Tuesday, March 13, 2012

Techie Tuesday: Taillight LIES

As an update to my G.I.T. Lit lighting system, I'm issuing this warning to cyclists, that claimed operation times for taillights, are wildly exaggerated. Simple math you can do in your head reveals them to be overstated by 10-20X - about equivalent to claiming you can drive cross-country on a quarter tank of gas. You can prove them false by simply putting freshly charged batteries in your light, turning it on, and waiting.

The longest duration of operation was found on my Blackburn Mars 3.0 at 12 hours, followed by the Planet Bike SuperFlash Stealth (not the Turbo) at about 8 hours, and finally the Turbo at 5 hours. This represents the amount of time before the light's performance is so degraded it's no longer effective. All times are for constant on mode. Expected Flash mode times should be roughly double, but the Turbo's effective run-time is about 8 hrs. 48 hrs. (see UPDATE note at bottom)

This research was prompted by an incident that happened a couple of weeks ago, when I arrived home and found that 2 of my 3 taillights were down, which made sense, as a van seriously crowded me approaching a 4-way stop intersection. My pilot training prompts me to take such breaches in protection seriously. Over the last 2 weeks, I've been doing a lot of testing of my taillights.

Simple algebra reveals claims of 100 hours of use to defies physics. Totally, completely, utterly impossible. As ridiculous as claiming you can drive cross-country on a quarter tank of gas.

To be clear, Planet Bike's Super Flash Turbo is a great light, and more importantly, PB is NOT unusual in making these claims. Blackburn, and most other mfgs make similar claims, which are equally ridiculous, erroneous LIES. 

Far from the 100 hours claimed by Planet Bike, the Turbo is effective for 5 hours on constant. In addition, while employing a 1W LED, that primary LED  cannot possibly be using 1 watt in the taillight's circuit.  The claimed performance would require TWO of GeoMan's best LION headlight battery packs - over $150 worth at mail-order prices.

Two AAAs supply 1.8 - 2.5 watt-hours of power max. Using 1W, max burn times would be 1.8 to 2.5 hrs. The light may be employing a 1W LED, but it's utilizing less than half that in the circuit. In my tests, it was completely dead in 7 hrs. As a guide, go with 5 hrs on constant, and less than 10  60 *  flashing, as flash mode uses 2 additional LEDs in combination with the main LED at about 50%  10% * duty cycle.

Common practice in designing conventional, high-power strobes, is to use the bulb beyond 100% of sustained power for short bursts, which I believe PB has done with a twist - utilizing only 30% of the 1W LED in constant mode , and 100% plus  in flash mode, but for only 8-10% of the time. Non-LED strobes typically run 20-120%, in part to minimize thermal shock, and in part to allow tracking by eliminating light voids. LEDs don't have thermal shock problems, but they do have heat dissipation problems. Using the LED at high output, but for short durations, or a fraction of rated capacity in constant mode, eliminates the need to use an aluminum body to dissipate heat - a strategy that will eventually fail as customers demand brighter and brighter lights.

At full brightness, based on a Fenix flashlight killing an AAA cell in 3hrs, this light's primary LED produces 20-25 lumens in constant-on mode, and to be clear it is lumens output, NOT watts input, that is the relevant metric.Because LED emitters are quickly becoming more powerful, as well as more efficient, it's important to realize that your 1W light will be brighter in a couple of years than it is now, and your current light brighter than those of a few years ago, even scaled for watts of power consumed.

These measurements should be DOT-regulated, as cyclists are using them on public, DOT-regulated roads to deconflict themselves from motor vehicle traffic. This is a serious oversight on the part of DOT - due I'm sure to the speed with which cheap, effective lighting has become affordable to millions of cyclists. Until then, carry spare batteries, or end up FUBAR.

My testing was done with  Sanyo eneloop 750mAh, Sanyo 900mAh, and Duracell 1000mAh batteries, with the expected increases in burn time. The latter are no longer available, but Sanyo now makes a 1000mAh AAA cell. I would recommend the eneloop as they don't self-discharge over time, but if you're using these lights regularly, and recharging every few days, go with the 1000mAh for 25% more time.

In addition to truth in advertising, I hope the next PB light will have 7 additional features.

1.) An amber LED on each side with a small wide-angle lens to keep motorists at bay when closing and passing
2.) 7.4v LION power via 10440 LION cells, which store 2.5X the energy
3.) 7.4 volts, so compatible with my MagicShine light, and Shimano's Di2 system
4.) An option to power the light from internal batteries, or from an external 7.4v system, or both with a fail-over to internal power if the external power fails.
5.) A full-power, wide-angle lens, constant-on LED for Tracking in flash mode.
6.) A medium Pace-Line power mode for riding in pacelines
7.) Flat ribbon cables that stick to, and hug the frame

A lighting system should be judged by how motorists react to it, not technical specs. It takes (at least) 5 lights to ride safely in urban environments. 3 taillights and 2 headlights. One white headlight, and one taillight with 180 degrees of coverage should be up on your helmet to keep high-profile vehicles from looking right over the top of you and running you over.

* UPDATE: 3/20/2012
I've been writing Planet Bike and we've had a nice dialog. They have admitted that the Turbo's constant-on mode burn time is 6-8 hrs, but insisted the flash time was 100 hours. Based on testing with a DuraCell Alkaline AAA cell, I would put flasher time at 48 hrs. I let it run for 60hrs, and it still flashed, but so weakly it would been seen at maybe 100-250ft, depending on ambient light. Inside the 48-60hr interval, the flasher is more effective in dark conditions than a reflector, but probably not more effective in a light-saturated urban environment. A couple of take-away points here.
  1. The Turbo's flash pattern is excellent, drawing your attention better than any other pattern I've seen
  2. It does this employing a pattern that is remarkably energy efficient
  3. It's also the most effective light I've used in constant-on mode
Since you need a flasher to Grab motorists attention, and a bright, 180 degree coverage constant-on light for tracking, this makes an excellent choice for lights #3 and #4, and their interchangeability is a big advantage.