For bike under-lighting at the Burn this year,

For bike under-lighting at the Burn this year, maybe I should ditch the crazy 10W-RGB-emitter-and-custom-MOSFET circuit idea and just use a couple of strips of 144 LED/m WS2812B. As much if not more light output, and a known quantity when hooked up to a cheap/simple 32u4-based Arduino (Beetle?).

Thoughts?

My own vote is for the LED strips. Light strips let you color your surroundings with patterns much easier. Besides I can glance at a single 5050 led and suffer no ill visual consequences. Glancing into a 10w LED leaves me glare blind with an LED pattern burned into my retina for at least a few seconds.

I’m ALL IN FAVOR of prefabricated, modular, in-stock, replaceable parts. And standard connectors. And having replacements on hand.

I’m like the Eli Whitney of LED projects.

Well, the point is there would be about 70 of the 5050s all going off at once (but pointed at the ground.)

Sounds so much easier though, I think I will go with them…

(with the added side-benefit that if I don’t get the project finished I can re-appropriate the strips to something else much more easily.)

Beetle has no on-board voltage regulator. For durability and survivability I might go with a Nano, which has one. I’m just being paranoid here. Really, it depends on how you’re going to power everything.

12V battery -> buck converter?
A pile of AAs?
Glow sticks strapped to a solar panel?

I’ve got a 12V SLA battery as the basis for the bike power system, and I just added a couple of 12V-5V 25A DC-DC converters to my Alibaba order. That ought to be stable enough? (But thanks for the heads-up, I hadn’t thought of that.)

Cool.

Also: Dan and I keep finding that a little below 5v is totally fine, but a little above 5v can send everything to deep playa in a whiteout without a camelbak.

We’ve seen things go bonkers with as little as 5.2v.

Best practices might include putting a high-current diode in series with your “5 volt” power supply. The forward voltage drop will be about 0.7 volts, and everything should still work fine at 4.3v.

But basically, I’d say measure the actual voltage coming out of your converters and if the voltage is even a whisper above 5v, drop it like Skrillex.

Yeah, my observation with my panel was that everything seemed happy enough and I’d get perfectly saturated whites as long as I stayed above 4-and-a-bit-V.

(+1 for “drop it like Skrillex”. Or possibly -1. I’m not sure. What is a “Skrillex” anyhow?)

So - having done more testing, i’m pretty sure that what’s happening when the voltage is above 5v is that the chips become more sensitive to timing slop. I think i’m going to need to spend some more time with a scope measuring, well, time. I think the problem might be that i’m sliding hi just a little bit too long when writing out zeros, which causes them to be interpreted as ones, which results in madhattery. At least, in theory. At least in one case that I was testing. (Though, with 3.3v outputs on the MCU, going too much above 5v probably does fuck with the led chips’ ability to discern between high and low voltage…)

I have really good luck running below 5v almost all my projects run on a single li-ion battery.

(And for those who may not know, lithium batteries typically deliver 3.7 volts.)

With my POV project, I use a single 3.7V Power Cell for the LEDs and drivers, and a small 5V boost to the controller. Li-Ion Power Cells are different than regular Li-Ion in that they can burst a much higher amount of amps when needed.

Cool! Would you post some links to what you use?

http://www.all-battery.com/Li-Ion18650_3.6V_1500mAhBattery-30256.aspx - look at the max discharge. That’s not a typo.

Oh, the actual max discharge for the 1400mAh version is right around 15A, the newer 1500mAh one is 20A. I got that straight from my Tenergy rep. (Funny story, friend of mine was looking to buy these batteries and kept complaining that everyone was out of stock. I said, ‘Let me contact my rep at Tenergy.’. The next e-mail from him was, “You have a rep at Tenergy? Where do you NOT have a contact?!” Made me laugh. And yes, I have a brand new set of batteries sitting at home right now.)

15A…wow.
Do you have a way of charging them in situ? Or do you remove them to charge them?

I build my circuits with a MAX8903 for charging. It accepts both USB as well as DC-IN for charging. So the battery never has to be removed from the unit … or rather, shouldn’t have to, unless it got damaged somehow.

Now what we need is a long, skinny, custom little board with:

  • DC-in barrel jack (power in), and USB port (power+programming)
  • MAX8903 charge controller
  • battery clip to hold an 18650 cell firmly in place
  • ATmega32U4 (or whatever)
  • a couple of solder pads for a few ATmega data ‘pins’
  • and at the other end, solder points for power and one data out line, suitable for direct connection to one LED strip.

Ideally, there’d be solder pads for wiring in a hard power switch, but which you could bridge across for continuous operation.

power/USB==> [jack charger battery ATmega pinpads LEDpads] ==> LEDs

Two thoughts:
(1) since power is more of a pain in the butt than wiring the LEDs, having a high-amp Li battery, the clip for said battery, and the charging circuit all in place would be actually a huge win.
(2) Did I basically just describe the boards you’ve already built, @Ashley_M_Kirchner_No ?!?