Beyond German Standards - road light experiments

I will post as I go on from this point.

The dyno light on the errand bike is in this thread. Which includes a link to a ride-by video. It is to be paired with the helmet light. The Helmet light's wall beam is in this post and the light and its hood in this post in that same thread.

The Mercian has a pair of lights now. The driver is an H6Flex, reduced in diameter a mm or so to fit the Marwi housing:



For those unfamiliar with thei light body, it has a cylindrical 1" bore, followed by a cone bore that drops to 22 mm ID at the truncated tail where a tailcone/mount fits. So the driver got copper leads removed from an old computer mother board soldered into the interface pads and angled inward so as not to short on the tapered rear of the body.

The Marwi on/off switch was opened, carved a bit and a NO MOM SPST switch from a Bflex and with the mounting legs removed, was siliconed in place. EL34 now has replacements that will save anyone else the trouble unless like me, they have some Scot in them!



A 1/8 " (3 mm) aluminum disk was fashioned cutting it 16 sided on a saw, then filing and grinding. An LED lead passage was notched in its bottom side. It was heat taped (supplied by George of Taskled with the H6Flex, Thanks BTW, if you read this George). Then the assembly was placed with the bottom edge as far back as it would tilt to clear the LED leads past the cutout in the rear bottom of the shell. Thermal load of driver about 1.5 W maximum.

Before taping and assembly:



A split ring copper sleeve was fashioned from a 1 1/2" copper pipe union and reduced in circumference so when fully closed with a hose clamp it can just be slid with some force into the body:



The tail was tapered and fitted as close to exactly to the angle of the aluminum driver mount as possible by eye. A 1" pipe cap (31 mm OD) was sanded to a flat surface as was the XM-L star. Mounting and wiring holes were cut and mounting holes tapped for No. 4 screws. Brass bevel head # 4 screws had their heads reduced in diameter using a drill as a lathe. They were prefitted to clear the holes of cuttings. The cap was then cut at the same angle as the sleeve with a thickness at its top edge equal to the thickness of the end of the cap to get the 12 mm mounting depth plus MCPCB for the Eva lens. The pipe cap/LED firewall was ground/filed/sanded to an inteference fit inside the copper sleeve. The sleeve installed first with some AA, more was places on the Al disc where the copper pipie cap would sit, the LED was mounted with a little AA, wired, then the assembled firewall carefully pushed into place. A spare hole allows a long # 4 screw to be inserted to pull the assembly out, should that be needed. The EVA lens bezel and O-ring compete that light.

The negative lead from the Eva LED goes to the other light. The Iris (sans mount) lensed light needed 19 mm of depth with a thin ring of silver solder (helps thermal path to bezel, too) under it to move it forward so its back side would clear the ID of the light body without modding the lens or beveling the body more at that point. You could put a Lflex driver in the tail of this light, but not the H6 Flex..

I took a 31 mm OD copper pipe cap and flared the open end of it on a cone mandrel using the vise as a press. Slightly oversized, I ground the flare back for an interference fit to the 32.2 mm ID. This shortened the cap which helped the fit WRT the lens. I used a piece of copper from a 1 1/4 " pipe union soldered behind the flare and ground/filed/polished to an interference fit for most of the skirt of the pipe cap. The bottom of the cap deformed with a central dish when pressed. It was pushed back with a wood punch and a layer of silver solder was laid down to get a flat mounting surface for the star this also decreased the depth of the cap and helped lens placement. A 3/4" pipe cap was cut to fit the coned back of the body and bolted to the backside of the pipe cap off center towards the top of the light. AA was used sparingly in the tight interfaces but liberally where the back of the pipe cap curved away from the body (10-100 X better path than air). This light has no temperature monitoring so MUST run at least as cool as the one with the driver. Not having a driver saves about 0.7 watt of input, but the cap did not quite bridge the ID transition point (thin) of the light body so I added the second reversed smaller pipe cap with its skirt cut into four fingers to provide ample thermal path into the tail. The Iris has a central hole, so a glass lens cover was used. This holds the bezel out from its fully seated position, so I may see a thin piece of polycarbonate for this job.

The competed light mounted:


By brianmcb at 2011-03-25

Rode for I mile last night with them between the 1400 and level 3 (350 mA?) levels. Somewhere around 900 lumens out the front at 1400 is more than adequate. The other level was not after what I have gotten used to having, but way more light up close thatn I remember from my old HID. Will try 2 A, at some point, of course. Curious if they will trip the 50 C cutoff in the driver at 3 A while riding, too. That level would be nice off-road. Will have to practice my lean down and switch to low level while riding technique. At 1.4 A, Each light is below the 5 watt German standards for a 12 volt light, but together over 8 watts. They were intended to be aimed with their hotspots on the road and hoods to block the higher angle spill from oncoming drivers and pedestrians and provide a light path to where the helmet light takes over.

