[Outbound]

Hey guys! Long time lurker who's been reading as a guest for a while now. Finally signing up.

Really impressed with the in-depth knowledge you guys go into, and want to help further knowledge a little more! I am an optical design engineer by trade, focusing on the automotive aftermarket developing off-road lighting and other specialty lamps for OEM manufactures. I have extensive access to simulation tools like OPTIS and wanna help develop a database of sorts to help people learn and chose the right optics for their bike lights.

I can tell some users are really already on top of their game in regards to spying the differences between reflector bowls (short/wide vs long and narrow), can discern about how more lumens is NOT always better (thank you!) and a focus on beam pattern quality over peak numbers. Though one thing I don't see too often is discussion of lux numbers. Very easy to take your own lux readings when comparing lights, and I think would make comparing lights even more analytical vs subjective. I use this one around the house when I'm comparing lighting.

https://www.amazon.com/Dr-Meter-LX1...UTF8&qid=1503533140&sr=8-3&keywords=lux+meter

If you take measurements in a consistent way (at 10 meters or 32.8' away) in a pitch dark environment, can help better figure out what's making good lighting. Obviously beam patterns play a MUCH larger part in peak lux readings, but can also help identify which lamps will let you see further than others. I've complied some comparisons of my own on niterider, Ion 700, and a whole gaggle of cheap chinese lights I've bought. Some surprising gems, some total **** lights, haha.

Anyways. Thought I'd introduce myself, and feel free to shoot me any questions regarding illumination engineering. What sources are great, what's coming in LED's, the stuff you only really get to hear about when you go to these large automotive lighting conferences in Detroit and Germany, how we design lights using raytrace software, etc. I'm an open book!

 

[Vancbiker]

Welcome to the forum!!!!

I've played around with machining parabolic reflectors and attempts to create shaped beams using modified commercial parabolics. None of them were worth trying to incorporate into a light.

Having someone around that knows what works or why something that one thinks might work won't, will be a great resource.

 

[MRMOLE]

Welcome!!! Sounds like I'm going to enjoy your contributions here.
Mole

 

[ledoman]

Outbound, nice to see you here.

I've question for you already. Here on the forum we had some discussions about efficiency of reflectors (+ some glass lense to cover the front) vs. TIR optics. It is hard to compare directly since reflectors can be OP or SMO and TIR lenses might have dispersion surface, but can you give us some basics which are better for some situations? Shurely at same size and same source of light (ie. leds we use here, XM-L, XP-L, XP-G,...etc)

 

[MRMOLE]

If you take measurements in a consistent way (at 10 meters or 32.8' away) in a pitch dark environment, can help better figure out what's making good lighting. Obviously beam patterns play a MUCH larger part in peak lux readings, but can also help identify which lamps will let you see further than others. I've complied some comparisons of my own on niterider, Ion 700, and a whole gaggle of cheap chinese lights I've bought. Some surprising gems, some total **** lights, haha.

Love to see the data you have on your lights. Especially interested in Niterider data since it's a very popular brand yet few if any of the regular posting members own them. I personally use a 2 meter distance for my light meter tests only because that matched the method of MTBR's Bike light comparison which is my largest data base on current bike light performance information.
Mole

 

[scar]

Welcome. What kind of lights are you currently using?


****

 

[Outbound]

Outbound, nice to see you here.

I've question for you already. Here on the forum we had some discussions about efficiency of reflectors (+ some glass lense to cover the front) vs. TIR optics. It is hard to compare directly since reflectors can be OP or SMO and TIR lenses might have dispersion surface, but can you give us some basics which are better for some situations? Shurely at same size and same source of light (ie. leds we use here, XM-L, XP-L, XP-G,...etc)

Great question, and can be a very complicated and variable answer. So the right answer is.... it depends. Like a lot of things in life, haha.

I will apologize in advance, I tend to kind of ramble and sometimes can go on a tangent. Always feel free to ask me to clarify something if there is any confusion.

Anyways. I’ll start with the basics. Materials. So as you already know, there are transmission losses when passing through a material, some materials such as pure glass and optical grade silicones have much higher transmission than materials like PMMA, or worse, PC. However they are only account for between 4-8% of the energy loss. What’s the major driver when it comes to TIR vs reflector discussions is the size of the TIR vs the design and sizing of the reflector.

