They said the LED light bulbs would last 20 years. Last week I had to replace 5 of them that failed all at the exact same time, 6 month old bulbs. Clearly there was some kind of power surge or disturbance that the power electronics didn't like; but yet here we are.

The LEDs themselves likely will last 20 years (though losing some of the brightness). The power electronics that feeds them off 110V (or 220V) is tiny because it has to fit into the lamp's base, cannot be too expensive for pricing reasons, but is still much more complex than the LED crystals. It has a much higher failure rate than the light-emitting parts.

I'm sure that electric car manufacturers will have the same things to say when people point to previous claims of extreme longevity.

Does that matter if we get off of oil in the interim? It seems the argument should be for continuous improvement in the manufacturing process and a closed loop recycling system for failed components and batteries.

`let lifetime = max(...components_lifetime)` is like `let size = to_greeklish(base10(hard_drive_capacity))`. Some people don't care, but the people who do care are not going to thank you when they optimize on the basis of the appearance of a promise with an actual performance that is opposed to its appearance.

I've analyzed some failed LED bulbs, and what fails is the power supply. The LEDs themselves were likely still good, but especially the cheaper bulbs have inadequate cooling and very cheap & crappy power supplies.

LEDs last almost forever if they work within their parameters (mostly the temperature is important). They degrade quickly if they overheat. And above let's say 5W they will overheat if they are made out of plastic.

So - these plastic ones are designed to fail, so why put a proper electronics there? So they skimp on power regulator as well (and that's probably what killed your LEDs).

Buy LEDs with metal radiators - they are more expansive, but they aren't designed to fail in a few months and should really last decades.

This may be pessimistic, and relating back to the higher level comments on the auto side. Watching how current automotive manufacturers have gone due to the bean counters demanding higher margins - I don’t know that we can expect them to default to the better quality hardware. If they do, I’d expect it to be during the formative/adoption years and again decline toward higher margins.

An electric motor is not a cheap newly made LED light bulb? We know how long electric motors last since we have been using them since 1832. They seem to last a very long time.

I haven't had your bad experiences with LED bulbs either though, mine seem to be lasting great so far.

Tesla drivetrains had a high failure rate possibly due to arcing through their ball bearing causing pitting. The Tesla problem is likely resolved, but I’m not confident these (non-Tesla) electric motors won’t have high failure rates too. Other simple things like battery pack power relays fail more often than you’d expect, costing thousands to replace a ?$50? part.

I’m a few years into my first electric and would feel more comfortable with a solid supply of reasonably priced parts and service manuals.

Give it time. As long as right to repair isn't killed off by lobbying (John Deere can piss off), the secondary parts market will slowly catch up.

There is an inherent risk to working on these vehicles w/out proper training--just like working on your home's electric has risks if you don't know what you're doing.

While this is true with regards to the motor, battery, etc. the rest of the car is the same as standard vehicles and will fail in the same way as today's cars do. So if the AC fails, some computer board fries, suspension gets messed up, etc. you're still back at the shop.

Yes, but those components (mostly) haven't changed. They're also (mostly) not part of the "standard" maintenance which dominates ICE vehicles post sale profitability. Another reason dealers don't want to sell EVs.

Agreed. They’re also likely to be components a nondealer may fix.

The motors yes, but the power electronics components in motor controllers, MOSFETs and IGBTs, have a notoriously short life, especially if heavy power cycling with short t_on and t_off occurs.

Those things are pretty easy to replace with a hot-air soldering station.

Unless they're conformally coated.

Why would they do such a thing? Conformal coating is good for harsh environments if the electronics are actually exposed to the environment (like liquids), but in a car that's not the case, it's protected in a box and usually put inside the cabin too.

Also, conformal coatings are very, very bad for heat transfer. Power MOSFETs usually need to be heatsinked; conformally coating them would work against that.