AFCI Breakers

[mbrooke]

Tony, that response suggests to me that you do not understand how our GFCIs and their test buttons work or that fact it is the only way to test them when they are not connected to equipment grounding conductor as is often the case in older homes.


The test button on a GFCI is a legitimate test of its function.


That said I have no idea if the above holds true for our AFCIs.

The test button on US GFCIs creates an imbalance via resistor, but that does not show or test how fast it takes for the GFCI to trip. And technically the amount of imbalance current the resistor places across the toroid coil is unknown too, ie is it 5ma or 50ma? The tests done in the UK actually ramps the current to a precise value and then calculates how long it took for the RCD to trip, a much more precise method because in the end you obtain actual numbers (both trip time and current value for that trip time).


I get the test button is legitimate from an NRTL/written perspective, but its not precise. I have pushed test buttons on GFCIs were it was obvious it took a second for the unit to trip.

The responses you have given thus far suggest that you do not know how British RCDs work nor how they are tested.

 

[GoldDigger]

The test button on US GFCIs creates an imbalance via resistor, but that does not show or test how fast it takes for the GFCI to trip. And technically the amount of imbalance current the resistor places across the toroid coil is unknown too, ie is it 5ma or 50ma? The tests done in the UK actually ramps the current to a precise value and then calculates how long it took for the RCD to trip, a much more precise method because in the end you obtain actual numbers (both trip time and current value for that trip time).


I get the test button is legitimate from an NRTL/written perspective, but its not precise. I have pushed test buttons on GFCIs were it was obvious it took a second for the unit to trip.

The responses you have given thus far suggest that you do not know how British RCDs work nor how they are tested.

The big problem with the British RCD testing method when applied to GFCI receptacles is that the test equipment will have to be connected to some other current carrying conductor (another phase or the PE wire) in order to generate a current imbalance. That is not available locally in many US installations of receptacle GFCIs.
Now panel mounted GFCI breakers are another matter entirely. For those, an external tester is practical.

 

[mbrooke]

I have faith in AFCI'd doing what they say they can do.......not what they say they cant do....

But in all honesty it has NEVER been established that arc faults are responsible for dwelling fires or even to what degree. Yes testing has shown under the right conditions a lamp cord or damaged cable will arc, but none of that has determined to what degree it takes place in the real world.

I know you have everyone's best interest in mind, and I am sure the CPSC, NEMA, ect and the CMPs do, but they are not infallible or immune to human error. People make mistakes, everyone does. At one point the best minds thought the earth was flat, and latter that everything in physics had already been discovered with the only thing needing more precise measurement; of course astronomy, relativity and quantum mechanics proved wrong. Its possible in an effort to target a very real problem (like electrical fires) that something was overlooked.

 

[mbrooke]

The big problem with the British RCD testing method when applied to GFCI receptacles is that the test equipment will have to be connected to some other current carrying conductor (another phase or the PE wire) in order to generate a current imbalance. That is not available locally in many US installations of receptacle GFCIs.
Now panel mounted GFCI breakers are another matter entirely. For those, an external tester is practical.

That's very true, however newer installation within the last 40 years will have an EGC, and if the tester isn't able to pass imbalance to trip the GFCI that right there shows an open EGC which is a code violation in itself. (Ok, 3 plug tester will show that, but it doesn't compare to an earth fault loop impedance test )


EDIT: BTW, I just want to add an important point: Most British 30ma and above RCDs do not contain any electronics, which in term means less susceptibility to failure unlike AFCIs which are packed with electronics.

 

[don_resqcapt19]

The test button on US GFCIs creates an imbalance via resistor, but that does not show or test how fast it takes for the GFCI to trip. And technically the amount of imbalance current the resistor places across the toroid coil is unknown too, ie is it 5ma or 50ma? The tests done in the UK actually ramps the current to a precise value and then calculates how long it took for the RCD to trip, a much more precise method because in the end you obtain actual numbers (both trip time and current value for that trip time).


I get the test button is legitimate from an NRTL/written perspective, but its not precise. I have pushed test buttons on GFCIs were it was obvious it took a second for the unit to trip.

The responses you have given thus far suggest that you do not know how British RCDs work nor how they are tested.

My understanding is that the resistor is sized to flow about 8 mA at 120 volts.

The UL standard for the trip time for an 8mA ground fault is ~3.7 seconds. The standard says the maximum permitted time to trip is

The maximum permitted time to trip in seconds is equal to the quantity (20/fault current in milliamps) raised to the 1.43 power.

