I know nothing about AFCI breakers, but I googled it and found out that it could be your long wire run's cable capacitance, electrical noise, damaged wiring, and a defective or overly sensitive AFCI. I discounted overloads because it only happened with a very light load.

Damaged insulation may lead to arcing even when the circuit is not in use...

You may try to "sniff" the path of that circuit using an AM pocket radio, real sparking/arcing would lead to elevated buzz/cracling around the affected cable (that is, what the AFCI responds to).

An AF breaker may trip over several different things, which may be hard to distinguish. Here are a few :

 - Arcing in parallel with the line (phase to neutral or phase to earth) through a place where the insulation is damaged and a conductive track is forming. The idea of arc fault breaker is, that the energy of the momentary shorts happening in the conductive track may be below what it takes to trip a normal breaker's electromagnetic release, unless they develop into a full high current arc (The odds of which get lower when the line voltage is lower, as in 120V vs 230V)

 - Arcing in series with a load. This is one of the least reliable detection functions of AFCIs, as some series arcs dont have as clear signature (current peaks in the microsecond range and 10's of A that parallel arcs have). Some of them act more like a fairly stable resistance. There are some series arcs the breaker is meant to allow (like brushed motors), but there it helps that most such appliances have at least an X capacitor in them, which prevents most of the interference from the motor from coming out towards the breaker. (vs. a bad connection upstream, where all the interference is directly in series with the breaker). I imagine that the breakers will have some time-delay which is sufficient to ignore thermostats, as long as they click reasonably quickly

 - Some devices plugged in still have parts under voltage even when they are off. For example, the AF breaker might (i guess ?) respond to self healing events of an X capacitor in an appliance line filter, even if it does have a mechanical on off switch and is switched off

 - Interference that somehow match what the breaker is looking for. This might be induced both from within the circuit and from upstream (like from another circuit in the house, back through that circuit's breaker, panel busbar, and the AF breaker in question). I think some factors contributing to this sort of effects may be : Long circuits in general, especially circuits that extend outside of the house, or if the circuit wiring downstream is somewhere separated - as in phase and neutral wires taking different paths towards a point in the circuit. (This is not a proper way to install circuits, but had been found many times especially in lighting circuits that span between floors or across big areas of the house). The interference may come from many sources, including appliances etc, there have been cases of SW radio transmission in the area tripping AF breakers in houses...

As you mention your circuit, it seems to tick quite a few boxes of the latter....



How is the circuit laid out ? With what type of conductors ?



Disconnect all loads from it, remove any built in electronics (such as GFCIs, motion detectors etc) and perform an insulation test between all conductors (phase to phase, both phases to neutral, to earth, and neutral earth)

If you find abnormal readings, get to the place where the circuit branches to different locations, disconnect them to test the insulation separately and possibly locate the part of the circuit which might have bad insulation



Most AFCIs do have some low limit on the current they detect for parallel arcing. If the circuit is not in use (or you can make it not in use), you can watch it for lower currents - disconnect the phase end of it, and insert a low range ampermeter in series to monitor the actual current draw of the apparently open circuit. (Add an incandescent lamp in series for protection, in case some load appears which is higher than the ampermeter's fuse rating, or if the circuit indeed suddenly shorts)

An intact circuit will draw tiny (order of <1 mA), but most importantly constant current, coming from capacitance between the conductors and maybe distributed leakage currents for some insulation materials

A non intact circuit may (if you happen to catch it) draw higher current (order of few mA), and this current will sometimes tend to fluctate over the course of hours or days in a pattern - As a conductive track is slowly formed in the place of bad insulation, then at some stage it burns clear and starts over. (The addition of current limiting in series may affect the process as it limits the available current for the burn clear stage)



If the phase wire goes along with bare earth in a cable, or any similar setup where leakage is likely to be towards earth conductor or towards external conductive objects, then parallel arcing faults will also likely trigger GFCIs. If your AF breaker is not also a GFCI by itself, then adding a GFCI at the beginning of the circuit (after the arc fault breaker) will add an additional layer of safety, and might show whether there is indeed an earth leakage (the leakage may be intermittent)



