UPDATE: My basic premise of "cost/benefit" and most of "why specifically window air conditioners" still applies. But after some helpful comments from @jay613 and watching this teardown video I have a much better understanding of exactly how an LCDI works. It monitors for leakage current using a sense wire but does not specifically check for the same problem (a ground fault) that a GFCI tests for. I have a hunch (but no time to research at the moment) that a circuit GFCI (e.g., GFCI/receptacle or GFCI/breaker) will in fact trip under many, but not all, scenarios where an LCDI would trip. More later, when I get a chance to research a little more. For now, I will
strike-through the parts that I expect to change.
I believe it comes down primarily to cost/benefit analysis and the type of usage of window air conditioners specifically.
There is a cost to almost any technical safety improvement. Over time those costs go down, but initially they are quite significant. This is certainly the case with GFCI and AFCI protection. So there is a balance between cost (how much to install in every necessary location) and benefit (how many fires, injuries, deaths (yes, lives have a cost in the world of law and government regulation), property damage, etc.). If it was not for the cost:
- GFCI would have been mandated for all areas with water ingress concerns as soon as reliable devices were available - i.e., not just kitchens and bathrooms but all the later areas - laundry room, unfinished basement, garage, etc.
- GFCI and AFCI, as appropriate, would be required with every panel upgrade, circuit change, etc. Grandfathering older stuff slows down general safety improvements but at a major cost savings - and in fact if the costs were added in then many otherwise very important safety upgrades might not get done, or would be delayed until there was no other choice.
There are certainly many other examples, and not just in residential electrical but in automotive engineering and other fields as well.
An LCDI costs more than a GFCI (actually, not sure - I suspect they did initially but with economies of scale a plug-LCDI or a plug-GFCI is likely comparable cost) which costs more than a simple cord and plug. So GFCI gets mandated (many years ago) for the most critical things - hair dryers and other appliances used by wet people around bathroom sinks, and later gets added to some more appliances but never, due to cost, gets added to low-risk appliances - lights, battery chargers, TVs, computers, etc. The high total cost for adding that $5 GFCI to all those devices that have extremely low risk of being used in a ground fault situation, combined with increasing implementation over time of GFCI protection in the power supply (GFCI/receptacles and GFCI/breakers) means that it just doesn't make sense to do that. All the more so with an LCDI that costs more than a GFCI (the device itself is more complex and it requires a more expensive cord).
Window air conditioners are a special case where LCDI has additional benefits:
They produce water. Few other household devices do that (dehumidifiers). People are plenty aware of (after many years of GFCI implementation in kitchens and bathrooms) of the dangers of "water + electricity", but many don't think about that when it comes to a device (air conditioner) that is not used with water (like a coffee maker or washing machine) and not used near water (like a toaster in the kitchen or a hair dryer in the bathroom). Actually, they do produce water, but that isn't what LCDI deals with.
- They use a lot of power. HVAC in general (window air conditioners, space heaters) are the largest plug-in devices for most people that run for more than a few minutes at a time. (Kitchen built-in appliances may fall in that category, but in most cases if not hardwired then the cord/plug is hidden so it doesn't get moved around and it doesn't even get touched by people in normal use.) No update here - because they use a lot of power, a cord problem is still a big issue.
- Many, but not all, users will move the entire unit twice a year with the change of seasons, which lends itself to cord/plug problems much more than with something like a refrigerator. Again, no update - still applies 100%.
and what I suspect is the biggest factor of all:
Window air conditioners are used in the oldest buildings. If a building gets a major renovation, HVAC upgrades are usually part of that. For new small buildings (shed, workshop, separate garage), a mini-split makes sense. New larger buildings (typical single family homes) have generally had air conditioning built in for decades, going back to before GFCI was even generally available (let alone required). Even many older homes that did not have central air conditioning when built had forced air heating so adding central air conditioning was relatively easy to do. But window air conditioners tend to be installed:
- In very old housing (especially 1940s and older) which has had only minimal electrical upgrades (e.g., a new circuit for the air conditioner - maybe, if the landlord can get away without it then not even that) and no (or few) GFCI-protected circuits. Even if owner-occupied, the cost to add central air conditioning is very high so that upgrade will fall behind relative to other more critical problems, which happen a lot with older homes.
- In rental housing with low-income tenants - the landlords only put in upgrades required by law and the tenants can't afford to do any upgrades themselves
- In rental housing many tenants consider themselves short-term (e.g., college students) so that even if they have funds available they are unlikely to put them into any major upgrades since the upgrades would revert to the landlord when they leave
So window air conditioners present a unique combination of
water (yet without being connected to the plumbing system), high power, movement and installation in older buildings.