does not have any GF detection in them at all.
 
GoldDigger said:
The arc detection is not a simple case of everything that is not permitted will cause a trip.
Instead the signature filters look for specific voltage and current profiles characteristic of arcs. But the analysis may also pick up on other similar patterns with normal sources. So the designers then tweak the filters to allow those similar signatures. When they get tripping on a new "normal" load that they have not seen before they do more tweaking.
With average luck, this will be a never ending process.

The requirement to avoid nuisance trips comes not from UL but from market forces. And with mandated use the only market forces are between brands.
Click to expand...
I was wrong but still kind of on the right track. The main thing is there are millions of possible load conditions the device may possibly encounter, it is programmed to accept or reject only so many of them and the rest are a crap shoot. That is one of the biggest problems I have with having AFCI use mandated in the code.
 
kwired said:
I was wrong but still kind of on the right track. The main thing is there are millions of possible load conditions the device may possibly encounter, it is programmed to accept or reject only so many of them and the rest are a crap shoot. That is one of the biggest problems I have with having AFCI use mandated in the code.
Click to expand...

Brings to mind another crap shoot twist. Product A and B both work with an AFCI, even together. Now add Product C, and we start tripping, but C works fine by itself or with either A or B, but not all three. :happysad:
 
GoldDigger said:
The arc detection is not a simple case of everything that is not permitted will cause a trip.
Instead the signature filters look for specific voltage and current profiles characteristic of arcs. But the analysis may also pick up on other similar patterns with normal sources. So the designers then tweak the filters to allow those similar signatures. When they get tripping on a new "normal" load that they have not seen before they do more tweaking.
With average luck, this will be a never ending process.

The requirement to avoid nuisance trips comes not from UL but from market forces. And with mandated use the only market forces are between brands.
Click to expand...

First off, thanks for the replies. They were all informative.

To an outsider looking in, AFCI protection appears attractive in theory. In practice, AFCI protection seems relatively worthless given its current and projected unreliability when exposed to even basic actual electrical environments. Turning the general population against electricians is probably not the best thing to do in the interest of long term electrical safety. After all, the electrician is typically the name/face associated with what's behind the panel door.

It is surprising that the NEC mandates what seem to be experimental circuit protection devices that do nothing to enhance the document's reputation when exposed to conditions beyond the experience of firmware programmers. Putting the NEC reputation in the hands of private industry firmware programmers was not a wise decision.
 
Dennis Alwon said:
I suspect it is the gfp in the afci breaker that is tripping. As I understand it the afci does not work unless there is 3 or 4 amps of current on the circuit.
Click to expand...
It is my understanding the the AFCI does not look for a series arc unless the current exceeds 5 amps, and does not look for a parallel arc unless the current exceeds 75 amps.
 
GoldDigger said:
Was that supposed to be 7.5A?
Click to expand...
I think 75 was correct.

That number is high enough that people will ask why the 15 or 20 amp breaker wouldn't be enough protection alone - but I think the idea here is it will recognize the arc signature and trip even if the duration was not long enough for the normal trip curve of the breaker to provide protection. How well does it actually work IDK, but I think that is the concept.
 
kwired said:
I think 75 was correct.

That number is high enough that people will ask why the 15 or 20 amp breaker wouldn't be enough protection alone - but I think the idea here is it will recognize the arc signature and trip even if the duration was not long enough for the normal trip curve of the breaker to provide protection. How well does it actually work IDK, but I think that is the concept.
Click to expand...
Yes, 75 was correct...one of the issues is that longer residential circuits many not even be able to supply that amount of current into the fault.
As I recall the AFCI needs to see the arc signature for at least 4 half cycles, so yes quicker than the instantaneous trip time for a breaker, even if the instantaneous trip was set that low.

There have been some suggestions to require a lower IT trip current (I don't the UL breaker standard even requires an IT trip), but a lower IT trip point many cause trips with inrush on motor loads.
 
Status
Not open for further replies.
Top