LED Headlamp

Ian

Esteemed Pedelecer
Apr 1, 2007
1,333
0
Leicester LE4, UK.
I've been working on this seasonal project for a few weeks but have been awaiting a shipment of LEDs from a Chinese supplier which finally arrived yesterday enabling me to complete the project.

It is a pair of LED headlamps built into dynamo lamp housings which come as standard on Ezee bikes. The reason for the pair is to give both my bikes convenient and bright lights using a large number of LEDs and running from the main battery. The lamp for the Sprint contains 55 LEDs and the faster Torq has 63 LEDs, arranged in 6 & 7 series chains of 8 and 1 chain of 7, the chain of 7 serving a secondary purpose as a voltage regulator for the 555 timer to provide the flashing function. Constant brightness is ensured by an LM317 regulator dropping the battery voltage to 28V, the current in each LED chain being then limited to 25mA by a series resistor.

The original reflector was removed from the dynamo lamps and replaced with an aluminium plate with holes drilled for the LEDs, the plate also serving as a heat sink for the voltage regulator seen fitted below. The metalwork was marked out by a method I often use when high accuracy is not required, that is to simply stick a computer generated template to the metal and use that as a cutting and punching guide.



The flasher module seen below is a simple 555 circuit built on stripboard and designed to give a flash rate of 2.5Hz and a duty cycle of 20% which ensures high visibility and very low power consumption. After testing the modules were coated in epoxy resin and encased in shrink tube to give mechanical strength for the high vibration environment at the front of a bike.



The lamp is controlled by a miniature toggle switch with weatherproof cover mounted in a very convenient bulge on the back of the lamp housing and arranged to give Flashing - Off - Steady. The red "blob" is the flasher module and the black "blobs" are the series resistors, all encased in shrink tube.



The Sprint lamp was finished yesterday and tested last night around a local unlit country park where it proved very satisfactory, the Torq lamp was finished today but heavy rain is preventing a proper test.

 

Ian

Esteemed Pedelecer
Apr 1, 2007
1,333
0
Leicester LE4, UK.
Cheers Frank, just stuck my head out to check the weather with a quick road test in mind, it has stopped raining... but it is snowing:(
 

flecc

Member
Oct 25, 2006
53,260
30,648
Nice job Ian, very professional, and no-one can say they didn't see you coming now.

It's been a grotty wet day in the south too, and yesterday was damp as well, but we've been spared snow.
.
 

Ian

Esteemed Pedelecer
Apr 1, 2007
1,333
0
Leicester LE4, UK.
Thanks Flecc. We had a dry day yesterday, even a bit of sunshine, but very wet today. The snow is only very light, melting on ground contact, but not what I expected when I looked out.
 

Tiberius

Esteemed Pedelecer
Nov 9, 2007
919
1
Somerset
Looks very good. How are they mounted onto the bikes? I'm starting with a Torq with no lights at all, just a reflector on the back.

Nick
 

giguana

Pedelecer
Nov 8, 2007
216
0
wow, I have some fairly efficient and bright red piranha led lights I got from America, they are machined onto to metal strips of 32 originally for the back of Cadillacs and they are very bright.I have lots to spare. LED is brilliant, I would like some 1/5w bright ones for the front... is it possible to simultaneously wired the batteries in series and in parallel to make 36V and 12V for the motor and lights?
 

flecc

Member
Oct 25, 2006
53,260
30,648
... is it possible to simultaneously wired the batteries in series and in parallel to make 36V and 12V for the motor and lights?
Not really, and if just some of the cells were used to do it on NiMh, the unbalanced cell charges that could result would shorten battery life and possibly cause a cell failure with a following burnout.

On Li-ion it's an even bigger no no.

With three SLA batteries it can be done, the one next to electrical ground feeding the lights.
.
 
Last edited:

Ian

Esteemed Pedelecer
Apr 1, 2007
1,333
0
Leicester LE4, UK.
Nick, Both lamps are secured with a short bracket to the mudguard fixing screw on the fork crown, you can see it clearly on this picture of the Forte. In both cases the lamps were original equipment on the bikes, the Sprint having a bottle dynamo on the front wheel and the Torq having a small LED bulb powered by the main battery. Dynamo lights and brackets are available fron decent shops.

Earlier Torqs had a bottle dynamo, later ones having main battery lighting but retaining the dynamo bracket on the left hand fork leg.
 

Tiberius

Esteemed Pedelecer
Nov 9, 2007
919
1
Somerset
Thanks Ian, that will help me to work something out. I have the bracket on the fork leg, but its not a good place for a light, and now I know what it was for.

Nick
 

Ian

Esteemed Pedelecer
Apr 1, 2007
1,333
0
Leicester LE4, UK.
wow, I have some fairly efficient and bright red piranha led lights I got from America, they are machined onto to metal strips of 32 originally for the back of Cadillacs and they are very bright.I have lots to spare. LED is brilliant, I would like some 1/5w bright ones for the front... is it possible to simultaneously wired the batteries in series and in parallel to make 36V and 12V for the motor and lights?
It should be possible to use 3 identical complete assembles in series on a 36V supply if you have room on the bike. If made for automotive use they should have the necessary voltage and current limiting components built in.
Using individual LEDs is more tricky as they require a series resistor to limit the current and care must be taken to ensure that the current remains within limits throughout the full battery voltage range. A single LED can be connected to a 36V supply through a suitable resistor (eg 1.5Kohms 1W) but it's rather wasteful, far more power being lost in the resistor than the LED.
 

