A discussion and study of orienteering headlamps. The article includes: a review of what typical orienteers currently use; the authors personal requirements of a suitable orienteering headlamps; investigations into some of the technologies used in lamp design; and details of the a project to construct a home-made lamp to the authors requirements. This article was originally posted as a forum item and was re-written to suit the style of an article. This article is likely to be added to as the project develops. Last Update 15/2/08
Night Orienteering Headlamps
Technology has moved on since the lightweight, high performance halogen bulb orienteering headlamps that many keen night orienteers currently us were designed. This presented an opportunity to investigate the possibilities of developing a new headlamp for orienteering (by way of a potentially useful (and straight-forward) electronics project). This project discusses the requirements (from the personal perspective of the author – a very experienced night Orienteer and (in the early days) headlamp dabbler!). A review of the designs and technology used in current orienteering headlamps gives a performance datum and offers some pointers towards area for possible improvement.
Review of Authors current Orienteering Headlamp:
The current headlamp set has been in use for approximately five years, although parts of it are adapted legacy items from earlier versions of home made lamps. The headlamp set consists of two main parts: the headlamp; and the battery pack. There are also some auxiliary items that are essential to the efficient use and function of the lamp set: battery harness; battery charger.
Existing Headlamp: 10/20W halogen Silva SL.
This is still very serviceable and functionally provides good illumination. The level of wear and tear on the headset is relatively minimal for most of the unit and should give many more years of use before it fails irrevocably. The front cover for the reflector, however, was broken and lost (along with the rubber securing ring) during an event a few years ago and is currently replaced by a very robust and satisfactorily effective (but crude!) homemade version. The homemade Perspex reflector cover is heavier than the original and very marginally effects the comfort. NB I’ve recently discovered (Nov 2007) that the reflector cover and ring are available as spared items from Compass Point for approx £6. This is a fairly modest cost (compared to the price of the lamp itself – approx £80) and the fact that they are listed as spare items at all may imply that they need replacing from time to time.
The two levels of bulb power are achieved by having two separate bulbs, both close to each other at the focus of the multi-faceted parabolic reflector. The bulbs a fitted with the filaments horizontal, and the 10W bulb above the 20W bulb. Only one bulb is lit at a time, and the positioning relative to the focal point means that the 20W “Main Beam” shines out higher than the 10W “Dipped Beam”. The head fitting gives adjustment to tilt the reflector. I position this such that the 20W beam shines straight ahead and the 10W beam (used most often) tilts down to land on the ground approx 8m forward. The design of the reflector and position of the bulbs gives a beam width of approximately 11 degrees, with enough light spill into the peripheral zone to illuminate the map and immediately surrounding terrain. The multi-facet nature of the reflector provides the very usable diffuse beam.
This headlamp replaced a previous homemade type based on the lens unit of an “Ever Ready” hand torch, fitted with a 4W halogen bulb. The beam pattern of this previous home made headlamp was quite narrow and had a greater range than the Silva head lamp, but peripheral illumination was much reduced. The Silva lamp was a functional improvement in many respects.
A home constructed assembly of five “D” size Nickel Cadmium (NiCd) cells. The cells were electrically connected in series (using short pieces of thick wire soldered to the cells). The cells themselves were industrial capacity (4.5 Ahr) items with tags on each end to assist soldering. The nominal voltage of each NiCd cell is 1.2V, giving an overall nominal battery voltage of 6.0V.
The discharge characteristic of the NiCd chemistry is almost that of an ideal battery: the voltage stays at a constant level for almost all of its charge, before falling rapidly at the end. The freshly charged cell voltages actually start slightly higher than 1.2V, at approx 1.4 (giving a starting voltage of the battery of 7.0V), but fall over the first 10 minutes of use to the 1.2V plateau. The cells are effectively empty of charge when they drop to 0.9 – 1.0V (4.5 – 5.0V for the battery) and should not be allowed to drop further than this or they may effectively be destroyed (and couldn’t be subsequently charged). One of the cells “died” and went short circuit (a symptom of age and over-discharge) approximately a year ago and was replaced. NB. This is not normally recommended, since cells usually age together as a batch – it is only considered a temporary fix (because I needed the battery, had a spare cell, and didn’t want to pay out for a complete new one at the time). This battery gave (when new) a theoretical lamp burn time of 2hrs 25mins on 10W and 1hr 10min on 20W. The design working life of this battery was approximately 7 years, by which time it’s capacity may drop by 25-40%. This was considered from the outset in order to have a full life usable battery for all normal night orienteering. Cell technology advances all the time and more modern cells are probably better.