We will see.

So let the experimentation/night riding and recording it begin. First up a revisit of the ride by witha focused camera and appropriate camera and headlight angles.

shoot... I was distracted by the Mercian.
Nice Campagnolo bits on it.

Did you discuss the LEDs and optics? I saw mention of EVA leds or optics, but I'm not familiar with them.

thanks for the write-up!

regards,
Steve K.

road light experiments, not yet

Sure that the bent-metal b&M holder visible in the lower photo is stable enough for the load?

Re: road light experiments, not yet

Marcturus said:

Sure that the bent-metal b&M holder visible in the lower photo is stable enough for the load?

Click to expand...

Do you have expereince of one failing? If I was off road, MTBing, the jarring of the weight in these lights would likely fatigue it at the bolt hole or just above in spite of the recurved side reinforcement. I am running 32 mm tires on Deep Vees (they bow out a lot), the front tire at 60-65 pounds. (I'm circa 200.) Not a 25 mm tire at 110 pounds. Also, the bike has a classic curved 531 fork which absorbs a lot. It was cheap enough I am not married to it. Maybe I will repurpose a piece of Al crank arm. I have one lying around.

Something ate my reply to your post. So lets try again.

Steve K said:

shoot... I was distracted by the Mercian. Nice Campagnolo bits on it.

Click to expand...

You have good eyes. Only a bit of the rear QR is visible.

Steve K said:

Did you discuss the LEDs and optics? I saw mention of EVA leds or optics, but I'm not familiar with them. Thanks for the write-up! regards,
Steve K.

Click to expand...

XM-L's were well discussed including optics in a thread on MTBR. Larger die that XP-G, more output at fewer watts. The Iris worked well as a narrow 35 mm optic. The Eva was not tested there but Luminous DIY uses them in its Eva-lution line. It appears to be a Narrow medium. Aimed down with the high spill controlled, it looks like they will do.

Glad to have a reader or two.

Brian

Well done. You pushed the Marwi housing to the limit. Will you install the driver friendly hood for this build?

Microa said:

Well done. You pushed the Marwi housing to the limit. Will you install the driver friendly hood for this build?

Click to expand...

Thanks. Haven't pushed them above 1.4 A yet. The old NiMH pack at 40 F wouldn't feed the over 2 A to helmet too, and stay above the voltage cut off, so no point trying 3 A even with the helmet on the other pack. Is giving a bike a light, well any new bits..like giving a mouse a cookie?

Yes. I have a 1/4 mile piece of well lit road with a parking lot at one end where I can set the camera up in to ride at as if the camera was a driver. I will hone angles of dip, length and style, and roughness (orange peeling of reflective underside) until I get what I need to be civilized and visible and recover some of the cutoff light usefully. I then have Kathryn as a final arbiter. And she's tough! :kiss:

Here is the first video of the twin light in action, though it is day, lights not tipped much at all and no hoods yet. I did swap out the glass cover for thin polycarbonate and the bezel on that light is seated as intended, now. My take is that the low mounting height and the narrow beams of these lights make the under 100 foot very good even though the camera is placed as if the car were only 1 foot from me. Promising. Very Promising. The lights star in the second set, only though.

http://img12.imageshack.us/flvplayer.swf?f=Pf0j

The Dyno light is in the last two sets. Oh and the Marwi Tailights do an afterburner effect in the first of the first two sets. Yes, I now have a sixpack of 'em.

If interested in the side visibility analysis of the video above see this post.

Embedding video's does not work on this forum, But I was able to click the link, so no reason to complaint

BrianMc said:

If interested in the side visibility analysis of the vide above see this post.

Click to expand...

As you noticed and analyzed yourself, the side visibility of your lights using this camera is not as impressive as the head-on visibility of both tail and front lights. I Guess you need a better camera to capture it, or need to enlist someone to ride your bike while you observe and just concede that it is not easy to show the visibility here in the digital world

But did you notice I got the camera level? :naughty: Tricky, with the round bottom key fob camera. It only took a meter of masking tape, 5 short videos and viewings iteratively, until I got the tripod set right to counter their built-in lean. The tape released a bit in the cold so the other went slightly off level, but redundancy saved the day!

The digital world could address this, but a high speed low light digital camera is not exactly in my budget. Where's an IMax cameara when you really need one? So artifacts of the method with the tools available need to be addressed. Or I might be able to borrow a higher megapixal camera for the job. I could just turn all the blinkers on solid. That would take out any frame rate issues. I move the camera closer so weak lights will hit enough pixels. Assuming the vehicle was slowing or was at legal city speeds, about half the distance would be 1 second away. My guess from the frames is that the side firing PBSF and the helmet light might show better but the side output of the others won't. A sunny day today. I can just about guarantee non-success shooting into a southern sky full of sun. I can try for night video instead.