You can try to swap AF breakers between this circuit and another, to see if there is some issue with the breaker. However, this test might be inconclusive. There might be a combination that's acting to trip it - There may be indeed something wrong with the circuit, but borderline, and the particular breaker might be with a little lower thresholds than another identical breaker. So if this test appears to solve the problem, you still dont know if it was interference from outside or there is really some very first stages of something happening in this circuit



If you can reconnect loads like groups of sockets permanently or semi permanently to other circuits, this will rule out parts of the circuit which may or may not be the problem (Just in case, do have an identical AF breaker on the circuit that you move the parts to). Also, it will reduce the amount of interference the circuit receives

The AFCI have quite some degree of selectivity where the RF arc signature signal is coming from. So for something being upstream it would have to be really strong. So I thing it is safe to assume thetrigger signal comes from the affected circuit.
You said the affected circuit is old. So I would really see as very likely there is something wrong with it (or something still on it).
A degrading insulation is the first suspect here.
But it could be also some cr@*py configuration, like Neutral shorted to earth. In the pre-GFCI time not that rare redneck "fix" for a broken Neutral wire. Or something not installed correctly.
Problem with arcing in the wiring and attempting to measure currents there is, on one side the current may go hard short circuit at any time, on the other hand even a few 10's of Ohm resistance in series (a light bulb used as an attempt to contain the eventual hard short circuit and not blow the Ameter) is enough to affect the arcing so it starts to act completely differently.
A clamp meter would be a better way than wiring anything into the circuit. Plus any even slight movement with the wires may greatly affect how the arcing behaves, you may easily hide the fault for quite some time orturn it into an immediate hard short.
The radio method has worked for me pretty well (to find series arcing, helped me to find quite a few dying switches and lose connections)

Maybe a question: Did youobserve any correlation between ow often and when the trips happen with things like temperature, humidity,or weather in general?

I have found some digital meters (including clamp) have some unexpected and undocumented minimum current threshold, below which they just dont show anything. Like an ammeter with 1mA resolution, but shows nothing if the current is below at least 10..20mA. From 20mA and up it shows everything correctly

The AFCI have quite some degree of selectivity where the RF arc signature signal is coming from. So for something being upstream it would have to be really strong. So I thing it is safe to assume thetrigger signal comes from the affected circuit.
You said the affected circuit is old. So I would really see as very likely there is something wrong with it (or something still on it).
A degrading insulation is the first suspect here.
But it could be also some cr@*py configuration, like Neutral shorted to earth. In the pre-GFCI time not that rare redneck "fix" for a broken Neutral wire. Or something not installed correctly.
Problem with arcing in the wiring and attempting to measure currents there is, on one side the current may go hard short circuit at any time, on the other hand even a few 10's of Ohm resistance in series (a light bulb used as an attempt to contain the eventual hard short circuit and not blow the Ameter) is enough to affect the arcing so it starts to act completely differently.
A clamp meter would be a better way than wiring anything into the circuit. Plus any even slight movement with the wires may greatly affect how the arcing behaves, you may easily hide the fault for quite some time orturn it into an immediate hard short.
The radio method has worked for me pretty well (to find series arcing, helped me to find quite a few dying switches and lose connections)

Maybe a question: Did youobserve any correlation between ow often and when the trips happen with things like temperature, humidity,or weather in general?

No, I haven't observed any correlation between tripping and the weather; the weather has always been very stable whenever it's tripped, and there haven't been any temperature "extremes" either.  If it trips again, I will have to remember to use the diagnostic test feature to see what "code" it is giving.  Would be interesting to know; I forgot to use that function the last time as I was just thinking about resetting it... 

