What Tech I Want, I Think (Early 2019)

It has been a while since I've written about a list of current tech items that have caught my eye.  I thought that it might make sense to given that CES 2019 just ended.

What I Want:

1. Humon Hex: it tracks oxygen availability in your muscles (specifically your thigh).  It provides more direct feedback on how hard you are working (which I'd always wanted to know).  Maybe I'm being lazy or just feel tired when my body is capable of doing more - this gives you a much better sense of how true that is.  Cost: ~$300.  Availability: Immediate.
2. Garmin Vector 3 Power Meter: I have gotten used to having legit power meter information from the Stages spin bikes at the Equinox at which I am a member.  Going for outdoor rides now are not as information rich.  I much prefer the outdoor ride (in the right weather), and already have speed and cadence sensors, so this is the next step.  It's a lot of money, though.  Cost: ~ $900.  Availability: Immediate.
3. K'Watch Glucose: A continuous glucose monitoring watch.  It requires a consumable which needs to be replaced every 7 days, but it looks super cool.  I'd love to wear for a couple of months to see how my activities, eating, drinking, working out and all affect my blood sugar levels.  I feel like I'd learn a lot over the first month or two of wearing it.  Maybe not too much after that.  Cost: ~ $150 for the watch, $100/month for the consumable. Availability: Unknown, but hopefully in 2019.
4. LIDAR Lite v3HP: A small 40m range LIDAR from Garmin.  I would love to work this into a Raspberry Pi 3 project to build a small device that would measure speeds on the road in front of me and then post the fastest on a website along with vehicle pictures (with ALPR, ideally) and a graph of the distribution of the speed of cars on the road.  Maybe someday I'll get to this project.  Cost: $150.  Availability: Immediate.
5. Naked Labs Body Scanner: I know I drink too much beer to keep my gut in check, but I think it would be super cool to get a long term view of where fat is being added and removed as you move through time.  So you should have a much better view of the aggregate body fat, and also the view of where it's coming and going from.  Also, you get your weight.  Cost: $1495.  Availability: Q219, I believe.
6. Air Quality Monitor: I'm not sure that Awair is the one, but I would like something that tracked indoor air quality that was inexpensive enough that I could have several of them in my house.
7. A compact point and shoot camera: a successor to my Canon S100.  Not sure what's best out there now, nor what the right amount of money is.  Sometimes, the camera on my phone just doesn't capture the magic.

What I Don't Want

1. MagicLeap AR Eyeglasses: Not until somebody shows me software that matters.  Cost: $2295.
2. 8K TV: Not until most content is 8k.  I bought my current 65" Sharp LCD 1080p TV in 2008.  It's still pretty great.  When it fails, I will buy a 4k TV of similar size (we'll see about OLED).  In fact, I'm rather hopeful that the Samsung modular micro-LED approach gathers some steam and we get some cool options there.
3. Bike head's-up glasses: Way too much money.  Way too early in the product maturation cycle.  Same for a motorcycle helmet head's up display.
4. USB turn-table: Maybe a cool mix of old and new, but I'm not a vinyl guy.  Spotify for me, right now.
5. Glowing alarm clock: My Garmin Forerunner 935 vibrating works just fine.  I wake up at a different time than my wife, so I think that lighting the whole room up would be disruptive.
6. Any voice-activated appliance: I have an Echo Show and I like it.  I just don't feel the need for a microwave that has this built in.
7. Any robot: call me when they are actually useful for more than vacuuming.
8. HTC Vive Pro Eye: but only because I don't have the space to devote to it right now.  Also, having bought the original Vive, I feel like there isn't enough software for it.

If I think of anything more, I'll add to the list.

Another Dead LED: Time for a Disection!

One of my first generation (for me) Feit "PAR" 30 LEDs just stopped working in my kitchen.  Which isn't a very good way of describing the problem.  It turned on, at super low brightness.  Not like L80 failure.  It all of the sudden went from a very bright bulb, to flickering occasionally, to flickering almost every time it was turned on, to not glowing brightly.  The bulb in question is about seven years old, but still disappointing that it is going.  Many other of these bulbs in my house are still going strong.

I'm not knowledgable enough to know what likely failed, but I thought that I would take advantage of the failure to see what was on the inside.  I did a destructive disassembly of everything but the circuit board.  I can say definitely that I did not expect the circuit board to be so complicated.  I'm half inclined to open one of my newer, much lighter LED bulbs to see how they've been simplified over time.

In any event, here are the pictures of the disassembly, step by step, with my amateur descriptions of what is there.  Further conclusions (if any) will be at the end.

This bulb is dominated by the aluminum heat sink that surrounds it.

Looking at the business end of the bulb - 7 actual LEDs

After removing the three screws you saw in the first picture, I was able to wiggle the top of the bulb, but couldn't remove it further.  Time to try the other end of the bulb.

The plastic lens popped off without too much effort.  It revealed four screws - three of which were required to remove the plastic insert plate.

Plastic plate removed, exposing the metal plate that has the actual LEDs sitting on it.  Some identification markings from Feit: ALPCB0589 Rev1 which doesn't mean anything to me, or the internet.  The positive and negative wires bringing the (presumably) DC power required to make the LED's go bright are exposed.  I had to cut them to keep going.

Now I was able to remove the aluminum heat sink from the bulb.  Here was my first surprise: the aluminum heat sink is only about half the weight of the bulb.  I had assumed it was the heavy part.  The core turns out to be pretty heavy too. 

To keep going, I had to start breaking the center core.  It was a hard, but reasonably brittle plastic that allowed me to dig down a ways.  I'd eventually get stuck just using my Nest screwdriver and had to bring in more destructive tools.

Just some perspective of the working part of the bulb - it is tiny, a super thin plate.

Maybe the plate is not metal: it looks like there is a super simple (in series) circuit on the surface?  Why would they choose to do it that way instead of in parallel?  Seems odd.

I've broken all of the inner core that I could with the screwdriver.  Inside there is silicon or some such substance around everything.  The circuit board is already way more complicated than I had expected it to be.

The other side.

Using some vice grips, I pulled off the E26 connector.

The vice grips were also useful to crack off the remaining plastic.  After that it was just clearing out the silicon.

Given the simplicity of the back of the circuit board, it came off in one easy piece.

Here is the top of the circuit board as much as I cleaned it off -- took way too much time even though you can see a ton of silicon still on it.  The right is the 120VAC input, the right is the DC (IDK what voltage) output.  I would guess that the exposed copper coils are a transformer bringing the AC voltage down.  I recognize some resistors and some capacitors elsewhere, but not sure what may of the other pieces are (like the three big brown things on the top and top right of the board).  I'm also not sure what the yellow covered device is (though presumably, it is an inductor).  In the end, this is almost certainly a constant current power supply shoved into the bulb.  It's just so complicated.

Why a constant current power supply needs multiple chips on the back is a mystery to me.

Here are my conclusions:

1. First-gen LED's were heavy due to both excessively large heat sinks and the silicon goop on the interior.
2. It's easy to remove the aluminum heat sink from Feit bulbs - do it and recycle.
3. New bulbs are not readily accessed like this one.  Likely they avoid both the big heat sink and the silicon goop.  When they start breaking, I'll dig in and see how they are different.
4. LEDs are pretty awesome.  Based on the marking here, I likely bought this bulb in the early part of 2009.  It's the early part of 2019.  I'll take a 10 year replacement cycle.  That being said, I'd gladly take longer.