It would be rare that anyone would hold the test button in that long, so if the device trips when you push the built in test button, it is very likely it is within the standard.

 

[kwired]

I?ve asked about testing GFCI?s on here before. If anyone was to suggest ?pressing the test button? as an acceptable test in the UK they would be laughed at.

Should they call the button something besides "test"? I guess some people think you need sophisticated equipment to test most anything or else you are a hack.

Maybe the AFCI manufacturers can add a USB port and change to the safety razor business model:
The breaker is cheap and you pay for software updates to keep up with nuisance trips?

I had a similar thought once, maybe not with the USB port but some ability to update firmware of the AFCI at the very least.

1) If you where the manufacturer of AFCIs would you rather sell 1 or 40 to each homeowner

As a businessman that is one are I mess up at. Just today I talked a customer into a wireless control device over a wired setup for what they wanted to do. Is a regular customer, but I could have made much more had we gone with the wired controls. (would have had to trench lines to another building, and run some raceways in the building on the "operator" side, instead will just be minimal wiring on the "load" side and a relatively inexpensive "key FOB" type of transmitter to operate from another building.

As an electrician you may get by with spending $500-$600 on materials but for the average homeowner it's going to cost a bit more than that.

For the average homeowner the cost doesn't stop when they buy the house. After the warranty is up on the home then every time there is a problem they have to call an electrician.

Here is something that hasn't been brought up. Testing the Arc Fault, I have a Siemen's package right in front of me. It says to test every month, especially in thunder-storm season.
How many of these do you think ever get tested? I wonder how many homeowner even know they are supposed to be tested. They do make sure to state that's it's required by code on the package.

I wouldn't have any issue with $500-600 added cost to the house either, what annoys me with a lot of junk they make these days, yes HVAC equipment, appliances, etc. are all junk anymore, is the maintenance cost over the lifetime of owning it. Older simpler items were maybe less energy efficient, but most of them didn't break down as often either. The AFCI trip issue for unknown reasons kind of fits in that category for the average homeowner - an added expense to have someone come look at the problem - and probably has to come back more then once because they really don't know what is happening to cause the problem

I rarely check my own GFCI receptacles (no GFCI breakers yet....), but I do test the GFCIs in public places that I visit. About 30% are not working.

I also push TEST buttons on emergency lighting. :angel:

Do you check to see if there is voltage at the GFCI's that don't trip? Should there be no power test will not do anything.

 

[mbrooke]

My understanding is that the resistor is sized to flow about 8 mA at 120 volts.

The UL standard for the trip time for an 8mA ground fault is ~3.7 seconds. The standard says the maximum permitted time to trip is


It would be rare that anyone would hold the test button in that long, so if the device trips when you push the built in test button, it is very likely it is within the standard.

That is true, and excellent info btw (I think you got distracted mid way through the second paragraph)


However, we don't know what the test current really is or the exact opening time. An microprocessor tester is far more accurate, and the ramping current effect is more thorough over a fixed, unknown resistor.

 

[kwired]

That is true, and excellent info btw (I think you got distracted mid way through the second paragraph)


However, we don't know what the test current really is or the exact opening time. An microprocessor tester is far more accurate, and the ramping current effect is more thorough over a fixed, unknown resistor.

True, but if it trips with any leakage at all, it is still better protection against shock hazards then standard overcurrent devices are. I still think that if we were to change the standard to 30 mA fault protection we still would have a pretty effective safety device at preventing electric shock and less trips that make you wonder what happened.

 

[mbrooke]

True, but if it trips with any leakage at all, it is still better protection against shock hazards then standard overcurrent devices are. I still think that if we were to change the standard to 30 mA fault protection we still would have a pretty effective safety device at preventing electric shock and less trips that make you wonder what happened.

True, any trip is better than no trip. But still better to have precision knowledge imo. Don't get me wrong though, IEC RCDs have a test button with a resistor as well, but foreign electricians often receive training that requires them to augment the standard push button test with a testing device.

Changing the trip threshold should be looked at closer; in adults 30ma fault protection would still work in theory, but more research would need to be done in regards to children. If children do have a lower threshold then we may need to keep 5ma as is. Anyone know what the threshold btw?

However, placing a 30ma RCD at the branch circuit origin for all 120 volt circuit would provide the same protection as AFCIs without nuisance tripping and in addition if IEC designs are employed its possible to do away with the electronics.

