Most commercial bike lights usually run on NiMH (Nickel metal hydride) and Li-ion (Lithium Ion) batteries. These batteries have higher energy densities compared to SLA (sealed lead acid), as well as have the advantage of being smaller and lighter. The initial cost to purchase these types of batteries (and the appropriate charger) are also usually much higher.
My choice for a light source is a quartz halogen bulb. Most high end commercial models utilize one or more LED (light emitting diodes). These new LEDs are very efficient but still cannot compare to the output of a single halogen bulb. The light produced by an LED is much whiter. Some people like this better than others. A yellower or warm light looks more light natural light and highlights terrain better at night. Some LEDs on commercially available bike lights are not user-replaceable so even though they could last 10 times longer than a comparable halogen, once the LEDs burn out, it’s usually hard to get spare parts. An MR16 bulb can be purchased from $1 to $5 for the most common wattage and has an average lifetime of 2000-3000 hours. Some long life models are even rated up to 10000 (although these are mostly available in 35watt versions and higher). One bike light made by Busch and Muller Germany has a lighting element rated for 1000 hours and costs €150 to replace. (Yes, that's 150 Euros for the light).
LED light output is harder to focus or collimate. Even with specialized optics, not all of the light of an LED based system is usable. An MR16 bulb already has an optimized reflector to concentrate most of the light whether it’s a flood or a spotlight. A 20-watt MR16 is normally rated at about 700-850 lumens and a 10 watt MR16, about 400 lumens. Very good LED lighting systems that achieve 400 lumens cost more than CAN$200. A 900 lumen model by Light & Motion costs almost CAN$700!
Some people would expect an LED to run much cooler than a halogen. While this is true to some extent, high powered models with multiple emitters produce considerable amounts of heat as well. The best high end LED lights come with some form of massive heat sink to dissipate heat. In most cases, the external housing is also made from machined aluminum to help in the cooling process.
Perhaps the biggest advantage of these LED-based lights would be the total system weight. Since all of them use either NiMH or Li-ion batteries, a complete unit will weigh 250-650 grams (battery and all). My system, which uses an SLA, weighs in around 1800 grams (almost 4 pounds).
I do not ride a commuter bike and will probably do very occasional street riding at night with the mtb. Therefore, as it stands, it would be quite impractical for me (cost-wise) to purchase a CAN$300 light and use only for a couple of times in a year. The chemistry or battery type also plays a factor in this practicality. Nickel Cadmium (NiCd), NiMH and Lithium batteries (including lithium polymer) lose a portion of their charge each day. In most cases, if these batteries are allowed to self drain, they would be unusable after about a month unless recharged. An SLA would retain its charge much longer and may need occasional charging every 3 months or so. So while NiMH, NiCd and Lithium ion pack more power relative to size and weight, the humble SLA still retains its viability for my applications.
Below is a table summarizing the pros and cons of some commercially available solutions versus the "spice jar" bike light that I made.
I do not claim that my system is better (both in performance and looks). I have nothing against LED lighting and as a matter of fact, I would probably purchase on or more of these LED lights if I had both the resources and actual need for them.
Note that on this comparison, the run time at maximum brightness (yellow highlight)is about par with the other models. The 2.5 hour runtime for the "spice jar light" or SJL is already derated /adjusted from the theoretical runtime achievable using an SLA type battery.
Run Time = Wh / W (Watt hour / Watts)
where Wh = V x Ah (Voltage x Ampere Hour)
V = 12V
Ah = 5
W = 20
Wh = V x Ah
= 12 x 5
= 60
60 / 20 = 3 hours
However, an SLA battery should never be discharged beyond 80%, so a 5Ah technically has 4Ah of usable power. Replacing values above for Ah gives a final run time of about 2.4 hours.
RT = [(12 x 4)/20] = [(48/20)] = 2.4 hours
Further adding to the issue with SLA is that the rating of 60Wh is with a drain of only 5% per hour or C=0.05. At a higher discharge rate, the capacity is further decreased. The 20 watt halogen bulb produces a load of 1.67 amps < C/40. This depletes the SLA much faster. Based on RT above of 2.4 hours and a derating factor of another 0.85, it would be safe to estimate that the actual runtime would be closer to 2.04 hours.
2.4 x 0.85 = 2.04 hours
All these derating factors are only estimates and may be even unnecessary since my required runtime is only about 1.0-1.5 hours. The 5Ah SLA should therefore have enough buffer capacity in terms of power.
