Page 5 of 5

Re: Bike lighting: how many lumens do you need?

Posted: 6 Sep 2024, 8:30pm
by Brucey
cycle tramp wrote:.......has any one tried using to lower powered leds together from the same dynamo? And if so what we're the results?
I did some experiments along these lines and I am able to share the results with you all. My experiments came in two different formats; on and off the bike.

1) In the first set of trials I used a StVZO-rated generator to power two cheap LED torches which must have had some voltage multiplying trickery inside them, because they used just one 1.5v cell up to ~0.5A current. As the battery weakens both the voltage and the current droop, meaning that the lights were brightest with a new Zn-C cell installed, but had a very long 'tail' of lesser output. I configured each torch with a diode in series as protection, then wired them back to back across the generator, so each lamp would see half-wave rectification. The idea was that i'd get some light at low speeds, but the full current output of the generator (~0.9A ac) wouldn't blow the lights up either. The half-wave rectification was intended to be attention-grabbing, and the two light sources were meant to provide better definition.
I lashed each torch to either side of a set of drops and went for a ride. The results were interesting, and although nothing blew up it was not a complete success. The definition might have been improved but the strobing was terrible. It was certainly attention-grabbing, especially at low speeds, but also completely unbearable. I daresay some smoothing capacitors would have helped, but I didn't have any. Luckily I am not susceptible to epilepsy, but that would almost certainly have set me off if I was.

2) the second set of tests were off the bike and used cheap dynamo LED lights.These came from the netherlands and were sold under the FALKX brand; according to the packaging they have a 1W rated LED inside. They use an exposed emitter and a fairly simple modified parabolic reflector to create a mostly torch-like beam but with a square-ish bright spot in the centre, which I would guess is about 30 lux if pointed about 10-15 m away, Because the emitter is exposed the light is highly visible, even in the daytime/sunshine. It is certainly bright enough to be a real annoyance to others if pointed the wrong way, just like any other bright light. There is a built-in reflector. As a 'be seen' light it is excellent, and as a 'see by' light it is quite passable, very good for the money (I paid less than ten quid each I think for mine which were the Er9a model, I believe). Looks-wise you would be forgiven if you mistook it for one of B&M's offerings, having a very similar stainless steel bracket and housing design. In short, it is a good light by any standards and for a budget light it is remarkably good. It reaches full brightness at about 5v, 0.5A. It is a 3W rated light, so I think it must use a voltage regulator or a Zener diode to turn any surplus generator output into heat. Until it taps out, it seems pretty much as good as anything at turning relatively small amounts of electricity into useful quantities of light.
Simply connecting two these wee beasties in series means that you need to go a little bit faster to make any light at all (if any part of the open circuit voltage waveform exceeds ~7V there will be some light; c.f. ~3.5V for a single lamp) but most generators will manage the ~10V 0.5A required for full brightness. If both lamps are pointed in the same spot, the beam is twice as bright, which sounds a lot better than it looks It is usually much better to make the relatively small centre splodge either twice as wide or twice as long (the latter being a lot more straightforward if the lamps are mounted via a single bracket). If two lamps are wired in parallel, the low speed output is very good, but the lamps never get to full brightness, unless the generator will produce 1.0A, which most won't. Just for fun, I connected a set of four lamps to the generator, in two parallel series pairs. Just like other other parallel setup, full brightness (which in this case requires ~10.0V and ~1.0A) is never likely to be achieved. However, it does make more light.
So the sweet spot is likely to be a pair of lamps in series, if they are like the ones I tried. These can (with the use of tube spacers and studding) be mounted on a single bracket and are easily aimed so as to make a usefully longer centre splodge. In terms of not dazzling oncomers, four lights is about four times worse than a single lamp, but two is probably OK. They are probably best wired up as ' master' and 'slave', with the latter interrupting one of the wires to the former to which the rear light is also connected. If both switches are closed, all the lights are on. Turning either switch off (opening it) always extinguishes both front lights, but if the rear light is capable of running happily alone (not all are), the 'slave' interrupts to ground connection of the 'master', and both the generator and rear light have their own ground connections (which some do) then when the 'slave' switch is open and the 'master switch is closed, the rear light alone should come on.

Re: Bike lighting: how many lumens do you need?

Posted: 7 Sep 2024, 11:17am
by Brucey
plancashire wrote: 5 Sep 2024, 10:38pm.....LED vehicle lights are used for the non-directional day-running lights, decorative do-dahs and the indicators, I think. The focussed beam is usually a discharge lamp but you can buy parts with a few LEDs arranged in the form of a filament. Correct me if I am wrong.....
that sounds about right to me; the brightest and most powerful vehicle lights (and therefore the choice for most headlights) remains discharge lamps. LEDs are more efficient, but so far a shaped beam of suitable brightness has not been possible with LEDs. Last time I
checked, ~30W was the upper limit for a focused beam using a LED light source. The limit is set by the need for cooling of the emitter; although LEDs are vastly more efficient, they still make more heat than light, and excess heat shortens their life considerably. 30W may be deemed 'adequate' if you are not going too fast, and increased efficiency is of interest in some cases. So I believe LED headlights are increasingly being used on EVs etc.