 

[kwired]

True, any trip is better than no trip. But still better to have precision knowledge imo. Don't get me wrong though, IEC RCDs have a test button with a resistor as well, but foreign electricians often receive training that requires them to augment the standard push button test with a testing device.

Changing the trip threshold should be looked at closer; in adults 30ma fault protection would still work in theory, but more research would need to be done in regards to children. If children do have a lower threshold then we may need to keep 5ma as is. Anyone know what the threshold btw?

However, placing a 30ma RCD at the branch circuit origin for all 120 volt circuit would provide the same protection as AFCIs without nuisance tripping and in addition if IEC designs are employed its possible to do away with the electronics.

In the majority of trip incidents there is fault current that trips the device before any person becomes a current path. 30 mA threshold probably catches most of those incidents anyway, the 4-6 mA class A standard is probably there just for those times when the person is the link that causes fault current to flow, and don't think you won't feel that 4-6 mA. Some think GFCI's prevent shocks, but is not true, they just limit duration when one does occur.

 

[mbrooke]

In the majority of trip incidents there is fault current that trips the device before any person becomes a current path. 30 mA threshold probably catches most of those incidents anyway, the 4-6 mA class A standard is probably there just for those times when the person is the link that causes fault current to flow, and don't think you won't feel that 4-6 mA. Some think GFCI's prevent shocks, but is not true, they just limit duration when one does occur.

Exactly. The 5ma is probably there in the event the person is the full EGC and can not let go. 30ma will catch most other cases and what the CMP should have required instead of AFCIs. 30 or 50ma for all 120 volts circuits and then a 5ma GFCI where needed. This imo would have given us one of the safest system on earth instead of AFCIs some of which do not even have GFP. :rant:


FWIW I have heard experts talk about that GFCIs came about due to a lack of ungrounded tools/no ground prongs. So 5ma may have been part of the equation.

 

[mbrooke]

1) If you where the manufacturer of AFCIs would you rather sell 1 or 40 to each homeowner

They would want to sell 40+ of them per home, in addition to self testing GFCIs or a dual function breaker that only costs a few cents to lower the GFP from 50 to 5ma but sell it at near double the cost.

2) If you where the electrician forced to install AFCIs that we all know nuisance trip would you rather the customer was put entirely or even 50% in the dark or limit it to just one circuit. Please no snap answer, put yourself in the customers position. Would you be happy if each time you tried using your tread mill 50%~100% of the power goes out?

Im with you on this. Id rather the customer have a headache one or a few circuits at a time (we all know it wont be just one circuit but several once the computers and laptops get plugged into each room) rather then having half or the whole home black out.

3) If you had to troubleshoot a tripping AFCI for free* would you rather it have one circuit or 20 circuits connected to it?


(*I say free as that happens often as customers call expecting warranty service on new installations and many ECs feel obligated to do so to keep their good reputation)

Id rather have one circuit, because there is absolutely no testing device on the market than can tell me what wave forum is bother the AFCI and where its coming from... good chance it will be coming from nearly all of them.

Lets dump AFCIs and find something else to use at the branch circuit level, be it RCD, GFCI or lower instantaneous trip ratings.

I agree. If we ditched AFCI we could go with the submain approach because a 30/50ma GFP will not nuisance trip. If the total combined circuit length exceeds say 1200ft it would be a good idea to have another or a 3rd sub main (excessive capacitive coupling can nuisance trip) , but in any case when done right we know they will only trip for a genuine fault.

Even if the submain idea is frowned upon a GFCI breaker for every 120 volt circuit would be the ultimate solution, but certainly wont be welcome by manufacturers:


GFCI breaker, $30 x 20= $600


or the manufacturer preferred options



$15, GFCI receptacle

self test for that receptacle, another $10, so 25x12=300

AFCI, $37x20=740

740+300= $1,040

or

Dual function AFCI+GFCI 52x12=624

AFCI $37x8= 296


$920

The advantage of having different terms, each offering its own price point. :happyno:

 

[Tony S]

Thanks!

Just for to show the US folks, do you have a pic of the consumer unit? Or one using an RCD that feeds multiple MCBs? I still cant believe this option was never offered for AFCIs. If folks in the UK saw Im sure they would be laughing even harder. :lol:

Believe me they are laughing.

