AFCI test rig and test design

[Robert Green]

I just received my AFCI 20A duplex outlet and downstream OBC protector:

Leviton AFTR2-GY SmartlockPro Outlet Branch Circuit Arc-Fault Circuit Interrupter Receptacle,

20-Amp, 120-volt, Grey

At first I was miffed about not noticing it was grey-colored, but then I decided it would be useful to know which outlet were AFCI "master" outlets at a glance.

I took a 14AWG three-wire cord and plug from an old A/C unit and wired it to the outlet which I installed into a plastic duplex outlet box. Now, anything plugged into the outlet would be arc-fault protected. But I soon realized that to test the protection of outlets that would be downstream, I needed to connect outlets to the LOAD terminals of the AFCI. So I took an outlet strip that had a broken plug, cut that off and attached the cord to the LOAD terminals inside the plastic box. Now, anything that generated an arc plugged into either the duplex outlet or the outlet strip (no surge protection - just six outlets and a switch/circuit breaker).

The next phase is to figure out some testing parameters and arc-testing situations. For the test item, I thought I'd use an old plastic hairdryer that can pull 1500 watts and that has a pretty sparky motor. Same for a 70 year old Black and Decker 1/2" drill (I'll have to take pictures - they just don't make things like they used to) that has a notorious spark.

I will also test those items and a 1500 watt space heater with old, corroded extension cords and outlets that have been replaced for "cause" in the past. (There *is* a reason to save dead stuff!). As Bud pointed out, the AFCI's need a fairly large "base" current (5A) to respond to an arc fault. I believe all three test items qualify. I am not sure I am going to create a carbon arc - only because I can't seem to find my stash of carbon rods at the moment. I might try to arc some pencil leads but that would have very low SAF.

I also want to test the rig I've built with GFCI's. To that end, I will use my push button GFCI tester both with the AFCI plugged into a GFCI outlet and a GFCI outlet plugged into the AFCI test rig.

From looking around it seems that cost-wise, GFCI breakers with AFCI outlets in the cheapest way to go. I can't figure out why the AFCI outlets are cheaper than the breakers, but that seems to be the case. My best guess is the size factor - they have to "shrink" the AFCI components to get them to fit into a breaker slot.

The outlet itself is a little larger than a GFCI but seems much heavier - will try to remember to weigh them. It has a bump on the back that would make it a tight fit for passthrough use. The unit comes with double screw clamps that are sort of like a back-stab outlet, but the wire is inserted under a screw down brass clamp (no hook needs to be formed as with some screw terminals so that's one less point of probable failure). The unit ships with the reset button tripped as a precaution.

The outlet is quite stiff - it comes with a child-proof sliding shutter on the hot slot and has a T-shaped neutral hole to accept 5-20P NEMA plugs (IIRC). Not sure if that means the outlet's going to hold plugs more securely or if the added force needed to insert the plug is going to result in - you guessed it - partially inserted plugs and potential arcing.

Any ideas about what other tests I might perform are welcome. I'm just not comfortable relying on an on board "Test Button" to determine whether the AFCI is still protecting against arc faults. I'd like to be able to come up with some test rig or procedure that will simulate real-world arcing conditions.

[gregz]

For some reason, I always like to arc the light switches. You put a load across a light switch, then very slowly make or break the connection. You can usually hear the buzz. Everyone has done that, right ?

Greg

[Robert Green]

Not me! That's a good one. I hadn't thought about wiring in the lighting circuits to protect against arc faults but you bring up a good point and another item to add to my checklist. I probably should break the AFCI outlet out of the single gang box and wire it into a duplex box test rig with a light switch I can "jigger" to create an arc.

I seem to recall at least a few of the switches I have are "snap" switches specifically designed to resist arcing. I'm sure I have something that will work in the "box of the undead" electrical equipment. I even have some switches going back to the to the two black cylinders stacked like stoplights where when you push the lower one in, the upper one pops out. I assume they're from 1940 or so because they were in the box I inherited from my junk saving father. (-:

Thanks for your input, Greg.

[bud--]

If you use carbon rods or some similar arc I suggest you use something like the 1500W space heater in series. That limits the current. Else you can have a current that is the available fault current at the outlet, which one of the articles said is 60A or higher (if I remember right).

Obviously this is a shock and fire hazard.

But that does not necessarily protect the house wiring. The NEC allows an AFCI receptacle to protect a branch circuit if it is the first device, and the branch circuit is wired through it, and the wiring to the panel is something like in metal conduit.

If you need an AFCI breaker for a SquareD panel it comes from SquareD and they can charge more. Receptacles can be made by anyone (although I don't know of anyone but Leviton that make them now).

You used to be able to reverse-wire (line-load) a GFCI receptacle and the receptacle would seem to work as expected, but the receptacle is not GFCI protected (downstream is) and a GFCI trip did not disconnect the receptacle (but did disconnect downstream).

The UL standard was changed. Now if reverse-wired the receptacle is still not GFCI protected. But a trip now disconnects the receptacle and it can't be reset. To protect from reverse- wired installation, the GFCIs are shipped tripped (and can not be reset).

The AFCI receptacle includes ground fault protection, and I suspect the same applies.

Child-proof receptacles are generally required for new construction and replacement in a residence.

---------------------------------------- A "snap" switch likely has an AC-DC rating. The silent switches are an AC only rating.

[TimR]

On Tuesday, July 16, 2013 4:34:17 PM UTC-4, Robert Green wrote: I am not sure I am going to create a

I haven't opened one since I was a kid, but a D cell used to have a carbon rod as the positive terminal, about 3/8 inch in diameter, and the lantern battery had one almost an inch across.