TIR sizing plays a large part in this because the smaller a TIR is, the harder it is to actually collect the light from a larger source such as an XML, or really any CREE source that isn’t the new FX chip. Therefore you tend to lose a lot more light through scattering. I ran a quick simulation of a 10mm tall, 22mm outer diameter TIR that’s designed to be a pure spot, put an XML2 chip at 1200 lumens and simulated at 20 million rays. an see in the righthand side next to the Flux measurement, 915 lumens, so divide that by 1200 original lumens coming from the source, and we have optical efficiency around 76.5%. Now to demonstrate the sizing effect, I kept the same designed spot pattern, but made the TIR 8mm tall, and an OD of 18mm. The new flux measurement is 810 lumens, so the efficiency is now 67.5%. Efficiency dropped almost 10% by reducing the overall diameter of the TIR by 20%.





Now let’s go the other way. Increase the TIR size by 20%. So now we have a TIR OD of 26mm and a depth of 15mm (outer diameter way more important). Now the simulation shows 1054 lumens, so that’s an efficiency of almost 88%! It’s fairly linear. Smaller = way worse, and bigger = way better.



So a savvy engineer can try and balance out the size he/she has to work with for the product, figure out the sources they are working with, and determine if two tiny TIR’s with mid lumens are going to be better than one big TIR. There are fabrication challenges with a large TIR (talking over 30-40mm size) as well as very expensive tooling due to things such as shrinkage. Generally when you get into a TIR that big, you have to have it custom designed (as one should for every optic IMO) and can start to better shape the beam pattern to the desired effect. I am always a fan of a large optic over large lumen sources. Whenever I see lamps with tons of tiny TIR’s and lots of LED’s I cringe. It’s such a waste of energy. Having to drive the sources harder, generate more heat, suck down more power, when could have gotten away with half the lumen output and still have the same beam pattern and intensity just by going with a larger well designed optic. However there is a point of dimmishing returns when it comes to TIR since at some point, the transmission losses account for the majority of the loss compared to optical efficiency. So a well designed light will start with the largest and most efficient optic, and work down from there.

---------------------------------------------------------------------------------

As for reflectors, there are two types as you know. The parabolic bowl types that you see on things like NiteRider and lots of generic lamps coming from China, then the downward firing reflector that you see on things like the Phillips SafeRide, and high end lights like the B&M IQ series. I assume you are questioning about the bowl type so I’ll mention those for now. In terms of losses due to materials, there is now the protective lens, and the loss through the aluminized reflector itself. These combined add up to about the same as a TIR loss in terms of materials, 3-6% but will be a little less. So what should really be focused on is the actual reflector design, not materials. The type of lens is almost inconsequential. Glass vs PMMA vs PC amounts to a 0.5-1.5% difference when talking a 2mm thick lens.

The big difference is how much light is where you actually want it. So these simulations are going to show why I hateeeeee shallow bowl optics. Keep in mind that on the measurements of the lumen amounts on the TIR optics, it was measuring almost 45* left and right, and 45* up and down, which is MASSIVE and totally beyond the actual FOV that a person would really experience. The real FOV, and what I should really be measuring for true opticial efficiency is around +/-40* horizontal, and -35* to 10-15* vertical. But for the sake of clarity, I’ll stick with an overall sense of when someone shines a light on a wall.

So let’s start with a bowl similar in size and such as the original TIR we started with. 10mm tall, 22mm diameter. You’ll see a surprising answer. It’s actually more efficient in terms of “to the wall”. 1017 lumens verses 915 lumens on that similar sized TIR.



However that’s measuring up to 45 in the sky…. Think about the last time you rode your bike. Were you craning your neck up 45* to look at a tree or something? If you measure out beam angles you’ll find that anything above 15* is semi-useless. Up close there is enough scatter from the light itself and light bounces on the environment that you’ll still spot objects such as a tree branch or something with plenty of time. So with this in mind… let’s revist that measurement and figure out how many lumens are actually “useful”, such as between the FOV horizontally, and the 15* up, and -35* down (35 or more and you are basically lighting up your bike tire excessively, another discussion for another day). Now we can see the actual useful lumens is 813, or 68% efficiency.