Re: Bike lighting: how many lumens do you need?

Posted: 26 Sep 2024, 4:38am
by DiTBho
PH wrote: 2 Sep 2024, 12:00pm The best lighting system IMO is to have two lights.
I find a decent dynamo light to be more than adequate for the majority of my riding,
I then supplement that with a very bright battery light.
two lights as redundant light
  • dynano powered
  • battery powered
interesting :D

but which dynamo hub? which one do you have?
  • Shimano? Nexus DH-3000-3N-QR?
  • Sturney Archer? XL-FD?
  • ...
3Watt?

Re: Bike lighting: how many lumens do you need?

Posted: 26 Sep 2024, 5:13am
by DiTBho
Brucey wrote: 6 Sep 2024, 8:30pm StVZO-rated generator to power two cheap LED torches
I'm making my own LED flashlights, with aluminum tubes used as a case and a heat sink.

Having no battery inside, it is quite light, not excessively lighter than other solutions because I am using several aluminum fins to dissipate the heat, however it weighs in the order of one hundred grams.

It's shaped to be mounted with a TUSITA clamp, which is a cheap Garmin clone, not bad, half the price.

Inside are a cluster of four high efficiency LED modules, a PWM controllable current generator and glass lenses. No battery, just a wire that connects both the power lines and the digital control line, PWM signal, at 10Khz, duty cycle 1% to 99%.

Ummm, the case is only "shockproof", it should survive a 1g drop without breaking the lens, but it's not splashproof or waterproof yet, and this will be a serious problem in the final stage.

Anyway, I am using the a Nexus DH-3000-3N-QR 3W front dynamo hub, just ... not mounted them on the bicycle yet, I am using them in the lab, connected to an electric motor that simulates pedaling at about 20 km/h, the output of the dynamo does not go directly to the LED modules but passes through a DC/DC switching circuit with a supercapacitor tank.

This helps when you stop the bike at a red light and the light stays on for 10 minutes.

The whole circuit is heavily feedbacked, it keeps the output voltage stable at 5V and 48V: 5V to power the digital system, 48V for the light system.

I chose 48V for the lighting system because in this way I can keep the cable section thin, and 48V helps a lot the module that drives the LEDs in current without adding other power losses to the line.

The efficiency of the module has been sized to 95%, so it heats up a bit and loses energy, but that's part of the game.

I could also insert an ambient light sensor to automatically adjust the light, but there is no linear mapping between "how much light to make" and "what value to give(1) to the light module".

So, I think I'll put a little knob on the handlebars, for manual adjustment

(1) from 1% to 99% of the maximum brightness

Re: Bike lighting: how many lumens do you need?

Posted: 26 Sep 2024, 7:49pm
by DiTBho
Interesting topic here, dynamo led project

Re: Bike lighting: how many lumens do you need?

Posted: 27 Sep 2024, 12:23am
by edocaster
Always interesting to read about dynamo and lighting projects.

Note that if you're aiming for 48V, I assume the DC-DC conversion is step up as well as step down? Because it may be hard to get 48V from the dynamo at normal speeds with any load. You may not need 48V for lighting, of course.

Re: Bike lighting: how many lumens do you need?

Posted: 27 Sep 2024, 2:19am
by DiTBho
edocaster wrote: 27 Sep 2024, 12:23am Note that if you're aiming for 48V, I assume the DC-DC conversion is step up as well as step down?
step up for the 48V line, step down followed by a LDO for the 5V line.

I am using a Shimano Nexus DH-3000-3N-QR as front dynamo hub, designed for rim brake wheels, it produces 3watts output at ~6 volts.

The 5V line does not need to supply more than 200mA, as it powers an MPU used to calculate speed and cadence, and a small display, which is the most power consuming part along the entire digital line.

Re: Bike lighting: how many lumens do you need?

Posted: 27 Sep 2024, 7:32am
by cycle tramp
Those who've be kind enough to read my demented ramblings (thank you, you're both very kind people) may remember my dislike at the intense whiteness that led lamps are apt to produce, and the way they wash the colour out of whatever they are shined upon....
..it was with interest that I read the following

https://theradavist.com/outbound-lighti ... ts-review/

..which suggests that light wave length may be more important that lumens, and hints that the way forward may not be more lumens, but a light which produces a stronger wavelength across the spectrum.

