Leviton Arc-Fault tripping on LED lighting.

M

makesends

New User
Location
Oregon
Occupation
Electrical/Electronics Tech
In my last job, almost all the lighting the customer wanted is LED, so I ran most of the house lighting on two 15amp lighting circuits. They also wanted the Leviton breaker panel (which, otherwise, I do too). The problem is that though neither breaker is overloaded (RMS current), at some point both breakers trip according to the number of fixtures/ amount of load on Arc-Fault only. It usually happens on starting, for example, undercabinet lights. The inspector says I must have arc fault protection even on the lighting circuit over living areas, even though they have no receptacles involved.

My question is, would a choke/reactor on the line solve the problem, or is there some other suggestion you might have? I'm guessing the breaker is reading the accumulation of switching mode power supplies in the fixtures as arcing. These people also have two bedrooms they will be using as offices (therefore, Arc Fault breakers on the related receptacles. I'm dreading what those computers and servers will do.
 
First look at 210.12(A)
Then go to Art 100. 'outlet'.
That will explain your inspector's opinion. He is correct.

What do the analytics of your breakers indicate?
LED can have a very high startup current that you will not catch with a typical meter.
My suggestion would be more circuits.
 
Do the lighting circuits have dimmers? Sometimes they can contribute to AFCI tripping issues.

I came across a video where a blank face AFCI was used instead of an AFCI breaker because it was apparently less susceptible to tripping on certain loads:

 
Parallel arc and series arc are very easy to distinguish on the diagram, but everything is series in reality. The distinction can get blurry.

Series is something sensed like a bulb that's turned on, but flickering because of a loose connection.
Parallel is something sense as a bulb that's off, but flashing briefly as it makes intermittent unintended connection and causing it to come on.

There is a legitimate arc every time you interrupt a flow of current with a mechanical switch but this should be ignored, but if it happens repeatedly in a small amount of time or it goes on for a long time, the AFCI will trip. The decision making process is depends how each AFCI is designed or configured.

The reason the guy in the YouTube video a few posts up mentions microwaves and clothes washers is that they have a current signature that some AFCIs interpret as one of the arc pattern. His idea is to get around it by using "some other" AFCIs hoping that they don't interpret the signals quite the same way.
 
Last edited:
Flicker Index said:
The reason the guy in the YouTube video a few posts up mentions microwaves and clothes washers is that they have a current signature that some AFCIs interpret as one of the arc pattern. His idea is to get around it by using "some other" AFCIs hoping that they don't interpret the signals quite the same way.
Click to expand...
And we shouldn't have to put up with this.

I'm not against what they are trying to accomplish with AFCI, but they have made the consumer part of their long term testing lab and many installers have had to eat costs on it and they still haven't perfected this technology yet. they maybe have replaced some devices for no charge, they probably have not reimbursed much or even any man hours for this nonsense though.
 
makesends said:
In my last job, almost all the lighting the customer wanted is LED, so I ran most of the house lighting on two 15amp lighting circuits. They also wanted the Leviton breaker panel (which, otherwise, I do too). The problem is that though neither breaker is overloaded (RMS current), at some point both breakers trip according to the number of fixtures/ amount of load on Arc-Fault only. It usually happens on starting, for example, undercabinet lights. The inspector says I must have arc fault protection even on the lighting circuit over living areas, even though they have no receptacles involved.

My question is, would a choke/reactor on the line solve the problem, or is there some other suggestion you might have? I'm guessing the breaker is reading the accumulation of switching mode power supplies in the fixtures as arcing. These people also have two bedrooms they will be using as offices (therefore, Arc Fault breakers on the related receptacles. I'm dreading what those computers and servers will do.
Click to expand...
A few years ago I was working with a guy who had rented the Dranetz German power quality meter (really expensive high end) and he was bewildered by the readings of an LED light bulb. It had a massive power spike and horrible power factor loss for a nano second when turning on. The current load switching on was 9X of the actual energy when on. I didn't believe it, what I was seeing. We ran this test on other LED bulbs, fixtures and found all kinds of different readings, but the same deal: Massive power factor loss and nano second current spike.
 
