Controlling RGB LEDs with only powerlines: Anatomy of a Christmas light string | Hacker News Reader

Controlling RGB LEDs with only powerlines: Anatomy of a Christmas light string

cpldcpu.wordpress.com

152 pointscpldcpu51 comments

This is modern electronic design in a nutshell.

The large-scale physical design is as simple as possible (just regular wiring), the individual components are all "smart" (Christmas lights each with their own integrated IC, doesn't get more embedded than that) and addressable via a communication protocol.

Imagine how difficult the wiring would be to have 100 addressable lights that didn't use this arrangement!

You don’t need individual data lines to address each LED. A serial bus and a microcontroller would be enough. The only difference is that here there is no dedicated data line/lines.

1-wire is a similarly clever protocol.

I was getting at, if you didn't have embedded electronics in each light - there could be some arrangement of wiring to control 100 LEDs individually (e.g. a 10x10 grid).

300 LEDs, even, because they're RGB.

In 10 years every one of these LEDs will be running an IPv6 stack, and they'll still be using the same wire-and-glue construction.

All of them on wifi… Imagine the utter hell of living in an apartment building surrounded by neighbours blasting every Hertz of the wifi-spectrum with hundreds of cheap devices.

This is now, haha. I can see over 200 APs from my bed (1/2ish are xfinity wifi).

Fortunately my place is so small a 100’ cat-5 cable reaches any point in it, so I just use wired. Wifi barely works.

Considering wifi-6, but I can’t figure out if it helps with this type of congestion or not…

In 20 years each LED will be a pod in a Kubernetes cluster, running Node.JS and the output color will be controlled through React interacting with blockchain smart contracts.

And they will still flicker annoyingly.

Each LED will have its own NFT.

I dunno... I2C bus was designed in 1982 and it is still in the wide use today, and no replacement is in sight (for the "low speed inter-board" use case).

The hardware world moves much slower than software one, and the if the designs work, they are rarely replaced.

Most addressable LED strings have 3 or 4 wires, with 1 or 2 being data.

The ones in the article have only 2 wires, positive and negative.

What is interesting is that one could in fact do this with a single wire. You might eventually see something like this for micro-LED display solutions.

The key is to drive a current and control the voltage at both ends of a string such that each LED detects modulation and bypasses (itself) or consumes power (in-line). That might be relevant where the strength of a single (e.g. steel) wire was used to support a long string. By driving relatively low (max RGB) current at higher voltages and time (e.g. PWM) sequencing the light modulation.

In a simple case, imaging there are 100 RGB lights, half of them excited on each polarity and bypassed with a 0.2V Schottky stacking to +/-10V minimum (for digital data) at <100mA. You might synchronize and send data to the smart lights each "zero crossing" and get away +/-50Vpp for the string by only having 10 (12.5%) stacked (inline) on at any one time (a controller for the LED FETs can stay awake on a capacitor). As long as you run at a few kHz you won't have noticeable flicker, capacitive and you can update with a few bits per light each cycle.

More complex designs could use FETs to improve data link efficiency of the minimum voltage and higher current (lower PWM times) could substantially reduce peak voltage to +/-5V or so.

This string is only "slightly addressable" - in a string of 100 LEDs, you can't display any arbitrary image.

With small changes though, they could have allowed arbitrary addressing, which would have made these LEDs applicable for a far wider variety of products.

One way would be to deliberately make the string of LED's reflect the signal back from the unterminated end of the string. Then, make each LED only respond if the reflection and original signal add up past a threshold.

The controller can then have an initialisation step where it 'discovers' all the functioning LED's and their position on the line (by looking at the power draw when commanding them to turn on).

The downside is you'd be sending GHz bandwidth pulses down a poorly controlled pair of wires, and sharp bends and stuff might leak rather a lot of RF.