Another home-made item, essentially based on a factory throw-out! I converted the redundant lab constructed special to type power supply to a constant current supply (adjusted to 350mA) by reconfiguring some of the internal components. It’s a linear power supply design, which means that electrical efficiency isn’t particularly good (it gets quite hot during the initial part of the charging cycle). In practical terms the maximum power consumption from the mains is probably no more than 12W for the 16hrs or so of charging – approximately 0.5KWhr – costing 5p(?) per charge! The inefficiency of the supply is not that significant in the overall scheme. The charger was built so that it should be impossible to damage the cells if I forget to disconnect them (charging current is below the maximum specified for continuous (or “ripple”) charging. One of the benefits of using old tech NiCd cells was that a simple charger (such as this design) could be used without risking the health of the cells.
Silva foam design, such that the battery is carried on you back between the shoulders. A 760g battery does not carry well in a bum-bag or clothing pocket! This harness does what it’s meant to and holds the battery firmly with an acceptable level of discomfort (not really much at all).
Authors Requirements for a Good Headlamp
The headlamp set described above certainly does ok for normal night orienteering, but it’s time to review the essential requirements of a night orienteering lamp. We can then go on to investigate and discuss what could now be produced that would fulfill the design remit of:
a) lots of light for a long period of time
b) good useable beam pattern
c) comfortable to carry
d) light weight
f) uncomplicated design
g) not too expensive!
Note: What got me thinking about all of this is the observation that even something as apparently simple as the design of a battery powered head torch has benefited from the evolution of technology. My observation came from a re-examination of my (now aging) battery pack, which is based round five Nickel Cadmium (NiCd) cells. I have found that NiCd technology is significantly heavier and more bulky than the more modern equivalent capacity Nickel Metal Hydride (NiMh) batteries that seem to be generally available now. I am interested to see how the advance of technology may’ve influenced other parts of the headlamp design.
Modern Trends in Personal Lighting
A bit of digging around on the web (primarily in the areas of caving, mountain biking and adventure racing) has shown that there is quite a lot of activity in the area of lightweight personal lighting at the moment.
Technological trends appear to be taking the leading edge designs well away from halogen bulb based designs, through the relatively complicated and expensive HID (High Intensity Discharge) bulb models, and now into the realms of LED (Light Emitting Diode) based lights. The light producing efficiency of high powered LEDs is now starting to surpass that of fluorescent tubes (and are easily much more efficient than filament bulbs).
LEDs (essentially solid lumps of plastic with electrodes embedded) also have an inherent resilience compared to any bulb made out of glass. On the negative side, there usually has to be some control electronics for the LEDs, and the higher powered versions have to be built in such a way that the LED devices themselves can be cooled (even with LEDs, most of the electrical energy from the battery is converted to heat, though nothing like as much as with light bulbs).
Basis of a project
There’s little argument that a 10/20W halogen bulb based lamp is easily adequate for almost all normal night orienteering, but it’s interesting to explore other possibilities.
It may even be possible to come up with a design for a high performance, cheap-n-cheerful design, made out of easy to obtain items that could make night orienteering with a decent lamp a more accessible for more people. The possible basis of such a design could be one of the cheap sealed unit low voltage halogen lamps that are currently popular as down-lighters in kitchens etc - just work out a good way of mounting it on your head and an affordable 12V battery, and you could be there! There again, things are usually not this simple!
The challenge is therefore set for a project to use modern technology (where appropriate, i.e. safe, functionally effective and cost effective) to satisfy the requirements above.
It would be useful to have feedback and thoughts about the different experiences people have with the various types of lamp.
A Forum item has been initiated to air these views.
I await your responses!
The intention is to add to and update this article as the project progresses (might be over several months!).
Headlamp Project Update 16 January 2008
I checked the performance figures of the currently available LEDs and decided to try components from two different manufacturers: Seoul Semiconductor Company (SSC) “P4”, brightness bin U; and Cree XR-E brightness bin Q5. The market changes all the time with these types of devices, but currently Cree offer the highest performing (indeed they have recently brought out an even brighter bin – R2). The SSC P4 and Cree XR-E are slightly different shapes and require different optics to optimise beam performance. I chose optics manufactured by the Finnish company Ledil.
I chose Lithium Ion (Li-Ion) 18650 cells (2400mAh) to make up the battery. A battery of 4 of these will supply 14.8V (nominal).
I chose a selection of driver circuits (they were cheap enough to try a few!) to output 800mA from battery voltages ranging from 4-18V.
All components have now been ordered:
Ledil Optics – web purchase from Holland
LEDs; Driver circuits; Li-Ion batteries and charger; switches; heatsink compound. These were web purchase from two different suppliers in Hong Kong.