So there are obviously some ways around this, can make the bowl oval like to try and make the beam pattern a bit elliptical, or can make the bowl deeper to try and collect a bit more, but then at that point you are starting to sacrifice beam width for the collection of excess lumens going where you don’t want it too. Others use a hood or something, but that’s just also kind of a waste unless the hood is designed to also reflect light onto the ground. Every hood I’ve seen is just flat black.

Going bigger doesn’t necessarily result in better performance either! So this picture shows the similar TIR size that gained us 88% efficiency, (15mm tall, 26mm OD) and we are at 940 “useful” lumens vs the TIR had 1054. Or 78% vs 88%. Can see the overall width of the pattern tightened up a bit, but the actual useful “on target” lumens didn’t.



--------------------------------------------------------------

So how do we boost that optical efficiency? We determine the ideal beam pattern for the riding that someone wants. Cutoff vs Flood vs spot vs whatever. Mostly determining the beam angles. Then you can backtrack those beam angles to figure out how much light you want in certain areas and redirect some to those areas. This is where powerful optical modeling tools are needed, and even then it’s not as simple as clicking a button like it was for me to generate these TIR and parabolic reflector models. It takes a lot of engineering, a lot of experience, and a LOT of time. I’m happy to dive into that topic a little later, share the math behind it, the perceived brightness vs actual, etc. Figure this is a good starting point.

 

[Outbound]

Love to see the data you have on your lights. Especially interested in Niterider data since it's a very popular brand yet few if any of the regular posting members own them. I personally use a 2 meter distance for my light meter tests only because that matched the method of MTBR's Bike light comparison which is my largest data base on current bike light performance information.
Mole

I'll compile some of the preliminary stuff I wrote down. But when I measure stuff without the integrating sphere (has it's place, but it's not the whole story) I focus on Lux at 10m (I strongly disagree with the method that MTBR does) and then measure beam pattern width at 1m or around 3' just because it's far easier to measure on the wall at that distance. Then I look for things like concentrated hotspots, what the lux readings are at the edge of it and can backtrack those numbers to get a rough idea of candella and intensity. I've successfully been able to then replicate that optic inside OPTIS with the sources they are using. It's fun for me.

Generally though off the top of my head. Ion 700R = 60 lux peak, tighter hotspot but nice wide spot optic. I like it personally. NiteRider 1400 had a peak lux of 65, wider hotspot though and about the same beam width as the Ion 700, in the 60* range for overall, and hotspot is around 10*.

 

[Outbound]

Welcome. What kind of lights are you currently using?


****

Currently, when I go on a ride (both MTB and road) I use a lamp from Sate Light that I had gotten when I was evaluating a lot of lamps. Large TIR optic, 700 lumens. Produces a very nice rectangular beam pattern with about a 40* width, and 15* total height.



Though I've been designing my own bike light using my knowledge from the automotive side, so once I get those prototypes going I'll be riding MTB with my MTB focused optic, and then put my road-going version on my road bike. Road one will basically meet low beam requirements for automobiles, while the MTB is focused more on beam width and homogeneity.

 

[ledoman]

Outbound, great answers, thank you for your detailed explanation. So we can conclude in general (over the thumb) the TIR produces more useful beam pattern for our needs, right?

It is great to have you here with all that knowledge and experience. Hopefuly the others would have interesting questions too. For me the next question would be SMO vs OP reflector surface. We all know the OP reflectors make beam looks nicer, less ringy, but again how much do we loose. Well it is just academic question since the lumens we have are big enough anyway.

 

[Vancbiker]

.......So we can conclude in general (over the thumb) the TIR produces more useful beam pattern for our needs, right?

IMO, no. Most cases, I'll take a reflector beam over a TIR. The best reflectors for my beam preference are nearly as deep or deeper than their diameter. A shallow bowl type reflector throws too much like out to the sides. A deep reflector gives a sharper cut off between the lit and unlit areas. I don't care for lighting up the trail to the side of the bike. For me that just creates shadows that are eye-catching and a distraction. If the trail is tight and twisty, that's what the helmet light takes care of. Look at where you want the bike to go, not at the sidelines.

 

[Outbound]

Outbound, great answers, thank you for your detailed explanation. So we can conclude in general (over the thumb) the TIR produces more useful beam pattern for our needs, right?