There's one part about the article that I'd like confirmation about: does this mean that the LEDs come out of the factory already configured as belonging to one of the 6 zones? This would be as opposed to, say, WS2812 LED packages where they are identical.

That's my interpretation, that they are randomly assigned:

> The control groups are randomly distributed across the curtain.

Thanks, I missed that line!

I'm really excited for the inevitable improved version of this that supports binary addresses and setting approximate PWM brightness levels.

Given the manufacturing process[0] linked, it could be a _little_ tricky to actually assign a known ID to each light in the strand, but either using the placement order when the strand is made or using a stochastic process to randomize the IDs and subsequently calibrate the strand should be viable, and result in a low cost strand with WS2812-like capabilities.

[0]: https://youtu.be/wCKlCUBsaT4

I love electronics and electrical engineering in general. But this article explains perfectly why it is less accessible than software engineering - the former requires a individual components, soldering iron(s), a digital oscilloscope, etc. while all you need is access to a computer with internet for the latter.

I think it has never been easier to enter electronics (especially embedded stuff) than it is now. You can get a starter set including RPI/Arduino/Breakboard/Components for less than $100. Also, with a breadboard no soldering iron is needed.

Try chemistry, mining, process engineering or even mechanical engineering in comparison?

We bought one of this kit for our daughter when she was ~10 year old. It has a the explanation of a few projects and the components to put in the breakboard. She liked it, and we bought some additional components, like a 7 segment display and leds of additional colors.

It's wonderfully easy to enter advanced electronics nowadays, but has accessibility really kept up with the rate of advancement (electronics getting smaller and more complicated)?

Totally agree. Electronics has never been easier thanks to low-cost microcontrollers like the ESP32. Most of the "electronics" you'll be doing is simply plumbing--connecting the GPIO outputs of your microcontroller to whatever peripherals you want to support. It's a good gateway into the wider world of electronics.

I am curious what the next thing up is that is more "legit". At this time I've used Arduino/ESP/Teensy. I bought an STM32 and Beaglebone but wondering what does something like a Boston Dynamics dog use? Probably a full on processor like i7.

Edit: oh it uses i5 with Ubuntu ha, plus separate graphics for the AI part

Chemistry takes some initiative these days, but the dark age after the golden age of chemistry sets is pretty much over.

Mining, I dunno, buy a shovel? Hobby prospecting - sluicing and panning for gold, mostly - was still a thing in the rural west when I was a kid. I'd bet you could find a forum or something if you searched.

Process engineering, I don't know what you'd want to do as a hobbyist.

Mechanical engineering - 3d printers and chinese machine tools are affordable. It's never been easier to buy small parts and fasteners. The free finite element simulation packages are getting pretty good.

Made me think of Cody's lab guy's sweeping some dirt from a highway and pulling palladium out.

>why it is less accessible than software engineering

Having experience in both, and started at rather young age (pre-teens both) - the hurdle is the "magic smoke" (that all electronics devices work with). Mistakes are very punishable with components and power supplies blowing up, while software mistake are extremely forgiving... unless you work on an 8bit 6502 that doesn't have properly implemented RST signal and cold restarts the entire machine instead, wiping off its memory, on top of lacking any permanent storage.

Mistake and learning electronics costs resources and time each time, and it's a lot more frustrating. Also bread boards (nowadays) don't really need a soldering iron.

you can always play around with the simulators

https://falstad.com/circuit/

As someone who tried to use falstad, spice, etc to learn as a kid, simulators never really worked for me pedagogically. I was only able to learn by using physical devices. Nothing really beats a breadboard for immediate feedback (whether reward or punishment).

The video of the manufacturing process is quite interesting. A lot of human labor is involved.

One downside to this type of Christmas lights is that these design parameters — i.e., make it possible to manufacture feature-rich novelties with a minimum of wiring and changes to the assembly line — result in a product that has no user-serviceable parts: when a few LEDs die, the whole product should be thrown out.

