I'm into starting threads on lights - so here goes!

Whilst it seems the Niterider lights are well made and work well, I suspect the digital control is a waste of money. Due to the way light bulbs work, you'll get less than 60% of the full light output when running a 12W bulb at a 9W setting (ie equivalent to a 7W bulb), and even worse running at 6W. If you really need these settings for longer run time, you'd be better off having another lower wattage bulb, which would give better run time for the same light output (or more light for the same run time).

But surely the point is, during a ride you may need a wide range of lighting. Some of the ride might be on road, possibly through town, to get to the trails. You'll only need a 6W for that sort of riding. But when you get to the trails you'll need something more powerful. Surely swapping bulbs isn't very practicle when your out riding?

What Chris is saying is that the 6W setting consumes 6Watts of electricity but produces a lot less light than a 6W bulb. Or if you produce the same light as a 6W bulb you'll need more than 6W of electricity if you know what I mean.
All the digital electronics do is reduce the voltage on the bulb which is a really innefficient way of reducing power output.
If you got some lumicycles you could have the usual 12/20W set up and for an extra £25 mount a third 5W lamp unit for the road sections for the same cash as the niteriders.
You can't "turn down" the power of a bulb by lowering the voltage across it in a linear way.
I.e. halving the voltage will halve the power consumption but might provide only 1/3rd of the light - so you are throwing away all that electrical energy of that very limited battery source.
The undervoltage options only seem to make sense in the "get you home when your battery is almost empty" option.

If you run a 12/20W lumicycles then you already have 3 illumination levels possible.

AFAIK the NiteRiders use PWM so the 6W setting illuminates the lamp at 12W but only for 50% of the time.

Disclaimer: Not an electrical engineer.

Aaah!
If so, very clever!

I would of thought for that money they would be using PWM. Which is as Roadkill says turns them on for a percentage of the time dead quickly. Which means effectively less voltage is applied to the blb without wasting any power. I'm not sure though but perhaps a 12V bulb with 6V applied to it produces less light than a 6V bulb with 6V applied to it and therefore is less efficient.

I wonder how variable that PWM is can you tweak it up and down as much as you want? it would be nice if you could choose whatever voltage you liked and overvolt 'em if you wanted to.

But why don't they apply themselves and put a dead clever charger in there which monitors the individual cells of the battery and charges them appropriately. Sure you'd need a couple more wires but the batteries would probably preffer it. They have to talk to me to get me to pay that kind of money though.

I obviously mean the lights would have to talk to me. Which would be pointless but at least then they'd be proper night riders :-)

It does use a PWM, Roadkill (is that your real name? - you must have funny parents). However that doesn't illuminate it for 12W half the time. You see bulbs work by heating, and if you only put current through half the time (as on a 6W setting), the bulb doesn't heat up as much, and because of the non-linearity I mentioned before will only put out about 30% of the light. It will tend to stay at a fairly constant temperature, as it's response time is far longer than the frequency of the PWM.
I tend to agree with Kevin's points - for less money than a Niterider digital you can get a Lumicycle with an additional 5W bulb, which should be about as bright as the 9W setting on the Niterider, but burn for much longer.

Typical, someone else replies whilst I'm typing out my response. Unfortunately Phil, it's not possible to use a PWM to increase the effective voltage of the pack, so you're stuck with the basic voltage of the pack as your maximum. However I believe Niterider do already overvolt their systems.

If you're really interested, you can buy a PWM, and in effect make your own Niterider digital system from any plain ordinary system - I got mine from Willie Hunt, who makes the only really efficient ones I'm aware of:
http://www.cs.indiana.edu/hyplan/willie/lvr.html
With his systems you can use a push button to dim the bulb, but that's not an option I use. Mine is used with a Lumicycle pack which I got built up with an extra cell (nominal 14.4V), to drop the effective voltage to 13.2V - this gives a constant brightness to the bulb through the run time (without over stressing the bulb at the start, and dimming at the end), and increases the run time slightly over the standard pack.

WARNING : Electronics student in training

Correct you can't use PWM to increase the voltage I was figuring either using a higher voltage battery or lower voltage bulb, which every method Lumicycle choose. If you wanted to increase the voltage you could try a Buck-boost converter, then you might aswell go the whole hog and slap a metal halide on it. Hmmm.

Now excuse me I must go and revise.

Like others I didn't post fast enough but here goes any way

I like Kevin's idea. You get loads of exciting of power combinations from turning on offer various light. 6W, 12W, 18W, 20 W, 26W, 32W, 38W

The turning the lights on and off idea sounds a bit too clever to be true. I think unless you did something very clever you would come up against the same problem outlined by Kevin.

I think that the problem is that the reduction in efficiency is because you are running the filament at a lower temperature. At a lower temperature the filament gives out a greater proportion of its energy as heat as opposed to light. I don't think it's possible that the filament is taken up to its most efficient temperature during each pulse of electricity, I don't think the filament could change its temperature that quickly.

I guess where you would get an increase in efficiency is that if he reduced the voltage across the bulb by connecting the resistance in series with it he would be wasting energy heating the resistor.

You've hit the nail on the head there John. PWM is only an efficient way of reducing the voltage. One thing that's just struck me is, if bulbs are manufactured to be most efficient at their specified voltage then why overvolt them. Surley a bulb designed for 13.2V is going to be more efficient/brighter than one designed for 12V. Or are we just moving the life expectancy goal posts?

The hotter the filament the more efficient, presumably desidn voltage is the manafacturers efficiency design life compromise for domestic applications

I've just realised that there seems to be about 100 people far more qualified that me to write that battery chicken article!

As a side note, the NiteRiders do have built in smart charger circuitry. Although it doesn't monitor individual cells.

The latest version has a display of run time remaining, which is a nice touch.

As for PWM/power/efficiency/filament temperature, I have no idea being merely a squished rodent and all. But the NiteRiders do work, the 6W setting appears to be indistingushable from a regular 6W lamp and the batteries last well. It's a tough sell though, at twice the price of Lumis. Not far off twice the weight either...