Re: Bike lighting: how many lumens do you need?

Posted: 27 Sep 2024, 10:41am
by Bmblbzzz
I'm not sure what that Radavist article can tell us really, except that colour is fun. It might have told us something if he'd used the same lamp for each colour, but as he's used a far brighter lamp (and presumably with different beam pattern too) for the green than than the amber and red, that rather spoils any conclusions that might have been drawn. Also, the beamshots are of a small patch of random desert soil and shrub. We don't even know what it looks like in daylight, and it's not going to bear much relevance to many other settings (roads, woods, fields...).

And then he says:
...amber light, which is just full-spectrum white light minus the blue spectrum...
If this is correct, then it's not even monochromatic! It's simply not quite full spectrum.

The final consideration, not relevant to offroad fun but relevant to most of us, is that lighting colours are somewhat constrained by considerations of road regulations and other road users' expectations.

Re: Bike lighting: how many lumens do you need?

Posted: 27 Sep 2024, 11:11am
by Cyclothesist
I took the article TL;DR as: A well directed white light is optimal.

All the 'monochrome' lights were worse than white for seeing and navigating the trail necessitating reduced speed. His amber light needs clarification as to whether it is monchrome - is it just white minus the blue end of the spectrum as he speculates, or a narrow wavelength band in the yellow region? If the former then it's polychromatic.
There is genetic diversity in human retinal cones affecting wavelength sensitivity and some of us have 4 cone types with enhanced colour perception.
https://en.m.wikipedia.org/wiki/Color_vision
https://en.m.wikipedia.org/wiki/Tetrachromacy

An article on visual perception and white LED colour temperature would be illuminating 😉

Re: Bike lighting: how many lumens do you need?

Posted: 28 Sep 2024, 5:00pm
by Cyclothesist
An interesting Chinese article here
https://www.researchgate.net/publicatio ... l_lighting
It's about optimum lighting for driving in long tunnels (apparently there are a lot of these in China). It seems a white with less of the blue spectrum is better. Maybe applicable to cycling at night where it does feel like you're in your own light-tunnel.

Re: Bike lighting: how many lumens do you need?

Posted: 28 Sep 2024, 6:06pm
by Nearholmer
Probably wouldn’t help on-road, but I find that one of the biggest issues off-road is the unidirectional nature of the beam, giving too much contrast, so I propose replacing beam lamps with large panels, maybe 1m c 0.75m, secured to the front of the bike, providing a diffuse, but still high intensity, light.

This may or may not catch-on (probably mostly catch on passing shrubbery).

More seriously: I fitted a very bright circular lamp with diffuser in my study, one that I think might be meant for bathrooms, or police interview rooms. That is only c250mm across, but gives a very high, but even, level of illumination of the entire room, and I can imagine might actually work on a bike in the woods.

Re: Bike lighting: how many lumens do you need?

Posted: 29 Sep 2024, 9:32am
by Cyclothesist
Two lights spaced apart on the bars is more practical. I used to have an old lead-acid battery twin halogen set for off road night riding. It was heavy but good. I think 2 beams give better illumination for depth perception.

Re: Bike lighting: how many lumens do you need?

Posted: 29 Sep 2024, 10:24am
by geocycle
DiTBho wrote: 26 Sep 2024, 4:38am
PH wrote: 2 Sep 2024, 12:00pm The best lighting system IMO is to have two lights.
I find a decent dynamo light to be more than adequate for the majority of my riding,
I then supplement that with a very bright battery light.
two lights as redundant light
  • dynano powered
  • battery powered
interesting :D

but which dynamo hub? which one do you have?
  • Shimano? Nexus DH-3000-3N-QR?
  • Sturney Archer? XL-FD?
  • ...
3Watt?
I definitely agree with the two light strategy, especially on the back. All my bikes have dynamos which are great and always there for the unexpected times. For example, middle of the day yesterday it started to rain really heavily and I was pleased to be able to flick the dynamo switch. You can get poor connections and cable issues so I’d not want to be completely reliant , especially at the back. Two battery lights at the back are also fine.

There’s not much to choose between the dynamo hubs and you do get what you pay for. My main commuter has a SON classic which is about as good as they get. The other bikes have SP and Kasai dynamos, both have been fine although are not used as much. Depends what you want to use them for and how much they will be used.

Re: Bike lighting: how many lumens do you need?

Posted: 29 Sep 2024, 12:25pm
by DiTBho
Interesting video on YouTube :D
  • "Are Expensive Bike Lights a Rip Off?!", see here
  • "10 New Bike Lights for 2023, NEWBOLER, XOSS, SoRider, Magicshine, X-TIGER, NATFIRE Front Lights", see here