It is great to have you here with all that knowledge and experience. Hopefuly the others would have interesting questions too. For me the next question would be SMO vs OP reflector surface. We all know the OP reflectors make beam looks nicer, less ringy, but again how much do we loose. Well it is just academic question since the lumens we have are big enough anyway.

Well you need to classify the optics into 4 camps basically. TIR (standard, what was discussed), Parabolic reflector (what was discussed), but then engineered TIR (such as the ravemen cutoff one, I'm impressed with that, or the Sate Light one that I use right now), and then most importantly, the engineered reflector, such as what's used in automotive OEM's, or high end lights like the B&M IXON IQ, or most of their IQ series (Busch & Müller: IXON IQ Speed), or the Supernova series (Homepage |Supernova Lighting Systems).

The latter two when designed properly will outperform any TIR or parabolic reflector. However where do they fall short? Cost. Because they are expensive to tool up, and expensive to design. The optical software I use is $50k for a license, and that's on the low end with minimal features. That's why a properly engineered lamp will be upwards of $250-300. However, you get a much more usable beam pattern, with more efficient optics in the term that you are getting more "bang for your lumen", not necessarily your buck P).

My overarching recommendation for a cheap light (under $80 or so) for a road bike would be to trend towards a larger TIR simply because you can focus the beam a bit better, and on a road bike your effective FOV is narrower than on a MTB trail where you are constantly moving the handlebars around left and right. However on a MTB trail, it's kind of a mixup, the reflector bowl with a segmented reflector (like Niterider) or an OP type surface can be good because you get large beam width, and because generally at lower speeds and twisty areas, you can get away with a lower peak intensity in the beam pattern overall, but a properly designed TIR can make this irrelevant too. So it's really not a catch-all.

That's why for optical design, there are so many things to weigh when it comes to how effective a lamp is. For a headlight, we can make premium headlights using a single OSRAM Black Flat 1x4 or 1x5 running at 1400-1500 lumens that will let you have a highbeam and wide lowbeam that you can feel safe hitting the highway at 80-90mph with, but it's because it's a tiny point source, backed by a large (relative to bikes) reflector. So more lumens isn't always the answer, and always happy to see people reiterate that when someone brings up a 3500 or 5000 lumen lamp.

Lumen ratings, how they are affected by thermal, all great discussion points that I know some of you guys are well versed in.

-----------------

Regarding the OP vs SMO, I can't say definitively because I've honestly never done any serious optical design with OP in mind. In automotive optics you are either boosting the optical efficiency as much as you can by segmenting reflectors, working with lenses, or shutters. I've never run across OP beyond some cheaper bike lights and flashlights. I couldn't even find a surface in my database that could simulate it! However I did make an extremely segmented bowl optic with a small radius to try and simulate OP vs SMO, and the results do seem pretty close so I'll share it.

As I mentioned in the other post, the optical efficiency between the two will be relatively the same when aiming at a wall. However as we mentioned, efficiency is just defined as the number of lumens hitting the targeted area. In that regard, they are exactly the same. However now what's happening is that you are smoothing that peak intensity out and spreading it across, something you all know very well. So what does this do for peak performance? Well first need to define what we are saying peak performance is. Is it the number of lumens hitting that effective FOV like I mentioned before? (+/-40* Hor, -35 to +15* vert), is it the peak lux (how far you'll see an object) or is it the brightness across the beam pattern?

With the OP, you'll see DRASTIC reduction in peak intensity, using the 1200 lumen XML2 source, and a "mid sized" OP bowl, we get a peak intensity of 30,000 Cd, but lets take a look at a cross section of the beam profile and compare that to a smooth one. The peak intensity of the “smooth” parabolic is 5 times higher, but the beam profile is much sharper.





So this means that when the handlebars are moving around you’ll notice that hotspot FAR more than the OP reflector. This is why the NiteRider bowl is actually pretty good. The peak intensity is a bit lower, but it smooths out the peak intensity so that a hotspot isn’t as noticeable compared to a smooth reflector. Great for MTB, but a lot of wasted light still going upwards where it’s not needed as compared to focusing that wasted light somewhere useful. While the smooth allows for a very peaky but intense beam, which might be good for road riding depending on what your needs are.

