Experiments with AFCIs and GFCIs

[ELA]

AFCI testing data

AFCI testing data

I finally got some time to start testing my QO115CAFIC. My ultimate goal is to better understand just what the test switch accomplishes. In the partial schematic I have seen it appears only as an input to the microcontroller.
Without knowing what the firmware does this leaves it open to being as simple as tripping the latch?
I hope to better understand what the test switch does later.

In working up to that I have tested several loads such as 3 different vacuum cleaners and 900 watt heater loads. The heater loads were in series with a cheap phase fire type dimmer set at 25%, 50%, 75% and 100% setting and the breaker held fine (no nuisance trips).
This was in an attempt to create a load that might "simulate a series arc" so that I can develop a repeatable test I know will trip the breaker when seen as an arc. This is was also to identify what kinds of loads would not false trip the breaker.

I further tested to confirm the GF leakage trip point of the device at ~40ma. It held at 38 ma for more than 10 seconds and tripped at 42 ma in less than 2 seconds.
It held with a 4 watt night light as a load (hot to ground) even though the cold resistance measured about 300 ohms. Hot current = 33 ma

Thus far the most interesting was its detection of a Neutral to Ground connection. The breaker tripped with a direct short between load Neu. and Ground just fine. I was surprised to see that it held with anything greater than ~0.2 ohms. So this appears to indicate that a pretty good connection (neu-gnd) is required downstream of this to appear as this type of fault.


Manually tripped the test switch and the only current recorded on the input hot line was a 3 amp 1/2 cycle pulse required to power the coil to trip the unit.

I will now work towards creating my repeatable arc fault. I considered using carbon rods in series with my heater load but that seemed a but messy. I am considering building a electronic "arc simulator" where I can control the waveforms shape.
I know some "Arc Fault Testers" are sold out there but I have not seen one that gave thorough specifications on its capabilites so I am leary of investing in one. Then again I am not sure I have the time and energy to build my own either. Could be fun Suggestions on possible (electronic) arc type loads welcome

 

[gar]

080924-2049 EST

ELA:

An arc is an approximately constant voltage drop vs current.

For a discussion of arcs see Fundamentals of Engineeing Electronics, by William G. Dow, John Wiley & Sons, 2nd Edition, 1952, Chapter XV. The voltage drop is made up of anode fall. plasma drop and cathode fall. Dow was the driving force behind what became the Willow Run Laboratories, and the U of M Electronic Defense Group. He had his special way of teaching which was to make you think out problems yourself.

You might make a resistive load in series with a DC supply. Make the DC supply such that it can be switched in series with the resitive load at a controlled time in the cycle, like 75 V. This might simulate a series arc.

.

 

[ELA]

Gar,
Actually it was this site that you referenced earlier that gave me the idea for a series arc waveform:

http://www.zlan.com/testing/testing1.htm

I had planned on using two back to back IGBTs and a micro controller to "Start late", "End early" and then introduce a few dips each side of the peaks as a place to start from. I like the idea of having such a test device around for (phase controlled) inrush testing in addition to this task.