I say 'should' not because I think so, but because this is literally what is written on the labels affixed to many such products. We went from lights featuring incandescent light bulbs which came with a few spares in the box (anyone could replace those), to LED lights with a few spares (finicky to replace, but doable), to almost exclusively fixed-in-place LED lights without replacements and labels attached that tell the user to throw the chain out when lights break. And now more and more complexity with integrated ICs.

Only those of us with a soldering iron can salvage some of them, and this new category seems like it might not be repairable at all (excepting the really clever EE folk like the author). Environmentally speaking, the increased complexity and focus on ease of manufacturing of such products seems to reduce their lifespan as well, with repairability already at a minimum.

This is all true, but people voted for it with their wallets. At all levels, people understood that when their electronics break they would be paying for it again, making the ownership of the product more like renting with recurring payments. People understood the term "planned obsolescence". But they saw the cheap shinny blinky lights and decided they wanted it. The consumer is far more guilty than the company that gave them what they wanted.

I don't think people understand the trades they're making. They see that one offering seems to do approximately the same stuff as another one, but it only costs half as much, and don't understand that the cheaper offering is one one millionth as fixable as the more expensive one. By the time it fails, they forgot the circumstances under which they bought it, so they may never connect cause and effect.

As a consumer, I find string light packaging almost useless to estimate (1) lifespan until it breaks, or (2) possibility I can repair it when it does —therefore I hesitate to spend more than I have to on the theory that at any price I’m lucky if string lasts more than 2-3 seasons.

Given the lack of wage increases in many parts of the world perhaps the consumers, as a whole, aren't buying the better stuff because this is all they can afford.

if I have a choice between a $10 X and a $20 Y, is it necessarily the case that Y will be a better product than X, last longer than X, and/or result in more wages being paid to the workers that made Y?

Aside from the points noted by other commenters about this being a choice most consumers cannot reasonably be expected to make, the notion that there is choice at all is simply false. Almost all Christmas lights sold in shops last year were of the current generation of fixed LED light strings sold in a plastic tub. The year before these were sold mostly in cardboard packages. For the basis indoor Christmas tree light product, this accounts for 95% of the product available, and almost all of the offering in physical stores¹.

There's no reasonable chance of voting with your wallet against those numbers. Sure, I might be able to get some fancy hand-crafted Swedish Christmas lights with a ten year replacement part availability guarantee on some website, and who knows, perhaps a very effective word-of-mouth campaign will make them quite successful! — and still not make a dent in the global market. Realistically, it won't have any measurable impact in the greater scheme of things. This product is driven by sellers and manufacturers going for the absolute bottom in price because it is technologically feasible, externalities (like the environmental impact) are not part of the cost, and most consumers have plenty of other worries and simply crave a bit of Christmas cheer.

1: Here in the Netherlands, although likely globally applicable.

If we must assign guilt, one the one hand I do kinda blame people for being thoughtless consumers, on the other we've been subject to a relentless and ubiquitous system of commercial propaganda (advertising) promoting that lifestyle.

What's the way forward?

> result in a product that has no user-serviceable parts

You seem to think this is a step backward?

If you remember old lights you had bulbs that regularly blew, and the bulbs had to be thrown away and replaced. If you couldn't find compatible new bublbs you had to throw the whole thing away. The LEDs are not replaceable... but they hardly ever break either.

The LED lights product lasts longer, and doesn't consume bulbs. Strictly better, isn't it?

> The LED lights product lasts longer

Not my experience. My living room lighting already blew up two times and no, it wasn't the transformer which would be easy to replace.

> and doesn't consume bulbs.

Instead I have to throw away the whole thing with all its coper inside.

> Strictly better, isn't it?

Not in my books

The old product was designed to use consumables even if it was functioning correctly!

I bought christmas mini LED string lights last year from home depot, specifically labeled for outdoor use. Took them out this holiday season, and over half of the strings didn't work anymore.