 

[MRMOLE]

Test Methods

I'll compile some of the preliminary stuff I wrote down. But when I measure stuff without the integrating sphere (has it's place, but it's not the whole story) I focus on Lux at 10m (I strongly disagree with the method that MTBR does) and then measure beam pattern width at 1m or around 3' just because it's far easier to measure on the wall at that distance. Then I look for things like concentrated hotspots, what the lux readings are at the edge of it and can backtrack those numbers to get a rough idea of candella and intensity. I've successfully been able to then replicate that optic inside OPTIS with the sources they are using. It's fun for me.

Generally though off the top of my head. Ion 700R = 60 lux peak, tighter hotspot but nice wide spot optic. I like it personally. NiteRider 1400 had a peak lux of 65, wider hotspot though and about the same beam width as the Ion 700, in the 60* range for overall, and hotspot is around 10*.

As I mentioned before I patterned my testing methods after those used by MTBR but only because it allows me to compare my results with what they get and since I can't afford to buy one of everything, allows me to have a much broader knowledge of available lights performance characteristics. I'm probably not going to change for this reason but am curious to hear your thoughts on 2 meters vs. 10 meters for the bounce test and center beam (max lux) tests I do. Appreciate your help with this.
Mole

 

[Outbound]

As I mentioned before I patterned my testing methods after those used by MTBR but only because it allows me to compare my results with what they get and since I can't afford to buy one of everything, allows me to have a much broader knowledge of available lights performance characteristics. I'm probably not going to change for this reason but am curious to hear your thoughts on 2 meters vs. 10 meters for the bounce test and center beam (max lux) tests I do. Appreciate your help with this.
Mole

It's mostly because it's basically the industry standard for automotive and for the STVZO, so I am used to it, and far easier for myself to make mental comparisons between true vehicle headlamp performance and a bike light performance. Though the "bounce test" still baffles me. Doing a peak lux measurement at 2m away is just as good as 10m away, further is "better" per say because it's less sensitive to where the meter is, mostly when taking beam comparison measurements. Such as 3m from the center, or 1m below the hotspot, etc.

You can backtrack your measured lux numbers (direct measurement, not the "bounce" test) using this calcuation to get the candella amount. Lux to candela (cd) conversion calculator

Then from there you have the candella of the beam pattern, or roughly thereabouts, and can generate the lux at different distances. (Candela to lux (lx) conversion calculator) What you can sort of do is compare your measured candella amounts for products that are FL1 certified since they should be publishing their peak candella, and can kind of figure out your meter offset from there. A lux meter is sort of like a car dyno, you can't really exactly compare two measurements from two different locations on different types of lux meters. A professionally calibrated lux meter is going to read differently than the $30 one I linked in my first post. However I like to recommend that cheap one because it's a a cheap meter that is fairly reliable, and that enthusiasts can pick up to become better at understanding lighting.

If you guys are interested, I can setup a diagram that sort of outlines my typical standard test procedure, or how to make a test procedure that should be easily replicated in your basement (so no 10m needed, could try 2 or 3m (6 or 9' roughly)), or a dark room in your house. Though the main problem with a smaller room is the bouncing effect, it'll throw off the actual lux numbers when measuring past the hotspot.

As with most tests, I'm not going to say you are doing anything wildly wrong (only if you start comparing data that's measured differently). Most important thing is to do something consistently. While I strongly disagree with the bounce test thing because it introduces a few variables and isn't really a measure of real-world performance.... as long as it's being done consistently and can be replicated, it'll count for something.

I'm lucky enough to have a goniphotometer at work, as well as a very large space that we use for testing automobile headlights, that maybe I can mount up some of the more popular headlights on here so you guys can get a true apples-to-apples comparison of lights.

I know I plan on using it to validate the bike light I am designing compared to my simulation results!

 

[garrybunk]

Glad to have you here and sharing your knowledge. I hope you're also sharing knowledge over at BudgetLightForum.com where users geek out much more than us here.

About the optic vs. reflector discussion - I know this isn't a "budget" forum, but I think many of us are somewhat budget minded, so the ideal optic/reflector designs are really out of reach for us I guess. We really want "bang for the buck" over "bang for the lumens" (again, many of us). I think us MTBR's are trapped in the situation where manufacturing costs are so high to build something for a market that is so small.

