The anatomy of a smart bulb

You get most of that with.good smart switches too. Reliability will never be quite as high, since it's hard to beat physically connecting and disconnecting the wires, but they're reliable enough that I've never had one fail.

But in terms of predictability and discoverability and fixed placement, a good smart switch is no different.

one issue with default on with power is that in the event of a temporary power outage, all of your lights will turn on.

Probably better than the alternative for most people but just something to be aware of.

> Apparently due to the way they work, in off state, there still will be a little bit of current leaking to the circuit

I've had this problem with dumb-on/off switches and LED bulbs. Stray currents are a known issue with LED fixtures in general.

Philips makes some bulbs that change color temperature to get warmer the more you dim them. Not as flexible as a smart bulb, but it gets you most of the way there.

Most hardwired smart switches close or open the circuit either electronically or with relays depending on supported load, so whether or not there is power to the bulbs is entirely dependent on the switch state. So yes, they can always match light state, ignoring the case of a burnt out bulb.

This applies to e.g. Lightwave switches.

Yeah, there's a lot of awful bulbs out there at the moment. Poor heat dissipation really does a number on the LEDs and smoothing capacitors.

On the other end of the price/quality spectrum, I have Hues that are still going strong after 6+ years of daily use. You can use them hub-free in HA using a ZigBee dongle.

It seems like a lemon market: https://en.wikipedia.org/wiki/The_Market_for_Lemons

I'm sure they know.

For sure they do. I spoke about these floodlights with the people in the shop I bought them at and they had a representative of the supplier come every month in a van to collect all the dead ones and give them new ones. I can't understand how they could sell them at such low price, replace 40%, and still make a profit. Perhaps they fix them themselves?

I noticed that virtually all LEDs driver resistors are two vastly different values, in parallel. Such that if you snip off the correct one, you are left with a slightly lower output light, that consumes half the power and last many time longer.

It's like a secret cheat code for people in the know. It's almost a conspiracy.

Just cut off the highest value resistor with a pair of snips.

Virtually all of them fail the same way. One(or more) of the individual LED diodes fail open. They are usually discolored as if they burned. I'm fixing them by replacing the diodes that failed (I desoldering them from other lights that failed). I always buy cool white Led lights and all of them use 1W diodes it seems (at least I think they're 1W, there is no writing on them).

I just purchased like $1500 worth of linear LED's from these guys for my workshop. I'm really sensitive to LED flicker and these were one of the only websites I found that specifically talked about flicker and claimed their LED's don't do it.

Does anyone else notice when LED's flicker at 60hz?

I was in a restaurant last week that had some of those LED's made to look like old time lights and the flicker in the room was terrible. If anyone moved their hand or the menu, you could see the strobe effect very easily.

I haven't. I should've probably mentioned I'm in Europe. That most likely excludes them for me.

I've had a look at their page, they don't have the fittings I get (gu10, E27 - probably these are not used in the US) and no European resellers as far as I can find (and no 4500K option, but that's a minor thing).

Although good that it's possible to get a free replacement, I suspect the broken bulb will still end up in a landfill or polluting some place in the world with more micro plastics. It's a real shame when manufacturers would rather make a worse product and pocket the few cents difference, environment be dammed.

This is the same in Poland (the country I moved to), but who can find receipts for light bulbs bought 6 months ago... I do keep them now (especially for more expensive ones like these Led floodlights). In fact that's how I got through so many of them. I bought them, a couple failed few months later, I'd take them back to the shop, they would just give me new ones in exchange. I'd install these, and they would break again in few months, rinse and repeat. I got so tired of having to swap them up high on a ladder I started fixing them myself and regulating the power supply for lower power (they have a brick style PSU).

Honestly that story is really not as simple as people suggest it is. There is a tradeoff between bulb life and brightness and standardization may have been good for the industry. Technology Connections does a thoughtful deep dive in to this story in this recent video: https://www.youtube.com/watch?v=zb7Bs98KmnY

>Those which fail are likely to fail due to inadequate cooling. With good cooling even large currents cannot damage the LEDs. As long as the temperature of the LED is limited there is no need to drive it less.

