Mtbr Lumens testing
We're done for now!
The good news is we used two labs, Light and Motion and Lezyne. Mtbr was present the whole time and all the data was saved directly into files. Lumens readings were recorded in 1 second increments for about 5 minutes.
We logged the lumens reading at the 31 second mark to follow the FL-1 Flashlight standard. This gets rid of the 'switch on spike' that some lights have. This reading is always slightly lower than the 1 second reading but it is a much better representation of the light.
We recorded 3 minutes of run time to show how the light behaved over time as it warmed up. We did not want to do a whole battery run because we were not confident that enough air was moving through to cool the light from the direct correct direction to simulate riding (front to back). Since the light head was sticking into the sphere, there's no way to cool it from the front.
Our lumen readings are more accurate this year than last year because we learned to calibrate each reading according to the lumen loss that each light head represented. Because each light head is stuck inside the reflective sphere, some of the light is absorbed by the light head instead of being reflected. This varied anywhere from 2% to 10% loss. After that was measured for each light head, it was added back in the lumen reading.
The Light and Motion Taz is a good example of light that was tested in both laboratories. It measured about 1205 lumens at each lab. This gives us great confidence that the numbers from each lab are reliable and repeatable.
Light & Motion Taz 1200 – 2013 Mtbr Lights Shootout | Mountain Bike Review
Reading from Light and Motion lab
Reading from Lezyne lab
We will compile all the lumen graphs in one article.
Main shootout page is here: 2013 Bike Lights Shootout – 50 LED Lights Measured | Mountain Bike Review
francois, I just wanted to take the time to thank you for putting together a fine review and for going through all the trouble to gather all the data for all the various light systems. I find the results of the lumen test fascinating.
As a regular user of the "lights and night riding" sub forum I'd also like the point out that some of the standards of the " FL-1 flashlight standard ", ( ANSI / NEMA) are not as useful for rating bike lights as one might think. In the review you adopted the FL-1 standard of rating an output just after a couple minutes. This makes sense for flashlights because usually flashlights are just used for quick on/off purposes. Bike lights not so.
I'd rather see a lumen rating after about 10 minutes. My experience with bike lights tells me that most lamps drop in output after the emitter has had a chance to warm up. This usually takes no longer than ten minutes. With some lamps it will happen much faster if the lamp is over-driven or not designed to dissipate heat well. Other lamps that may be under-driven or better engineered may take longer to stabilize. This is real important to know when making a decision on which lamp to buy. Lamps that are better engineered to handle heat may appeal to people in warmer climates. As such a 10 minute read-out would be the better way to go.
While lumen ratings are really important I'd also like to point out that without a "Peak Beam Intensity" rating you are only getting about 33% of the picture. The *PBI rating would tell you "how intense the sweet spot is on a lamp at a given distance". This would help the buyers judge the throw of the lamp which is another 33% of the picture. The last 33% is supplied by your beam shots which help the buyer judge beam pattern ( or beam width ) All three together they can supply a really good idea of what a given lamp is capable of doing.
( *Note: Both lumen and Peak beam Intensity rating for bike lights should be performed with a given amount of air moving over the lamp. Real important because unlike flashlights, bike lights are designed to be used "While moving". ) Testing bike lights without moving air is like testing car engines without oil.
Peak Beam Intensity... interesting. I looked it up on how to measure it:
"Peak Beam Intensity
Light intensity in candela (cd) at the brightest part of the beam. A lux reading in the brightest portion of the beam is taken at some distance (2m, 10m, or 30m) with the light on its brightest mode and tightest focus at some time between 30 and 120 seconds of turning the light on. The lux reading is multiplied by the square of the distance in meters to get candela. Regardless of the distance at which the reading is taken, you should end up with the same value in candela, therefore the measurement is independent of distance. A lux reading taken at 1 meter distance is the same as candela (since 1 squared is 1)."
It could definitely be useful but it could be misleading as well when viewed without an accompanying beam pattern photograph. Maybe I'll publish it only on the beam pattern photo. The narrower the focus of the light, the higher the PBI will be (all other things equal). And narrow=bad for cycling.
The good news is there is a new crop of dual-beam lights which are engineered for throw as well as width.
It would be cool if I had a light meter that remembered its max reading for a given cycle. That's because these lights are very difficult to aim perfectly at some distance. I could just move the light around and the light meter will find the max.