I'm using the same lux meter (after seeing horrible #'s from a cheap HS1010). I really should take lux throw measurements with my lights. (I started to back with the HS1010 and that's when I discovered it was way off, but never revisited it). Would it be a lot better/more reliable to do these lux readings outdoors vs. inside?

-Garry

 

[Outbound]

Glad to have you here and sharing your knowledge. I hope you're also sharing knowledge over at BudgetLightForum.com where users geek out much more than us here.

About the optic vs. reflector discussion - I know this isn't a "budget" forum, but I think many of us are somewhat budget minded, so the ideal optic/reflector designs are really out of reach for us I guess. We really want "bang for the buck" over "bang for the lumens" (again, many of us). I think us MTBR's are trapped in the situation where manufacturing costs are so high to build something for a market that is so small.

I'm using the same lux meter (after seeing horrible #'s from a cheap HS1010). I really should take lux throw measurements with my lights. (I started to back with the HS1010 and that's when I discovered it was way off, but never revisited it). Would it be a lot better/more reliable to do these lux readings outdoors vs. inside?

-Garry

Well, I guess begs the question, what is considered "out of reach"? Generally what I've found in developing my own lighting is that the good battery pack, as in samsung/panasonic cells, hard case, silicone straps, etc. is what makes up half the cost of the high end stuff. It can cost almost $25-35 from the manufacture itself, add in the high cost of shipping heavy cargo when buying 1000 of them, not to mention the tooling for a custom hard case, and can understand why companies like Lupine, NR, and others have battery packs pushing $100-150. But can safely bet they are going to hit their runtime and battery output numbers.

I've tested out some "claimed" 5400 Ah battery packs that were under $7 from a manufacture, and they'd be half the claimed output, just like claimed lumens. I know some of you guys do some excellent work testing out battery packs just like I have my electrical guy do. With load testing and actual logged measurement, which is freaking awesome to see.

So I guess begs the question as well. What if there was a good, well designed lamp that you could get without the battery pack? That'd probably bring pricing down to something that's more "budget minded" for the lighting enthusiast who already has a bunch of good battery packs. I know Lupine, NR, and L&M use proprietary connectors, which I understand... from a business sense.... but hate it from the tinkering point, as I am sure you guys do too.

--------------------

Regarding the lux testing. If you have a large enough space like an open basement then testing down there can be pretty reliable. However if it's a small 10x10' white walled room, it's going to be pretty inaccurate and better off doing it outside.

However like I had mentioned before, the best thing is simply consistency. If you are outside and already have a 3-5 ambient lux reading from surrounding light like a streetlight or something then that's not a worry at all. It doesn't need to be pitch black. What you can do is just take your ambient reading at a few points you plan on measuring, and then subtract that from your actual reading to get a fairly consistent reading across the board. Some lux meters you can zero out, but I don't believe the one I linked can.

If doing it outside, I'd mark off a consistent distance that you can setup, shine it against a large wall or garage door. Doesn't have to be 10m, it can be 3m or 5m. The key is to accurately state what distance you measured at. That way it can be backtracked to candella, and someone else can compare their measurements to whatever distance they measured at, or whatever distance they want to potentially know the light intensity would be.

 

[garrybunk]

Right on about the battery packs, which is why we requested lighthead only on our custom build (which in turn makes it harder for the average person to deal with, even though they sell Panasonic packs which they could lump in as a complete kit offer). I totally agree on the quality battery pack, this is no place to cut corners.

-Garry

 

[Staypuft1652]

Outbound: thanks so much for sharing your expansive knowledge!

 

[angerdan]

Specialized Flux Pro

Very interesting and unique public project!
But why create something which is already available by Specialized (Flux Pro & Expert)?

 

[Outbound]

Very interesting and unique public project!
But why create something which is already available by Specialized (Flux Pro & Expert)?

Just because someone else has had the same concept doesn't mean I don't think it can be improved on. That was one of the few lights I saw that had the right idea, but I felt could be improved on with a deeper reflector and a wider spread.

It seems that the light wasn't the best in terms of reviews, and honestly, haven't heard of anyone actually using one except for a few european sites. Hence I set out to develop my own.