[N8N]

if you want to get a carbon rod out of a cell (a "battery" is multiple cells... yeah I know I'm being a pedant) you want a "heavy duty" cell not an alkaline one.

[Stormin Mormon]

Still true, in the case of carbon zinc cells.

I haven't opened one since I was a kid, but a D cell used to have a carbon rod as the positive terminal, about 3/8 inch in diameter, and the lantern battery had one almost an inch across.

[TimR]

So what's inside an alkaline one?

[Robert Green]

Thanks. That sounds like an excellent idea. I recall reading that as well.

I'm souring on the carbon-arc testing based on what you've pointed out and how far away it is from real-world faults that are likely to set off the AFCI.

I've only got the NEC updates guide so I am not sure what the codes are. I assume the reason the feed to the first AFCI receptacle should be in conduit is so that if an arc does form before reaching the protection device, it's safely enclosed and not like to spark a fire. The circuits I want to protect are all new ones I installed to power devices that I didn't want to run through old, cloth covered wiring runs rated (when new) at 15A. They are circuits for window ACs, toasters, microwaves, space heaters and the workbench outlets powering the RAS and other power hungry tools.

That sort of makes sense. (-: I am reluctant to use AFCI breakers because of the pigtail issues (the circuit box is crowded enough) and with outlet AFCIs I can quietly remove them (yes, I know they are fixtures and that's technically a no-no) and replace them with standard outlets when I move.

Here's another question. What does "Spec Grade" mean on the Leviton AFCI? Is it just marketing hype?

Yes. After I posted that message I did some more reading and Leviton says what you just did. It's shipped tripped to make sure it's not installed incorrectly. Good idea but confusing if you expect it to work out of the box.

Do you know how the plug-in GFCI testers with the pushbutton work? Are they creating a current mismatch between the two conductors?

So it seems. I wasn't kidding when I wrote that ironically, making a plug tough to insert completely could lead to arcing. Perhaps it's because it's so new but it's a real chore to insert a three-wire plug into the outlet. While I have a spring scale that can tell me how much force it takes to remove a plug from the outlet, I don't know of any way to measure the insertion force. Perhaps placing the outlet on a bathroom scale and then noting what the scale reads when the plug is fully inserted would give me a ballpark figure.

Hmm. Do you know why that is? Is it related to arcing or some other issue?

Thanks for your input, Bud, it's been very helpful. I am sure my wife will appreciate your "talking" me out of carbon rod arcing tests.

[Robert Green]

one almost an inch

Oh yes, now that you mention it, that's probably a good source for the carbon rods. Time to get out the hacksaw just to check. (-: I think I will skip the carbon arc test for the reasons previously stated but knowing whether there's still a carbon rod in a D cell is worth investigating.

[Robert Green]

cells... yeah I know I'm being a pedant) you want a "heavy duty" cell not an alkaline one.

Got both. Didn't realize they had different internals. Thanks!

[hrhofmann]

On the subject of trhe over and under push-button light switches, those dat e back to the early 1930's, probably even earlier. My grandparents had the m in their 1901-built house in Woodside, Queens, NYC, in the 1940's, and th ey were old and worn looking even then (both my grandparents and the light switches). The house had been converted from gas lighting to electric by r unning the electric wires in the old gas pipes in many of the rooms. The ho use is still there as far as I know, 39-27 56th street.

[bud--]

But think of the fun....

Correct. (It is a little broader than just metal conduit, but I am too lazy to look it up.)

The current NEC generally requires AFCI protection in a house for new circuits that don't have to be GFCI protected. (And gfretwell recently wrote that the 2014 NEC adds laundry and [another one] that will have to be AFCI protected.

They kinda become non-code compliant.

I think my post in a recent thread got into use of AFCI receptacles.

The NEC now wants AFCI protection on most new house circuits and extensions of existing circuits where the extension is in an area that requires AFCI protection.

You can also get regular receptacles and switches that are "spec grade". It is a higher quality and should last longer. Might be particularly good in a kitchen.

The other grade is "consumer" or "builder" (and there may be an intermediate grade).

The connect a resistor from the hot wire to the ground wire. The resistor is sized for about 5mA. The current on the hot wire is 5mA higher than on the neutral. They do not work if there is no ground.

GFCI receptacles and breakers have the same resistor but connected from the downstream hot to the upstream neutral (or vice versa). The current on the hot and neutral are about 5mA different. The ground wire is not used and the test will work on a receptacle without a ground.

When switching AC, the zero-crossings help extinguish the arc of the opening contacts. DC can arc much farther and you want a fast acting (snap) switch. If a switch has both AC and DC ratings, the DC voltage rating should be considerably lower than the AC rating.

AC only switches have a motor rating at something like 80% of the normal rating. Not likely AC-DC switches do.

[Robert Green]

(-:

I'm going to have to content myself with making things marginally safer but not as safe as they could be. I'd like to have GFCI/AFCI protection on every circuit but for now only the ones that pose the greatest threat will get upgraded.

It makes sense. I just wish a complete upgrade wasn't a job of titanic proportions.

I didn't realize they need a ground to operate correctly. I should have saved the manual.

You answered my next question before I asked it!

I see. I had noticed the disparity in switch ratings before but had no idea until now why they should be rated differently. Thanks for the edumacashun.

[Robert Green]

These switches came from a house in Brooklyn, NY of about the same vintage. One button is white (for on) the other is black (for off). They still work, too.

Same deal for gas lights converted to electrical lights. Heated by coal until the 40's, IIRC.

[hrhofmann]

I'd bet those old switches would arc, as I don't think they had a snap-action mechanism, more like a knife-switch opening and closing.

[Robert Green]

snap-action mechanism, more like a knife-switch opening and closing.

That's a good idea - I have some old porcelain-base knife switches that I can use to create an arc.