Going back to the table, the maximum brightness of the single MR16 bulb at 12 volts is around 700-850 lumens. This rating varies with bulb manufacturers. However, even at the lowest 700 lumen output, the SJL is still almost double the output of the dual-LED cygolite mitycross 400. Another advantage of the MR16 is that it is quite tolerant to over-volting. By increasing the input voltage for the lamp, the output is also proportionately increased. Since the SLA is not capable of being used for over-volting, this method of increasing light capacity is achievable only with a different power source. To obtain a 13.2 volts, it is common to use either a NiCd or NiMH power pack. These usually come in the form of double A sized batteries. Eleven pieces of 1.2 volts 2500mAh Energizer batteries will yield a 10% increase in brightness or 1180 lumens. Ten to 11 AA cells will give a theoretical runtime of about 1.08-1.5 hours depending on battery capacity. Adding a 12th AA cell to the power pack will give 14.4 volts. Another alternative would be to use Lithium ion cells (each is about 3.7 volts). Therefore, 4 of these will provide aroud 14.8 volts. Overvolting the MR16 bulb to this level will yield about 1555 lumens, This is brighter than any LED or HID system. The only problem with providing such massive amounts of power through the bulb is a drastic reduction in bulb life. A bulb rated for 2000-3000 hours when overvolted may fail in as little as 250 hours. If one rides for 1.5 hours daily, that's still almost half a year. MR16 bulbs are cheap (about $2-$5) so replacement cost is not a big issue.
What some DIY'ers fail to account for is the resistance of the wire they use for the entire setup. If voltage measured at the bulb (when the light is running) is compared to the voltage at the battery end (again, when the light is powered up), a substantial voltage drop should not exist.
For example, a fully charged battery with 13 volts on the terminals vs 12 volts when measured at the bulb end. This delta V of 1 volt computes to an unacceptable resistance of 1.3 ohms.
Using values obtained when the SJL is running, we get:
Delta V = 12.45 - 12.15 = 0.3V
V= 0.3V
W= 20W
Delta V = I x R (ohm's law)
P = I x V
I = P / V
Delta V = (P/V) x R
R = (Delta V)/(P/V)
R = 0.3V / (20VA / 12.45)
R = 0.3V / 1.60
R = 0.1875 Ohms
This result is fairly acceptable. If it was higher, then the resistance of the wire may be too big such that it produces a voltage drop. This is usually remedied by a lower gauge (thicker wire). Also note that smaller wire (higher gauge number) may suffer from heating with current passing through them, thus further increasing resistance.
The next table shows some of the better known high-end models for bike lights. These are mainly designed for offroad and professional use and are very robust. The Busch and Muller Big Bang is the first Gas Discharge lamp certified and approved by the very strict German stVZO. It's probably one of the best lights available, but at $1100 plus taxes, it's about the same price as a good bike.
The light and motion Seca 1400 @ $770 has a trail and downhill rated outpout of 1400 lumens from 6 LED emitters. It's small and has a runtime of 2.5 hours at the brightest setting. My only issue with this light is the plastic housing which may not be as sturdy as the metal housings of comparable lights. However, L&M has been manufacturing lights for both biking, caving and diving applications for a long time and I'm pretty sure that their material of choice has been thoroughly tested.
If I were to purchase any of these lights at retail price, my first choice would be the Cygolite TridenX. At $350 it's not cheap but it sits right in the middle of the cheapest and most expensive lights in terms of features and specs. Good runtime, high brightness and metal housing. It's more than adequate for professional downhill use and definitely more than enough for the casual biker.
Going back to the SJL (spice jar light), the final cost of about $25 more than compensates for its weight. Two kilograms is very heavy considering that most bike light systems are featherweights at around 500 grams. The SLA battery contributes to 90% of this weight. Fortunately, I have access to a lithium ion system. Eight 3.7 x 2200 mAh Samsung lithium ion batteries would be more than sufficient to produce a pack of 14.8 volts and 2200 or 4400mAh. Fourteen volts will overdrive the MR16 to HID level output. The first four cells could be wired in series two make the first pack and a second similar pack could be attached in parallel to increase capacity. Alternatively, a single 2200 pack could be used and the second 2200 pack could be swapped in once the first pack is depleted. A single 2200 pack should be good for more than 1.5 hours with each pack weighing just a little bit more than a similarly sized AA pack. With either litium ion cells (18650 x 4) or NiMH (AA x10)weight is reduced to less than 750 grams for the entire setup.
Version 1.5 of the SJL is already in the works. This will utilize 4-LED emitter array on an MR16 format. The new "bulb" is rated at around 250-360 lumens and while it is about half the brightness of the halogen MR16, it will be using only about 80 percent less power (or 4W). This will be very close to the SLA's C=0.05 and should provide at least 15 hours of continous runtime. I have no practical use for such a lengthy runtime, therefore if light output is bright enough, I may switch to Li-ion pack. A 2200 pack should give a theoretical runtime of 8.14 hours @ 14.8 volts.
Another advantage of my SJL design is that the bulb holder portion is interchangeable. Two small screws hold the lamp holder. A second lamp holder assembly will be used to hold the LED MR16, this way, it would be easy to switch back and forth between the halogen and LED emitters. The connection to the battery is also easily adapted to use any power source, so it would be fairly easy to switch between the SLA and a NiMH or Li-Ion pack.
