I can’t solder to save my life. I’ve just proven it.
So after some more advice, I went out and bought a new tip, some tip refresher and a brass “sponge”. After forty-five frustrating minutes I’d managed to fuck up half the joints I attempted, destroy the brand new tip and give myself a headache from inhaling the fumes.
Clearly, I need to find another solution. The FCI “Clincher” connectors look good except for two things: to get a good crimp, you need about 1cm of of the cable inside the crimp part housing. The exposed solder pads of the LED strips are 1cm long in total, so for every usable length I have to sacrifice three LEDs. Second, the crimp tool costs over USD$2000! Buying one’s obviously right out, and god knows where I can rent/borrow/steal one in Singapore.
Solutions on the back of a postcard please…

I can’t solder to save my life. I’ve just proven it.

So after some more advice, I went out and bought a new tip, some tip refresher and a brass “sponge”. After forty-five frustrating minutes I’d managed to fuck up half the joints I attempted, destroy the brand new tip and give myself a headache from inhaling the fumes.

Clearly, I need to find another solution. The FCI “Clincher” connectors look good except for two things: to get a good crimp, you need about 1cm of of the cable inside the crimp part housing. The exposed solder pads of the LED strips are 1cm long in total, so for every usable length I have to sacrifice three LEDs. Second, the crimp tool costs over USD$2000! Buying one’s obviously right out, and god knows where I can rent/borrow/steal one in Singapore.

Solutions on the back of a postcard please…

I’ve finally got the RGB LED strips I bought from Sim Lim the other week to work. The secret? The damn pads are labelled wrong. For future reference, if anyone else is having fun and games trying to get these strips working (the only marking on them is “ZJ168”), this is the pinout:

Di = +12V (Arduino pin Vin)

Ci = MOSI (+5V, normal logic level—Arduino pin ~11)

St = Clock (again, +5V—Arduino pin 13)

Li = GND (Arduino GND, obv.)

The 12V and ground are on the far sides of the strip, with the data and clock lines in the middle (as I thought made sense originally.)  They’re LPD6803-based and work great with the Fast_SPI library written by Daniel Garcia who’s been graciously helping me out via email.

Next steps are cutting this little strip up into a “demo” display (maybe t-shirt mountable) and writing a little code to transform 2D coordinates into a linear pixel address. I’ve also learned a little bit about soldering technique and, armed with a new tip, will reconsider my options for putting everything together (maybe using 0.1” header sockets rather than Clincher connectors.)

I made an technology

Not strictly true, but I spent an evening porting this javascript code for converting between HSV, HSL and RGB to an Arduino library. I’m not sure if I should be embarassed or relieved about how much C++ syntax I’ve forgotten since second year.

Now that’s covered, the next part of the project is to get a 60Hz interrupt reading those RGB values out of some kind of buffer and writing them to the LED strip.

It’s alive! Or at least, it can hear. My little mic-and-amp breakout boards were delivered on Thursday and now I’ve got the Arduino reading from one and triggering the LEDs. The ultimate aim is for the project to be able to automatically sync the light sequences with the beat of the ambient music.

I’ve moved my little 3-channel LED switch from the breadboard onto a much neater little slice of Veroboard. And bugger me if the damn thing didn’t work first time, too!

Crimping all the connecting cables is going to be the real problem though. It’s damn hard to do without three hands. Have to think more on that one.

Crimpin’ ain’t easy

Since I received my FCI “Clincher” parts from Element14 today, I headed down to the Singapore Hackerspace after work to commune with my fellow nerds and do some more work on the my Burning Man project prototype. After a bit of experimentation I found that the nose of some linesman’s pliers was the best thing to crimp down the Clinchers’ contacts with, and if I tried to bite the crimp down further than the exposed copper, it shorted the red and blue lines, so I could only get green, white and various shades of purple. Unfortunately I buggered up at least one end of all of the five strips while working this out, but this is why we make prototypes.

Next job is to wire some proper cables and maybe solder up a little board with the “power” electronics on it, rather than the gimcrack arrangement of test leads and protoboard I’m using now. The help I’ve receieved from the Electronics and Robotics Stack Exchange site has been invaluable in getting me this far in my journey.

Apparently all you have to do to get a rainbow sequence is traverse the edges of the RGB colour cube. Who knew

In other news, the flame effect is ready for testing. But where? It’s almost certainly 100% safe, but an ang mo fiddling with wires and pipes and little balls of flame anywhere in public view is probably going to have a very exciting evening.

Anyone got a shed?

Burning Man 2011 project #1 finished!

I still need to get the battery sorted but I’ll do that when I get over to the States (batteries look like Semtex to airport luggage scanners and I can do without that excitement.)

These are the bike underlights. Couple of LED strips and 700mA constant current drivers from Deal Extreme, encased in perspex pipe and sealed with E6000 (a recommendation of the fantastic Dr. Jon’s Guide to Electronics on the Playa.)

They’re really, really bright and hopefully they’ll reflect well off the playa surface at night.

… and from this point, it’s all software. And I can do software.

My pulse widths, they are modulated.