This seems like a logical conclusion considering lower power extends their life and the diodes that fail are usually discolored (as if they overheated). However, I'm not entirely sure it is as simple as inadequate cooling. All my high power LED floodlights are installed outdoors. I had 3 failures in the middle of winter when outside temperatures are way below zero (I remember these especially well, because climbing a tall ladder in freezing -20C wind is not something you forget easily). My theory is that they initially fail short that's why they appear burned. Also better cooling definitely extends their life, but it is not the only factor. The LED floodlights I mentioned were all housed in flat cast aluminum casings with find. The pcb is made of aluminium (with a very thin white insulating layer and copper traces on top). There is thermal paste between the back of the pcb that is raw aluminium and the case. I can't imagine a more favourable cooling solution. And still these floodlights are amongst the worst LED experiences I had. I tried all sellers available that didetry to charge >6x the price.

>I do not doubt your experience, but it is possible to choose better LED lamps

Perhaps it is location dependent. The lights I had a good experience with were purchased in the UK, but I've since moved to Poland. I'm happy to order from other countries within the EU (I even wanted to get the same LEDs from the UK again, but they don't sell them anymore).

The brands I see around here are either noname/picked up from thin air, kobi, samsung,philips or sometimes Osram. I have to admit I haven't got much experience with Philips or Osram. All the others are the same (including Samsung Pro). I would've bought more of the Philips/Osram lights, but none of my local stores carry them in cool white, ordering one or two online seems a waste of the shipping while I'm not ordering 20 without checking them out first... (if someone says "Philips is great" I'll get a 10 pack) Also I noticed people who like warm rather than cool light seem to have a better experience overall. Two of the best LEDs I have are warm, and they look like the famous Philips Dubai bupbs(I couldn't get any cool whites with the right fitting). They have been living outside for last 7 years, hooked up to a PIR next to a cat's food&water so they are activated a lot. I've replaced the PIR sensors, but the lights are still fine. Again, these were purchased years ago.

If anyone knows any good brands available in Europe, please mention them.

> I have bought 8 years ago about two dozen LED lamps

> All of those still work flawlessly

That's the point: you (mostly) can't buy LEDs made 8 years ago new. The LEDs I have from a decade ago weigh three times as much.

> "You have 12VAC, you can buy a 16V cap and that's all"

12VAC has an RMS measure of 12 volts but a peak voltage of ~17V. (I'm not saying that no one's done exactly that, but when they did, they were operating outside the rating in the first 1/120th of a second of operation.)

>A widespread DC socket standard for LEDs would go a long way towards this.

I can see California behind a tree rubbing it's hands together.

I had some "down lights" in an apartment I lived in in Australia that did this

Annoyingly for a single (small) room it had *three* switches, each controlling two lights. The ones on either side of the room were perhaps 8 feet apart, with two more at the near end of the room (about another 4 feet from the first two)

Replacing them with compatible smart bulbs of any kind due to them being DC proved surprisingly difficult in Australia. Though I suspect countries with larger markets would fare better

This is such a simple and obvious idea, I don't know why nobody does this. I guess they can't agree on a standard, but the first company to create this would probably also set the standard (and the second one would set it again).

Depends if you are calculating the costs in the summer or the winter.

My point is you are comparing to cheap garbage that took components and rare metals from around the world shipped on tankers to China where they were produced in harmful conditions, then shipped back around the world where they work for a year and are thrown in the landfill to sit for 1000.

The cost of those lightbulbs is not nearly as easy as the efficiency sticker on the box.

I'd be willing to bet the embodied manufacturing energy in a shitty Walmart bulb is greater than the energy it will save over its short 18 month lifespan.

Also frankly landfilled electronics are harmless.

Climate change is on track to seriously mess up the world over the next 100 years. In that time, a bunch of these in landfill will..still be there, doing nothing.

It's like the absurdity of worrying about land filled plastic waste: it's not actually the problem. It was oil in the ground before, now it's much less flowable oil in the ground again. It's the stuff getting everywhere else in the biosphere which is an issue, not material stably stored underground.

Which means they were 100% efficient during winter.

This isn't necessarily true. I tried to measure the PWM of my Cree and Philips smart bulbs, and what I found is that for RGB mode, they used PWM, but for white mode (which appears to use a 2700k led and 6500k led) they current dim both leds, and then use PWM to pulse the warm and cool leds on and off to achieve the target warmth. The Philips bulb actually syncs this so while the warm led is on, the cool led is off, and vice versa, so even though there is flicker, the total lumens eminating from the bulb is constant.