They have replaceable "bulbs", but HD doesn't sell them, "because they are never supposed to break".

Found replacements that sort of work (different shade of white) a another store for some of the strings, but not for others, and had to throw them away. Very sour with the experience.

I had that happen but all of the issues were related to the individual led bulb connectors getting loose and were easily fixed by reseating them. Dunno if you had a similar issue but wanted to mention it.

>My living room lighting already blew up two times and no

The reason is not the LEDs themselves but... money. The LEDs are overdriven to show higher lumen numbers (and wattage) on the box. That results in overheating both of the LEDs and the driving circuits - one of them fails. In short it's the good old - planned obsolescence.

It's absolutely possible to have LEDs that last for 50k hours (while losing 20% of their brightness) but don't fail. Sometimes even adjusting a resistor value (to reduce the current through the LEDs) would do the trick.

That could be an interesting hack to the power supply, turning down the power a bit.

Not the kind of thing that gets people motivated though, "let's make this product less powerful" isn't a common modding activity.

>let's make this product less powerful" isn't a common modding activity.

Actually it is for 'consumer' LEDs products, like a top 1. Undervolting GPUs and laptop CPUs is rather common too.

Big Clive does this sort of thing on his YouTube channel fairly often. A few examples:

* Adjusting a current regulator to drop a 10W lamp to ~2W - https://www.youtube.com/watch?v=YeWnHnZ-ojw

* Adjusting a capacitive dropper-based lamp from 5W to 0.5W - https://www.youtube.com/watch?v=RKRz7EplgDE

* Making a "Dooby" (inspired by the Philips Dubai lamps[1]) lamp - using a series capacitor to make various lamps run at 0.1-2W - https://www.youtube.com/watch?v=ISTB0ThzhOY

[1] They're designed to have a higher energy efficiency and a longer operational life - https://www.youtube.com/watch?v=klaJqofCsu4

Is there something wrong with the power in your house? Maybe slightly over/below the nominal voltage? I've been buying only LED lamps and bulbs for the past 10+ years and literally never had a single one break for me. Not even the stupid GU10 ones which get stupidly hot - they just work and work.

And yes, I'm sure many of us remember pulling out an old string of christmas lights, only to discover it doesn't work and then spending two hours looking at every single bulb close up to figure out which one is blown, and pray any of the replacements work. The LED strings just work forever.

Voltage is stable around 230-233 volts. There may be power spikes I cant reliably measure. I also do not operate big inductive loads (motors) which cause power spikes when stopped.

Also I never had those issues in the past with other appliances. But alas it is anecdotal and someone has to be statistical outlier.

This is just anecdote versus anecdote, but I've never had to replace a LED because of them blowing out. Just recently, I replaced a LED that has been turned on for 24 hours a day every day for a whole 5 years, and it was still working. You may ask, why replace it then? well, high quality LED bulbs might not blow up, but they lose brightness over time. Still, I'd say 5 years of mileage for a light has is never turned off is a solid return on investment, particularly when considering how little energy they consume compared to previous light bulb technologies.

The same thing also happens with the LED strips within monitors, monitors get replaced when they become too dark.

I would be suspicious about the quality of the electricity in the place you live or work at when you see this stuff break at too high of a frequency. Not directly related to LEDs, but somehow I had electronics like computers or just plain old regular incandescent light bulbs break far more often in a house I used to live in, and I used to blame bad luck with the devices I bought, until I moved into another apartment and things actually started to have acceptable life expectancy.

Don't forget those old incandescent strings could draw up to 100 watts each!

> And now more and more complexity with integrated ICs.

I dunno. Which is more likely to fail? The older electromechanical "blinkers" they used or a solid state IC?

I do get it though. Society is addicted to disposable things. But I'm not sure the problem is these things being non-user-servicable. When these things go, they go... Usually the cabling gets fucked on them long before the LED's.

Reply on news.ycombinator.com