Love the Lumen graph! Adding the PBI at 30 meters would be helpful as a reference. Some lights have plenty of lumens, but they just don't send it down the trail enough to be useful at speed.
Thanks for all the hard work!
Thanks for the "labor of love". The information you've compiled and shared is very informative and greatly appreciated.
For my preference, narrow is good for cycling. I only want a slight spill from the bar light, a dimly lit ~45 degree sweep. I want the rest of the light down the trail in a brightly lit 10-15 foot wide spot at 100 feet. If I need to see outside that, I'll use my helmet light. Lights with wide beams create moving shadows to the sides that I find distracting in my peripheral vision.
Originally Posted by francois
That's why the PBI reading needs to be compared to both the beam pattern and Lumen output to be useful. It is also why no one statistic or beam photo tells the whole story. Some people want a more intense hot spot while others want more of a flood beam pattern. Lastly some want that mixture of flood and throw. Having all three parameters available would significantly help the consumer determine which lamp is best suited to his or her needs.
Originally Posted by francois
Here's an example of how that would work: Suppose a person already owned a lamp like the Mj-808E. Having used the lamp they would be familiar with the beam pattern and usable reach of the lamp. Now suppose they used the 808E on the bars but decided that while the 808E had the needed throw they now wanted more peripheral light while still maintaining the same throw. All they have to do is look up the PBI rating of the 808E and they now have a number to help them judge throw. Then they look to the beam photos to see which lamps have wider beam patterns then the 808E while still maintaining the current PBI rating.
Once those facts are known it then depends on which set-up has the best features or best bang for the buck value before the final decision is made on what to buy. Anyway, that is how I look at it but WTH do I know....
Just want to say that those graphs will be most appreciated here and that you're doing a fine job testing the lights mtbr has.
Cat has a valid point here. I totally agree. I, for one, am a newbie in night riding and just started looking and doing some research on these kind of lights around 1 month ago or even less and this has happened to me. A lot of people (judging from forums) go for chinese clones has their 1st bike light because they want the best lumen for $ ratio BUT they do not know exactly what kind of beam pattern and intensity of light in different areas they need... So, the 808 clone or even the original (they seem to have the same beam pattern, despite all the other differences) is an excellent reference for us when looking over your measured lumens and beam patterns. However, as Cat said, having also the PBI would definitely complete this because we would then have another number (how bright a light really is on its strongest spot) we could use to compare with other lights. Our eyes have seen it in real use, so we would have a number for that.
Originally Posted by Cat-man-do
We would have all of the variables we needed to make a final decision.
I am not criticizing or your work - far from that! - I'm just saying it would make it 100% complete (at least for newbies).
The latest spreadsheet is here:
Guys, make me some cool graphs.
p.s. The Magicshine 872 has been downgraded to 920 lumens instead of 1200. After further testing, it would drop precipitously to 920 lumens. No amount of high speed fan could get the output back up. It does seem like they are overboosting the light for the first 30 seconds to make it look good in photos and measurements.
1200 to 920 is quite a drop. Thanks for that Francois.
Magicshine...why am I not surprised?
A pretty good indication that the heat path from the LEDs to the body is poor.
Originally Posted by francois
Yep, the Lumintrek looks like the exact same light and it measured at 859 Lumintrek TrailBlazer 1600 – 2013 Mtbr Lights Shootout | Mountain Bike Review
Originally Posted by Vancbiker
And the Fluxient 4xR5 1600 looks the same too and measured at 879
Fluxient 4xR5 1600 – 2013 Mtbr Lights Shootout | Mountain Bike Review
So that made me dig a bit deeper into the Magicshine 872
This proves my point about needing more run time to judge lumen output. This is why I suggested running for 10 minutes ( with fan ). Initial output is always higher while the emitter is cold. Giving the emitter time to stabilize ( at a working temperature ) gives a more accurate representation of what the lamp is capable of. ( you do need to use a fan though if running more than 2 minutes. ) Also informative to take a before and after temperature reading on the lamp head as well. An infrared thermometer is useful for doing that.
Originally Posted by francois
A little bit of looking at the Cree datasheet gives some interesting info..
Looks like the LEDs are running at ~135C (235F) to produce that type of output drop. Wouldn't expect that to be very good for them, though the max is 150C.
Last edited by fc; 12-17-2012 at 09:55 PM.