Another take home lesson is that the bright vest even in overcast is 1: LARGE, and 2: BRIGHT 3: CHEAP and 4: NO BATTERIES! Only the direct beams from bright headlights beat it and the brightest tailights seem close to equal in attention getting.

Another lesson that would have been obvious in a thought experiment, is that you don't have many flashes in a second or two to get a driver's attention and even at 15 mph or 24 kph you travel 44 feet (21 m) in two seconds. To me, that says the tendency here for motorists to try not to stop at intersections just doesn't give them enough time to see and register a fast moving bicycle and it explains a lot. Such as why a bright Xenon flashing front light in the day doesn't do much since it takes a second on fresh cells to charge the capacitor and fire and slows as the voltage drops.

A son or spouse would be the perfect person to talk into helping, but they don't fit my bikes at all. Which is mostly a good thing. However, I can just prop the bikes up at the end of my neighbors drive and move them as needed if the point is just to check the broadside light output. At this point, I have a very good idea of the angle of viewing as the cyclist approaches and recedes.

Had a nice ride in a heckuva wind. I find rides stimulate the mind. So when I saw that my neighbor so cooperatively placed his large trash receptacle to the curb today for tomorrow's pickup, I saw a bike and helmet stand I could view from my driveway. The helmet situated atop the lid would be the height of it on my head as I am riding by, the bike about a foot (30 cm) higher. The sun was about 90 minutes before sunset and some light clouds meant that it was close enough to the light levels of an overcast midafternoon of the video and not the brilliant sunshine of minutes before. The bike was about 6' (2 m) from the left-bound passes in the road so I adjusted my position relative to where the camera was accordingly. The Helmet light, the helmet PBSF, the side firing PBSF facing me, and the side light from the near side Radbot 1000 were all visible. The amber side light from the helmet headlight was brightest (visible but not pronounced), next the side light from the helmet PBSF, third was the side firing PBSF, and last was the side light from the Radbot 1000 ( a mere point of light, but there). Moving back I could get something like the video.

Then it hit me. What we have is like the old eye charts before the fancy optometric equipment. If you could not see the 'Big E' at 20' (6.1 m) they moved you to 10' (3 m) (personal experience). That means if our video camera is still a bit inadequate to stand in for the human eye, we just move it closer say 50' so it records what the human eye sees at the 75' (23 m). Problem solved. That will also allow conversion of my ride away and toward into human eye distances. If the camera need to be half as far to see as well then the 1/4 mile results of the old videos are really 1/2 mile for the human eye. Very neat conversion.

Since the side firing PBSF should be second, or even first, they are mounted too low even at the sightly higher position where they were observed this afternoon. Likewise, the Rabdbots need to be as high as I can put them.

I over extended myself a tad on the return into the gale, so the mods and the night shots will wait another night.

The Trike Mast thread had a mfr link quoting statistics on bike-car accidents in the US or maybe Canada and the US. About 45% were from cross traffic at night, and about 24% from left hooks, totalling 69.5% of all car-bike night accidents. So my concern is valid. The manufacturer of a side marker light contended that this was a side-marker light problem. In part, I agree, BUT a good amount of spill (not glare but easily visible from a headlight should deal with most of the left hook issue and cross traffic from the right. The left cross traffic will be out of the headlight's range if you are not going to blind oncoming drivers. They contended that passive relector systems allowed these. Not sure what percentage were riding ninja with all reflectors removed. It would not surprise me that it was a significant portion.

So my take is that a well designed headlight system will greatly reduce left hooks and right side pull outs without being antisocial to oncoming motorists. My steerable (head turn) helmet light allows me to reach the left side cross street motorists unless a large vehicle blocks the view and they bunny tail behind without looking. I have had enough starting to roll then stop when the helmet light was aimed at them to accept these anectdotes as evidence of a degree of effectiveness. You can't counter every threat, you can only do what you can do. Still, you can be distracted, so the fiber flare is an interesting device that might help that 45% cross traffic problem.

Need more data. More and better (closer) video.

More testing here. Not sure that the embeded videos play for non-members so here is a clickable link to the video:

http://img857.imageshack.us/flvplayer.swf?f=Pj43

It looks like the current hood is about as good as it gets. Replacing the helmet PBSF with a Radbot 1000, would help a little, and I can leave off the side firing PBSF's as they do nothing for real safety.

Interesting video.. my impression is that unless the light is aimed right at you, it really doesn't grab your attention. The flashing does help, as does the bit of wind causing the helmet to move a bit. I think it's just hard to compete with the ambient light, even on an overcast day. Movement is much better at getting someone's attention, as are bright colors. A nice bright safety green helmet cover might be a more effective solution for the daytime. This lets you use the ambient light instead of competing against it.

regards,
Steve K.