Pictures as requested:

A 17^th edition high integrity board.
Working from the right.
Main isolator
MCB with no earth fault detection, this would be used for smoke detectors or maybe a garage supply where a local RCD would be fitted.
RCD 1 section where half the lighting and half the sockets are connected
RCD 2 section where the other half of the general circuits are connected

An all RCBO board
Every outgoing way has its own earth fault detection

Mr. Brooke, you really are a wind up merchant ;-)

http://forums.mikeholt.com/showthread.php?t=165856&page=3

I?ll not send the e-mail, you?ve seen the drawing.

 

[romex jockey]

All this belief posting really isn't productive.

Isn't there some NRTL test procedure for afci's ?

One would imaging the details of it's function to follow suit.

In fact, just what test procedure in (2008) allowed the debut of the combination afci?

How did it assume the status of combination, which was assumed prior to this?

~RJ~

 

[iwire]

Believe me they are laughing.

If there were laughing that would only because they do not understand the test button is enough with the our GFCIs.

 

[mbrooke]

If there were laughing that would only because they do not understand the test button is enough with the our GFCIs.

I disagree, UK RCDs are even simpler than our GFCIs, a detailed report is better than a test button alone. Kind of like testing a diesel back up generator 30 minutes a month vs starting it up then 95% load banking it for 2 hours a month.


I however will let Tony have the honor of explaining it.

 

[mbrooke]

Believe me they are laughing.

Pictures as requested:

A 17^th edition high integrity board.
Working from the right.
Main isolator
MCB with no earth fault detection, this would be used for smoke detectors or maybe a garage supply where a local RCD would be fitted.
RCD 1 section where half the lighting and half the sockets are connected
RCD 2 section where the other half of the general circuits are connected

An all RCBO board
Every outgoing way has its own earth fault detection

Mr. Brooke, you really are a wind up merchant ;-)

http://forums.mikeholt.com/showthread.php?t=165856&page=3

I?ll not send the e-mail, you?ve seen the drawing.

Thanks, that is awesome! Beautiful work btw


May I ask, what purpose do the white wires serve coming from the RCBOs and then connecting to the earthing terminal?

 

[iwire]

I disagree, UK RCDs are even simpler than our GFCIs, a detailed report is better than a test button alone. .

Unless you can show me some information showing that the trip level of a GFCI 'drifts' out in the field the yes or no results are plenty.

A report would just be engineer porn.

 

[mbrooke]

Unless you can show me some information showing that the trip level of a GFCI 'drifts' out in the field the yes or no results are plenty.

But it is possible in theory, especially if the trip mechanism becomes slow to unlatch. There is the argument the resistor pulls 8 ma however in reality its not known or set in stone for every manufacturer. Take the picture for example where the resistor is 10k, pulling 12ma. 12ma might work for a higher value above 6ma but not below say 8ma.

A report would just be engineer porn.

In my world, that would be more than enough reason

 

[kwired]

Exactly. The 5ma is probably there in the event the person is the full EGC and can not let go. 30ma will catch most other cases and what the CMP should have required instead of AFCIs. 30 or 50ma for all 120 volts circuits and then a 5ma GFCI where needed. This imo would have given us one of the safest system on earth instead of AFCIs some of which do not even have GFP. :rant:


FWIW I have heard experts talk about that GFCIs came about due to a lack of ungrounded tools/no ground prongs. So 5ma may have been part of the equation.

IIRC first NEC requirements for GFCI's was for swimming pool equipment and then for bathroom receptacles.

All this belief posting really isn't productive.

Isn't there some NRTL test procedure for afci's ?

One would imaging the details of it's function to follow suit.

In fact, just what test procedure in (2008) allowed the debut of the combination afci?

How did it assume the status of combination, which was assumed prior to this?

~RJ~

The AFCI manufacturers sold their concept on the NRTL's just like they did the NEC, and I believe most improvements they have made to their products has been by their own findings not suggestions from any NRTL. So in essence the manufacturers made the device, set the standards, set the testing requirements (or at least had a major role in setting testing requirements for listing) basically they have total control over design, approval, and introduction into codes, they put enough money and time into gaining this control that nobody can afford to challenge anything they have done - that is why you have all these people that have ideas of what is wrong with them and even show demonstrations - but they don't have the financial backing to take it to the level to actually challenge the AFCI manufacturers. Not saying either side of that is totally correct, but maybe if they could both eat at the same table - maybe we could get a useful and reliable product for the purpose sooner instead of later.