The headlamp is becoming quite international! Reasons for sourcing from abroad are that the choice of components is better and prices are cheaper. Note that orders worth over £18 from Hong Kong could be liable to import duty, VAT and a post office admin fee (£8!)- this has to be factored into the overall cost.
So far all the main items have arrived apart from the driver circuit boards.
Practical construction has not been possible until very recently because of other commitments (Jan Southdowner!), but I’ve now started to try some of these items out.
I mounted a LED on a flat aluminium sheet (as a heat sink) and tried powering it off my existing constant current NiCd battery charger. The results were blindingly encouraging! Speaking numbers: the charger passed 330mA through the LED, which had a forward voltage drop of 3.28V. This gives an input power of almost exactly 1W. I fitted one of the Ledil optics over the LED and achieved a nice bright spot on the ceiling! It is difficult to absolutely compare lights, but comparing this to the existing 10/20W halogen lamp, the single LED seemed to have about a third to half as much light output.
This is a very good result, since the initial design plan is to run 3 LEDs at 3W each (at 800mA Vf can be expected to be approx 3.7V). Thus the single LED should be 2-2.5 times brighter, and there will be 3 of them – potentially nearly twice as bright as the halogen lamp for half the power drain – good to see that there seems to be some truth in the originally estimated figures!
The next stage (whilst waiting for the driver circuits to arrive) will be to refine the mechanical design and layout of wires a little so that 3 LEDs can be mounted together and I can experiment with the different optics (with various beam widths). Following on from this, I will start to look at the construction of the battery.
Headlamp Project Update 15 February 2008
Most components now available. First Build: “TechDem 1”.
The real fun has now begun!
Enthused by the results of the early LED circuit experiments in January, I’ve now started to look at the broader requirements of the project: I have tried a few ideas for the battery pack; tried out two of the three different types of driver boards (unfortunately still awaiting delivery of the final (and most favoured) one); started to address some of the practicalities of construction in a build designed to demonstrate the technology in a usable headlamp – “TechDem 1”.
Like most engineering projects, this one has to progress within various limitations:
Money (I’m having fun with the project, but want to keep costs to a minimum, especially whilst still determining the eventual product – spent £60-£70 on components so far and don’t want to spend much more at the moment);
Time (Other commitments mean that I have limited time to spend on this at present – perhaps 10 hours/week. My time doesn’t come into the cost calculation, however, as it’s a hobby project. In spite of limited time available, I’m conscious that the end of the current night-O season is rapidly approaching – I’d like to have something to try out this season!);
Materials (Other than the components that I’ve bought specifically for the project, I’m using materials that I have to hand with no extra cost implications – much of this is scrap/now redundant items that are being given a second use with some simple adaptation);
Facilities (I don’t have a dedicated workshop and am using standard domestic hand tools, electric drill, plus general purpose electronic soldering equipment and a standard digital multi-meter.
Customer Requirements and Expectations (I’m the end customer, and close contact with the progress of the project means that I should have a realistic idea of what to expect and accept!).
The Lithium Ion cells that I’ve bought to use on this project are a new technology to me and I’m being cautious about how the battery pack is assembled, charged, and discharged. The energy density of the cells is high and they don’t (as I understand) tolerate much abuse – an inadvertent short circuit, over charge, or over discharge could destroy the cell and potentially release a lot of the stored energy in one go – I don’t want that!. I’ve settled on a two cell design (for use with the driver circuits that I currently have, that cannot be powered by more than two cells in series). My preference is to charge each cell individually in a commercial charger specifically designed for the purpose – it’s therefore not practical to have the two cells permanently electrically connected. My current solution to this is to use a connector arrangement between the lamp and battery, where the cells are only connected in series when the lamp is plugged in – I’m therefore using four pole connectors, with a series connecting loop on the lamp side of the connector pair – a simple, cheap, and elegant solution (should mean that it’s also reliable). Selection of the connectors was a bit problematic, but I eventually settled on adapting a lead used for internal power connections in desktop computers – this gave me connectors with four poles, that could carry the current safely, and had a latching mechanism to prevent unintentional disconnection.
The types of driver boards that I have selected were originally designed to fit into mid-price range LED hand torches. I obtained them from a supplier offering them as “LED flashlamp upgrade components”. I settled on drivers that take the supply voltage (from the battery) to provide a current drive to the LEDs at a lower voltage – these are called “Buck converters”. The low cost buck converters that I’ve chosen do not contain the additional circuitry to regulate the output current very closely, and I was expecting some variation of output current with input voltage (e.g. as the battery runs down), in reality, the slight decrease in battery voltage (and corresponding current) over time made little difference to the perceived brightness. The two different drivers currently available are only able to drive two LEDs at their maximum input voltage – this determined a two LED design for the technology demonstration model “TechDem 1”.