herding_cats said:
A few years ago I was working with a guy who had rented the Dranetz German power quality meter (really expensive high end) and he was bewildered by the readings of an LED light bulb. It had a massive power spike and horrible power factor loss for a nano second when turning on. The current load switching on was 9X of the actual energy when on. I didn't believe it, what I was seeing. We ran this test on other LED bulbs, fixtures and found all kinds of different readings, but the same deal: Massive power factor loss and nano second current spike.
Click to expand...
Tungsten filament lamps have just as high or even higher inrush on a cold start. They are linear current devices though and likely not going to cause as much issues with GFCI/AFCI's
 
herding_cats said:
A few years ago I was working with a guy who had rented the Dranetz German power quality meter (really expensive high end) and he was bewildered by the readings of an LED light bulb. It had a massive power spike and horrible power factor loss for a nano second when turning on. The current load switching on was 9X of the actual energy when on. I didn't believe it, what I was seeing. We ran this test on other LED bulbs, fixtures and found all kinds of different readings, but the same deal: Massive power factor loss and nano second current spike.
Click to expand...
kwired said:
Tungsten filament lamps have just as high or even higher inrush on a cold start. They are linear current devices though and likely not going to cause as much issues with GFCI/AFCI's
Click to expand...

Here's the thing though, while software driven devices allow for post-production defect removal (affectionately called "firmware update"), they're insanely unreliable for critical applications.

Solid state landline phones with push buttons may contain some digital circuitry, but those devices are not embedded systems crap driven, they may experience issues that you can't ever resolve, but they're not prone to the reliability issues of modern software driven smartphones that suffer from occasional latch-ups that require rebooting.

Modern software based design allows for firmware updates and adaptive learning, but they suffer significantly higher chance of glitches, like software hang-ups that makes the device fail to trigger. False negative is simply unacceptable for AFCIs, but without software driven system that puts it at risk of such, AFCIs suffer from hard-coded vulnerabilities to some false positives that can only be resolved by redesigning the circuitry.
 
I had a microwave on a dedicated circuit that kept nunance tripping AFCI. I Changed it out with the microwave in the other break room. Just swapped. Everything new and all equipment the same, Problem went away. I was never sure what caused it, identical microwaves, identical panels, identical AFCI
 
Joethemechanic said:
I had a microwave on a dedicated circuit that kept nunance tripping AFCI. I Changed it out with the microwave in the other break room. Just swapped. Everything new and all equipment the same, Problem went away. I was never sure what caused it, identical microwaves, identical panels, identical AFCI
Click to expand...
Now the question arises. Which one was working properly?
 
ptonsparky said:
Now the question arises. Which one was working properly?
Click to expand...
Yeah good question. I couldn't find anything wrong. I know it happened right as soon as the start button was touched. The only thing I'm not sure of is if they were both on the same phase of the 208/120 wye.
 
herding_cats said:
A few years ago I was working with a guy who had rented the Dranetz German power quality meter (really expensive high end) and he was bewildered by the readings of an LED light bulb. It had a massive power spike and horrible power factor loss for a nano second when turning on. The current load switching on was 9X of the actual energy when on. I didn't believe it, what I was seeing. We ran this test on other LED bulbs, fixtures and found all kinds of different readings, but the same deal: Massive power factor loss and nano second current spike.
Click to expand...
A few years back I was working on a project in our shop which used (4) 24VDC LED lights powered by a 24V traction battery in a "Tugger" (factory vehicle for pulling carts). They were relatively small lights, and drew a combined ~4A when ON. Long story short, I come to find out the inrush for those lights was ~140A !! It caused problems with the relay that turned them on and off (it would typically weld the contacts closed).

Like Herding_Cats describes in post #12, I found the whole inrush event was incredibly short. The spike was gone by ~0.25 msec. I assumed it was poor design that had no - or nearly none - current limiting when first charging up whatever caps were in the internal LED power supply. They simply appeared like a Black Hole when the 24V first came up.

It was an interesting problem to have to solve, made worse by the fact they are DC powered. I can try to write up a separate thread about it - if anybody would find it interesting. I still have the data and scope traces I'm sure.
 
You must log in or register to reply here.
Top