|
The Radon headlightNEW: see our Scurion lamp!
As you read in the “no more carbide” article, now seems to be the time we can really go into caves using electric light only. After hesitating for a long time, I now decided to build a “complex” LED lamp. In contrast to the polaris design, which was maximally simple to make it as reliable as possible (sacrificing efficiency), Radon is a high efficiency, high power design. It is based on a Luxeon 3W LED (using the highest binning to get maximum light output) and a high efficiency synchronous step-down converter.
A small PIC microcontroller is used to implement different power levels and battery supervision. One might think you could simply build the electronics and put it, together with the LED, in an old headlight case (like an old Duo or Zoom lamp). Unfortunately, this does not work. First the LED has to be cooled adequately, which requires a heatsink (note the cooling fins). Second, the electronic circuit is very sensitive - if the electronics get wet, it will most probably stop working (if your Duo died as mine did, you quickly rule out the possibility of using the Duo case). This is the main reason why I built the Polaris lamp. It works even when totally flooded (although it would corrode quickly. But a “main” lamp needs more power and effeciency, which requires more sophisticated electronics. For the Radon lamp to be reasonably reliable, it needs an absolutely waterproof case. HousingUsing the Radon concept requires a watertight enclosure for the electronics. These are the main features:
These two images show the current construction of the Radon case. The front part is changed, I added cooling fins to decrease thermal resistance. Electronics |
|
Setting/ Input Voltage |
800 mA |
260 mA |
81 mA |
|
9 |
92% |
89% |
83% |
|
8 |
92% |
90% |
83% |
|
7 |
92% |
91% |
85% |
|
6 |
92% |
92% |
89% |
|
5 |
92% |
92% |
90% |

In case you enjoy your lamp so much you don't want to stick it in the mud, here's another idea...
As mentioned above, you can use any battery with the correct voltages. The cell voltages for common batteries are:
|
You could use four or five NiMH cells. Caution! Freshly charged Nixx batteries can have rather high voltages. The other alternative are two Li-Ion cell packs. I prefer this possibility because the energy density is higher and the handling easier (only one instead of four or five). You must use Li-Ion packs with supervisor electronics (like the BP-511) to avoid deep-discharging (and destroying) the cells. Li-Ion cells have another huge advantage: You can easily tell the remaining capacity from the cell voltage. This makes it possible to automatically reduce the light output when the battery gets near empty. This avoids you find yourself in the dark abruptly when the cells are empty. For Li-Ion cells, I will implement a power scheme that massively prolongs lamp runtime by reducing the LED current (even 10 mA give usable light output). This gives you time to realize that the battery is empty and find a good spot to change it (not somewhere in a narrow crawl or in a 100m shaft).
For backup purposes, you could carry a 9V alkaline or better a 2CR5 lithium battery with you, should you unexpectedly run out of batteries. If the electronics doesn't work anymore, you can even connect it to the emergency cable.
NEW: Just found a BP-353 (3.75 Ah, 2 Li-Ion, 25 Wh) at ebay for 10 EUR! This would mean 8 hours of light on maximum setting!
I implemented an automatic power reduction for Li-Ion cells. If the battery voltage reaches 7.1 - 6.8 - 6.5 - 6.3 V, the power setting is reduced to the next lower level. This ensures you're not standing in the dark abruptly. This graph shows how this is working:
I connected a freshly charged BP-511 to the lamp. The upper graph shows the battery voltage, while the lower graph diesplays the LED voltage. This is not so interesting, more relevant would be the current through the LED. But as the LED is current regulated and the current would have been more difficult to measure anyway, I just show the voltage across the LED. You see that after roughly 2 1/2 h the battery is nearly empty and the current is reduced from 800 to 260 mA (final values will be different). After another 30 minutes, the current is reduced again (81 mA) and after that, the last two PWM settings are activated (27/8 mA), giving another two hours of light (dim, but better than nothing) from an almost empty battery. After six hours, the BP-511 battery protection circuit switches off and you're in the dark.
Obviously, this scheme does not make sense for NiMH batteries. For these, I'd just disable this power management feature and the LED current would fall continuously when the battery voltage drops below 4V. The last visible glow is emitted from the LED at about 2.5V.