First attempt at a usable headlamp design using a selection of the new components. A two cell battery and two LED lamp design was made due to the limitations of the available driver boards. The two LED experimental lamp unit was built into a commonly available aluminium diecast box (approx 50mm square by 30mm deep). Some thinking (and filing!) was required in order to fit two LEDs,, Two LED optics (Medium: +/- 15 degree, and Diffused spot: +/- 9 degree), the driver board, switch and wiring into the relatively confined space of this box. Results of Initial tests using the higher output driver board have been mixed. With the battery fully charged to 8.4V, the driver was supplying in excess of 1.1A to the LEDs – this is overdriving the LEDs and would result in shortening their working life. After a while, the current stabilised to approximately 1.0A (the maximum specified limit) and the run time was in excess of 2 hours at this rate – pretty good! The downside of all this power (just over 7W) was that 5W had to be dissipated as heat, and the lamp box got very hot (peaking at over 70 degrees centrigrade before I introduced additional external heat sinking to protect the electronic components!) – you would not want that mounted on your head! NB. I’ve since learnt that a rule of thumb for heat dissipation from these sort of lamps is that you need approx 3sq inches of surface area (in ambient still air) per watt of LED power. The external surface area of TechDem1 is only about 12sq inches. In terms of illumination, an unscientific measure is that it floodlit (to a level useful for running) the whole of our back garden and seemed to give out an amount of light somewhere between the 10W and 20W settings of the halogen headlamp (albeit with a very different beam pattern and colour – the halogen beam was much more orange and less wide). Next step will be to try the other driver board, and replace the wider angle optic (30 degrees) with a second 15 degree one). The peripheral “spill” slight from even the single narrow optic was sufficient as a fill-in without using a specific wide angle beam. It is great to see that a LED lamp with only two LEDs (my original plan was to use 3, or even 4) and non-optomised beam pattern is already equivalent to at least 15W of halogen power (and using a battery that is much lighter and smaller than the NiCd one used with the halogen lamp – a harness mounted battery may be un-necessary in the final lamp.
Update on the next step:
Substitued the medium angle optic with a narrow one (+/-5 degree) and changed the driver circuit to a five mode, lower output one (measured as pushing around 0.5A into the LEDs on the high setting). Initial indications are very good (still heated up, but nothing like as much) – not quite as bright (due to less current), but should still be bright enough for night-O. Run-time measurements need to be made, but the short (1hr) test made so far shows that the lamp should easily shine for 4 hours on its high setting with the two cell Li-ion battery. A conservative calculated estimate of the light power output (assuming 15% loss through the optics) is 200 lumen (theoretically, the same output from a good 10W halogen lamp). A quick “garden test” comparison was made between the TechDem1 and the 10W/20W halogen lamp (albeit with a duff battery!) – A very smooth and even beam, with a less concentrated spot. Very useful side spill. Very white light (no funny colours when reading the O map), Should be an almost ideal beam pattern for night orienteering – only real deficiency is the lack of a distance penetrating spot, but for a 4W (or so) headlamp very promising.
Even better, I managed to design and make a suitably shaped piece of metal to attach the lamp unit on to an old headlamp band, and sorted out a method of keeping the optics in place – I now have a usable lamp. First real trial will be the Wimbledon Common South East Night League event on Saturday 16th February.
Mica SL 10/20W Halogen + 4Ah 6V NiCd battery
Run time 1.25hr on 20W, 2.5hr on 10W
Weight Overall 990g
Lamp only 226g
Battery Only 764g
TechDem1 + 2.5Ah 7.4V Li-Ion Battery
Est. 3hrs + on High; 6hr + on Low
Weight Overall 272g
Lamp Only 168g
Battery Only 104g
Output of TechDem1 on high approximately equivalent to 15W halogen (i.e. midway between 10 and 20W settings on old lamp). Power consumption of TechDem1 est 4-5W. Maximum lamp body temperature (still air in ambient room conditions) 55 deg C.
Revision 4 - 15/2/08 Added “Headlamp Project Update 15 February 2008”
Revision 3 - 16/1/08 Added “Headlamp Project Update 16 January 2008”
Revision 2 - 30/11/07 (formalised version to replace original article with extra details added).
Revision 1 - 16/11/07 Initial Version of article (NB the original article forms the start of the discussion thread on the forum).
Posted by Peter Chapman on 16th Nov 07