Advice on a single LED dynamo powered headlight

You might want to look at the newer XML instead of the MCE. It comes in 6V or 12V versions.
For road optics go as tight as you can with these bigger leds.
See pilom.com for circuit ideas and to see the tradeoff between 6 and 12V.

znomit said:

You might want to look at the newer XML instead of the MCE. It comes in 6V or 12V versions.

Click to expand...

Reading through the XM-L datasheet I noticed on page 3 the forward voltage is only around 3 Volts and it's a single die device (not an array). As far as I know a dynohub will supply 500mA at most so I don't see how that can work well without a buck coverter which would decrease voltage and provide higher current from the hub. Am I missing something here? With the MC-E I was planning on using all 4 dies in series with Circuit 7 from pilom.com.

LE: I found the multi-die XM-L devices after a bit of searching on Cree's website but unfortunately my local supplier doesn't carry any of them presoldered to stars.

rac said:

Reading through the XM-L datasheet I noticed on page 3 the forward voltage is only around 3 Volts and it's a single die device (not an array). As far as I know a dynohub will supply 500mA at most so I don't see how that can work well without a buck coverter which would decrease voltage and provide higher current from the hub. Am I missing something here? With the MC-E I was planning on using all 4 dies in series with Circuit 7 from pilom.com.

Click to expand...

Yeah theres a 4 die version of the XML:
http://www.cree.com/led-components-and-modules/products/xlamp/arrays-directional/xlamp-xml-easywhite
With four dies in series you might want to go to circuit 8 for the low speed performance, or just go with the 6V version, should still be plenty of light.

I had considered using a MC-E with an Iris optic myself. My thought was to manually or automatically switch from 2 LED segments in series to 4 LED segments in series as the dynamo speed increased. It might even be possible to take advantage of the different positions of the LED segments to make the low speed LEDs act as a low beam, and the extra two LEDs might act as a high beam.

The only thing that killed my interest in it was that the optics had a much wider beam pattern than I like. With the XR-E's that are in my current bike light, I'm using Ledil Rocket smooth spot optics. These have a FWHM angle of +/- 4 degrees, and I'm quite happy with this beam width for the mostly rural roads on my commute. The best I could find for the Ledil Iris, which is listed as 9 degrees. Actually.... looking at it again, I'm not sure if the Iris would be much different than the Rocket SS. The Iris sheet says "9 deg" while the Rocket sheet says "+/- 4 deg" It's surprising that they don't use a consistent terminology. My earlier thought was that the Iris beam was twice as wide as the Rocket, but now I'm thinking the two beams might be nearly the same.

At the same time... I find it confusing that the Iris datasheet doesn't give a graph for beam intensity for the XM-L. It's probably worth giving the Iris a try if you do use the XM-L. I never did find something that looked like a better option.

My only other caution would be regarding wiring all 4 LED segments in series. This will produce more light at high speeds, but will increase the speed at which the LEDs start to produce light. If you do much riding at slow speeds (i.e. up hills), then you may want to just use two or three LED segments in series (or include a switch to change between 2 and 4 leds).

Steve K.

Adding to what Steve said, the other thing to consider is standlight capability. If you have a stop-and-go pattern on your ride, either dense urban conditions or lots of short hills with no extended slow periods, a large standlight can smooth out those slow spots and get around the problem of running a high voltage for your LEDs. If, for example, your system only operates between 5.5V and 8.5V (such as with 2 XMLs and a red rear light OR a 6V XML and a red rear light), a standlight will never discharge below that 5.5V, and even when stopped will spend a lot of time dropping from 6V to 5.5V.

This makes it easier to run higher voltage systems, the tradeoff is that you may have a harder time getting up to full brightness if you don't tune your system to your ride. That being said, 2xXML at .3A are brighter than 1xXML at .45A, so the upgrade is worth it in most cases.

Also, consider a tuning cap. Putting a 100uF (actual value may vary) bi-polar capacitor before your rectifier (circuit 5) can boost your output at moderate (15-18mph) speeds by 50% or more.

Thanks for all the tips. Regarding the MC-E version of the Iris, Ledil lists the FWHM angle as ±5.5° for the spot version and ±15° for the medium. I suppose the 9 degree XM-L version should be ±4.5° but there's no sure way to tell other than mailing them. I was attracted by the MC-E as a novice builder because it seemed to put out a decent amount of light, four dies mean I can set it up in various configurations (4 series, series - parallel), it is readily available from my local supplier (Farnell) and they stock a good selection of lenses for it too. As for the XM-L, Farnell doesn't stock the 4 die version on a star, which means I can't really use it. I haven't found any four die XM-L stars available from other online suppliers either. XR-Es are readily available, just as the MC-E.

My commute is part urban, where there's plenty of street lighting and rural roads which can be pitch dark at times. No hills around though, so my speed will be in the 20-30 kph range for most of the trip. I'd like to integrate a standlight and I'll also look into tuning capacitors, but I thought I'd start by building the light itself (LED & optic) before further experiments. The single LED requirement is mostly related to the enclosure: it will be easier and cheaper to make a round aluminum housing which only requires a lathe. Being an absolute beginner, a low complexity project like this seems to make sense. The general goal is to end up with something better than an IQ Cyo which I can buy for 60 euros.

rac said:

The general goal is to end up with something better than an IQ Cyo which I can buy for 60 euros.

Click to expand...

For yourself, have you defined "something better" yet?

Marcturus said:

For yourself, have you defined "something better" yet?

Click to expand...

Yes, I'd like a beam with at least as much throw as the Cyo while having a bit of spill too. I realize the comparison is a bit apples & oranges, the Cyo has a custom reflector with beam cutoff while my light will use a symmetrical beam. However, blinding incoming traffic won't be much of an issue on well lit city streets. As for the darker country roads I ride, traffic is pretty light to nonexistent in the evening. Supernova for instance are also using symmetrical beams on their standard E3. I gave the Cyo as an example since it's one of the more affordable off the shelf products which a lot of people seem to be using. I'm not looking to have all the nice features of a commercial light such as a good standlight right away; the main concern right now is the choice of emitter/optics.