An AF breaker may trip over several different things, which may be hard to distinguish. Here are a few :

 - Arcing in parallel with the line (phase to neutral or phase to earth) through a place where the insulation is damaged and a conductive track is forming. The idea of arc fault breaker is, that the energy of the momentary shorts happening in the conductive track may be below what it takes to trip a normal breaker's electromagnetic release, unless they develop into a full high current arc (The odds of which get lower when the line voltage is lower, as in 120V vs 230V)

 - Arcing in series with a load. This is one of the least reliable detection functions of AFCIs, as some series arcs dont have as clear signature (current peaks in the microsecond range and 10's of A that parallel arcs have). Some of them act more like a fairly stable resistance. There are some series arcs the breaker is meant to allow (like brushed motors), but there it helps that most such appliances have at least an X capacitor in them, which prevents most of the interference from the motor from coming out towards the breaker. (vs. a bad connection upstream, where all the interference is directly in series with the breaker). I imagine that the breakers will have some time-delay which is sufficient to ignore thermostats, as long as they click reasonably quickly

 - Some devices plugged in still have parts under voltage even when they are off. For example, the AF breaker might (i guess ?) respond to self healing events of an X capacitor in an appliance line filter, even if it does have a mechanical on off switch and is switched off

 - Interference that somehow match what the breaker is looking for. This might be induced both from within the circuit and from upstream (like from another circuit in the house, back through that circuit's breaker, panel busbar, and the AF breaker in question). I think some factors contributing to this sort of effects may be : Long circuits in general, especially circuits that extend outside of the house, or if the circuit wiring downstream is somewhere separated - as in phase and neutral wires taking different paths towards a point in the circuit. (This is not a proper way to install circuits, but had been found many times especially in lighting circuits that span between floors or across big areas of the house). The interference may come from many sources, including appliances etc, there have been cases of SW radio transmission in the area tripping AF breakers in houses...

As you mention your circuit, it seems to tick quite a few boxes of the latter....



How is the circuit laid out ? With what type of conductors ?



Disconnect all loads from it, remove any built in electronics (such as GFCIs, motion detectors etc) and perform an insulation test between all conductors (phase to phase, both phases to neutral, to earth, and neutral earth)

If you find abnormal readings, get to the place where the circuit branches to different locations, disconnect them to test the insulation separately and possibly locate the part of the circuit which might have bad insulation



Most AFCIs do have some low limit on the current they detect for parallel arcing. If the circuit is not in use (or you can make it not in use), you can watch it for lower currents - disconnect the phase end of it, and insert a low range ampermeter in series to monitor the actual current draw of the apparently open circuit. (Add an incandescent lamp in series for protection, in case some load appears which is higher than the ampermeter's fuse rating, or if the circuit indeed suddenly shorts)

An intact circuit will draw tiny (order of <1 mA), but most importantly constant current, coming from capacitance between the conductors and maybe distributed leakage currents for some insulation materials

A non intact circuit may (if you happen to catch it) draw higher current (order of few mA), and this current will sometimes tend to fluctate over the course of hours or days in a pattern - As a conductive track is slowly formed in the place of bad insulation, then at some stage it burns clear and starts over. (The addition of current limiting in series may affect the process as it limits the available current for the burn clear stage)



If the phase wire goes along with bare earth in a cable, or any similar setup where leakage is likely to be towards earth conductor or towards external conductive objects, then parallel arcing faults will also likely trigger GFCIs. If your AF breaker is not also a GFCI by itself, then adding a GFCI at the beginning of the circuit (after the arc fault breaker) will add an additional layer of safety, and might show whether there is indeed an earth leakage (the leakage may be intermittent)



You can try to swap AF breakers between this circuit and another, to see if there is some issue with the breaker. However, this test might be inconclusive. There might be a combination that's acting to trip it - There may be indeed something wrong with the circuit, but borderline, and the particular breaker might be with a little lower thresholds than another identical breaker. So if this test appears to solve the problem, you still dont know if it was interference from outside or there is really some very first stages of something happening in this circuit



If you can reconnect loads like groups of sockets permanently or semi permanently to other circuits, this will rule out parts of the circuit which may or may not be the problem (Just in case, do have an identical AF breaker on the circuit that you move the parts to). Also, it will reduce the amount of interference the circuit receives

Very interesting; thanks for the long explanation!  Nothing in particular seems to give it issues, and the circuit is so long trying to isolate anything would be like trying to find a needle in a haystack.  It literally runs pretty much the entire length of the house.  The bad (hackjobed) splices were fixed by an electrician the last time they were at the house, so at least I can rule out that one "portion" of it.  All of the wiring is copper and is a mix of 12 and 14-gauge (it's a 15A circuit on a 15A 2-pole AFCI).