Ian

Esteemed Pedelecer
Apr 1, 2007
1,333
0
Leicester LE4, UK.
Nick, The bracket on the leg was for the dynamo itself, not the light. The light fixes to the fork crown, in the centre above the mudguard (just in the right place for the control cables to rattle against).
 

Jeremy

Esteemed Pedelecer
Oct 25, 2007
1,010
3
Salisbury
I posted a constant current supply circuit on the other LED thread that neatly gets around the variation of supply voltage problem a lot better than a series resistor. It's nice and easy to build and seems to work well. The only things to watch are the maximum voltage drop across the 317 and the maximum dissipation. Generally this isn't a problem with the right number of series LEDs connected, unless using a very high battery voltage.

Jeremy
 

Ian

Esteemed Pedelecer
Apr 1, 2007
1,333
0
Leicester LE4, UK.
Yes, but for multiple series chains you need multiple current sources. I used a single constant voltage source (28V) feeding several chains via resistors (100R), one of those chains being tapped to supply 10V for the 555, it's simple, it works and it was cheap as I had a bag full of LM317T's, but no room in the lamp housing for 8 of them. It could be argued that a single current source could be divided with resistors, the problem there is that if an LED in one or more chains fails, then current in the others will increase.

When advising people who's electronics abilities are unknown it's best to keep it simple. in the case of my last post a singe led + resistor is as simple as it gets and a single LED fed from from a 36V is going to experience less than a 15% current variation anyway over the expected voltage range. Not that I would advocate a single LED for any serious lighting.
 

giguana

Pedelecer
Nov 8, 2007
216
0
I would be tempted to use this style of LED
Find 5 X LUMILED (1W) WHITE - SUPER BRIGHT LED - UK SELLER on eBay within LEDs, Components Supplies, Electrical Test Equipment, Business, Office Industrial (end time 28-Nov-07 19:58:39 GMT)

-the data sheet says about 3.5V each, so 4 in a series going to one battery should be really efficient I guess... apparently it's best to have as many LEDs in series as possible with a minimum resistor on the end, although if they are sufficiently undervolted no resistor is needed...best thing is to use them a bit undervolted so that they are very bright but not too hot and last a long time. I have read that always using one resistor for every series is the norm but I can't remember the reason, something to do with stability, the resistance of the lights changes over time.
 
Last edited:

giguana

Pedelecer
Nov 8, 2007
216
0
Having a strong light doesn't make for good visibility though, at night in the rush hour on a narrow road with lots of cars in a queue, a cyclist overtaking them is completely invisible, the single light of a cyclist is the same as the row of single lights of headlamps from the queue to oncoming traffic, I really recommend getting some "reflective visibility tape" and wrapping it all around the bike, especially around the pedal bars.
 

Ian

Esteemed Pedelecer
Apr 1, 2007
1,333
0
Leicester LE4, UK.
LEDs such as those Lumiled 1W types, also called luxeons (which is a trade name) tend to have a wide beam angle, typically 120 degrees so need additional optics to make them suitable for headlamp use, their relatively high current requirement also complicates the provision of an efficient power supply when used with a high voltage battery. These are not impossible obstacles but I feel that it's simpler to use a large number of small LEDs that have a native beam angle close to the ideal. Buying LEDs can be a bit of a lottery though as some suppliers exaggerate the brightness of their products more than others.

The voltage quoted for LEDs is not a voltage rating, it's the forward voltage (Vf) drop at the quoted current and will vary between LED types, individual LEDs of the same type and at different temperatures. If a series resistor is used as a current limit then it must be high enough to prevent variations in Vf from causing excessive current to flow through the LED which in turn will heat up the LED, the increased temperature causing Vf to become lower meaning more current and more heat and so on leading to a destructive condition known as thermal runaway.

I agree with your comments on visibility, although a really bright flashing light helps a lot, and an LED lamp can be operated at a much higher current in flashing mode, making it flash very brightly indeed.
 

giguana

Pedelecer
Nov 8, 2007
216
0
that's a good point I forgot that high-power LEDs can be diffuse, the beam is really soft and gradual although it is fairly bright in the middle...I want something like this-

http://dmcleish.com/CPF/McR-8/Arc-Pelican-Beam.jpg
with focused light in the middle and diffuse light all around so it is possible to see grates puddles potholes just close the bike, and all the branches and things around the bike and with a focused strong beam going forwards to see far ahead...
must be really difficult trying to get it just right.
 

Ian

Esteemed Pedelecer
Apr 1, 2007
1,333
0
Leicester LE4, UK.
must be really difficult trying to get it just right.
I don't think any one light will be ideal for all conditions, on dark paths I'm constantly adjusting mine depending on how far ahead I need to see, I guess the ideal would be an easily adjustable spot light for the distance plus a flood light for the area immediately ahead.
 

Jeremy

Esteemed Pedelecer
Oct 25, 2007
1,010
3
Salisbury
Yes, but for multiple series chains you need multiple current sources.
Which is *exactly* why I previously took the time and trouble to post the link (in the LED thread) to the application note with the simple circuit, using a 317, that does drive multiple LED strings safely....................

Jeremy