SF Bay Bridge + 25,000 algorithmically controlled LEDs | Hacker News Reader

SF Bay Bridge + 25,000 algorithmically controlled LEDs

extremetech.com

54 pointsevo_932 comments

Similar large scale public art projects have had significant economic impacts for their host cities; the Gates project in NYC (which was up for only 2 weeks, compared to 2 years for the Bay Lights), generated $254m in economic impact.

http://www.nyc.gov/portal/site/nycgov/menuitem.c0935b9a57bb4...

Please consider supporting Bay Lights - its an amazing project .

Cisco must be happy about this. It's got to be the largest real world rendering of their logo anywhere!

This looks beautiful, going on the video. Its similar to light shows in other world class cities (Like in London, and the Eiffel Towel in Paris)

The ultimate screensaver!

Extremetech, please remove Onswipe, its only effect is slowing down loading when it's less appropriate.

ditto for every site that uses that piece of junk. Onswipe has to be the worst thing to happen to the web in years.

It's at least not as bad on an iPad 4 as it was on an iPad 1; it doesn't crash Safari immediately.

It still crashes my (still existing) iPad 1.

> [...] only $11,000 will be needed to power the display per year. That might sound like a lot of money, but an estimated increase of $97 million dollars in local business will probably change your mind.

$97 million increase from lights on a bridge?

Yeah, of course. Because of lights on a bridge, tens - maybe even HUNDREDS - of thousands of tourists are going to flock to the city over the course of the year.

What's so hard to believe about that?

But aren't all those tourists bringing in $97 million dollars just taking $97 million out of their local economies? People don't just pull out more money from their ass when something new gets built, they have to take money away from another spending habit to re-allocate it for the new one.

> But aren't all those tourists bringing in $97 million dollars just taking $97 million out of their local economies?

I don't think that local San Francisco businesses will be crying over someone spending $20 in SF vs. spending $20 in Podunk, Tx.

Also, it doesn't necessarily get pulled out of their local economies. The money I pay to AT&T, github, etc. isn't exactly powering local businesses.

Would be cool if they showed how much traffic was on the bridge, going across the bridge, or if there was an accident by flashing bright to those approaching the bridge so they could avoid it

The cynic in me says this will get co-opted by advertising in a wink. Imagine the Nike swoosh floating across the Bay Bridge.

The artist has 100% control for the first 2(?) years (and he has been vehemently anti-advertising - there have been numerous offers already). It will need to be reevaluated in 5 years when the bridge needs to be painted - as the lights will interfere with a new paint-job.

Does any hardware hackers out there have a source for the large scale switching boards that are at least semi-affordable?

I'd like to tinker with switching on/off few hundred leds at a time but haven't found a source for a board yet. Bonus points if it'll power RGB colored ones.

As a lighting designer my first thought was exactly the same, this is great, but how are you controlling it?

I image it's similar to the technology they use for those annoying huge LED billboards. I've investigated the driver boards used on those, but they are several thousand dollars.

Take a look at Buck Pucks or one of the many similar easy to use drivers whose names escape me at the moment.

Correct me if I'm wrong, but it's my understanding that the Buck Puck is just a single series driver that provides a fixed current.

I want to be able to wire up around 100 leds to a board and then connect the board to serial and/or USB and be able to change the switch on/off and ideally change the RGB of any of the leds independently.

When I was in high school (or shortly thereafter) I was working on a circuit that would use 74-series logic to let me clock out 64 LED states 8 bits at a time through my PC's parallel printer port, but got distracted by more pressing projects.

There are some serially controlled LED lighting strips, where each LED shifts its state to the next LED when it is given a new state. Thus, to set the state of a strand of 100 lights, you write a new state to the first LED 100 times.

Yes, I'm aware of the strips and have played with the ones from sparkfun. Those are really cool but I want to individually wire and control tiny led's onto a device. Picture a christmas tree that you could do effects on using the lights. Actually what I want to do is embed them into the plaster on a wall during a renovation to make it twinkle and change colors.

It's not at all difficult to build. However, if you really want the LEDs controlled individually, the problem will be the large number of connectors and wiring.

If we assume that, as in your other post, you want to connect 100 RGB LEDs, then you need 100 four-terminal connectors. Assuming you are using 0.1" spacing (I find smaller spacing connectors to be unreliable), each connector will require at least 0.5"x0.75" of PCB space (you need to leave room to be able to plug/unplug the connector). Next each LED will also require a mating connector and 4x whatever length of wire you need and a mounting method.

The first time I had a client who asked me to make up wire harnesses of 60+ wires, 4 feet long, it was a huge surprise just how labor intensive it was to build a dozen of them. And copper is heavy

Yes, I'm not looking forward to the wiring/soldering job that's for sure. An affordable switching driver is what I was missing, but have just found one. Thought I'd better circle back here and drop a link in case anyone else has a simular need: http://www.brilldea.com/product_LEDPainter.html

"but an estimated increase of $97 million dollars in local business will probably change your mind."

Curious to know how they calculated that number.

SELECT impressive_number FROM out_of_thin_air;

Exactly, although you need a function to make the number look precise. (ex. $97 million but NOT $100 million)

True enough.

It'd be very interesting to collect a large pile of estimates like these and see if they come even close to following Benford's Law[1].

1: http://en.wikipedia.org/wiki/Benfords_law

I would imagine that it's close to the inverse - a big pile of numbers that start with 9, and many many fewer that start with 1.

Why does everything in SF stop at midnight? This city is lame.

Because people go to sleep so they can get up and invent the future.

I'm really surprised and impressed that with such an installment only $11000 is needed to power it per year. LED is such an impressive piece of tech.

Ha ha whats funny is I thought the article said it would cost 8 MILLION DOLLARS to put some stupid cosmetic lights on a bridge.

I see they are on Causes but perhaps also using Kickstarter would help make up the remaining funding.

Reply on news.ycombinator.com