Yeah, the aiming part means the headlights and tail lights work for oncoming and overtaking traffic. The helmet turn works, too. The ANSI vest glows in sunlight and looks radioactive with a low sun shining on it. My next helmet won't be black, but I have 2 year left in this before I retire it for loss of resin resiliency. (Mfr's say 3 years, but apparently 5 is paranoid enough). Maybe I should paint this one to match the ANSI vest.

Here is a Video using the headlights at the top of this thread: http://img88.imageshack.us/flvplayer.swf?f=Mtym

The writeup is here. Enjoy.

Taming the double XM-L headlights for the road. The idea was to use narrow throwers with less aggravating spill and aim them below the horizon then cut some of the spill off with hoods that was headed toward oncoming drivers.

Untamed: Yeah, I tried the walk toward the stationary bike as if I was a driver with the lights in a level or slightly up angle so the hoods did little and the 'small' percentage of spill is too much. Not laser beam or HID level if the lenses delivered wider hot spots but like meeting a car with lows as it is cresting and you are on the crest of the same hill not yet headed down. As a brief flash of passing cars htat is OK, but steady as I pedal towards them? Not so good.

So I started with the same black plastic film canister-derived shade which gave the length, minus 10 mm about 7/16" to tape them to the body of the lights. Same length and shape I tested on the triple XP-G dyno light and which you can see in the Another Dyno thread. I did not bother with any reflective inner lining.

The hood shape and length is close enough to perfect for the Eva-lensed light by my eye even at 1.4 A. The Iris one is a real thrower with more intrense near the hot spot spill that the Eva. It was meant as a middle beam. With matching hoods, I found an angle that gave the stacked hot spots ending at 40-50 feet (about 15 m) ahead of the bike but it requires I switch to half power to be social and it is the Iris light that is the problem. With the H6flex in duomode, I can, and have reached down to switch to low power on a ride before the hoods, and got a good response from the driver I met. So that might be the modus operandi to be civil.

I hope to get a video camera with higher resolution and lower light capable that can document the lights and be a recorder for nice rides or bad drivers. In the mean time I want to zero in on my best options for taming the Iris lensed light.

If someone wanted a two light hi/low beam LED setup analogous to the less powerful halogen setup, as these lights were originally, then I'd use separate drivers with an LFlex in the high beam one and an on/off switch on the bars. Then you can aim it at the horizon or just below like a car high beam and use up to 3 amps if that is what you want to do. Or you could fashion a mechanical dipping mechanism. Dropping to 0.7 A does the trick with the hoods as they are (circa 500 lumens OTF, which isn't too shabby for a 'low' is it? Especially with another 500 in the helmet sharp cutoff and aimable light. That drops the twin headlamps below the German standard of 5 watts for a 12 volt light system (11.1 and 12.0 v packs, used) but with another 10 watts in the helmet light, the standard isn't so much broken as blown away.

So with one solution in hand, I will try both a deeper skirt on the hood and/or a longer hood and/or a fresnel lens to see if I can do better.

I think I have enough spill to be readily seen, but not enough to be excessive for all or almost all drivers.

I have three general road situations with oncoming vehicles:

1. Narrow suburban and urban residential streets
2. Wider 2 lane roads with no room for a bike and car in a lane at the same time.
3. Shoulders of 2 and 4 lane highways.

A fourth road situation is being built: a wider street plus a fog line and 2.5-3' (1 m) of road beyond that to the curb. That puts an oncoming vehicle's driver about 15' away (a bit under 5 m), and as far as my walk toward the bike lights test goes, it gets just enough spill that they can see me to not left-hook me. Two lane roads with shoulders place the drivers about another 1-2 feet away, more if they drift toward their fog line. The drivers are far enough to my left that the spill is not too intense at 1.4 A. 3 A might be a different story.

The oncoming lanes in the 4 lane roads with center grassed divider are so far to my left that I'd REALLY have to work at getting glare into their eyes. 3 A should be no problem, if I wanted that much light. Job done there.

It is the narrower roads which put me closer to the oncoming traffic that are an issue. In some cases their door mirrors are within 3' of me, the driver maybe 5' (1.5 m or so) from the lights. For those situations, I can power down a level and the amount of spill looks like 1/4 not 1/2. I assume that is because my irises are still set for the higher output. Again I don't find the spill a problem in a walk towards the lights and it worked on the road already. I can aim the helmet light and/or power back up if I think they haven't seen me and need a signal.

So I think I am ready to document and compare to vehicle lights. My cameras are not up to this task. So I am looking at a Drift HD170. It has a night mode that is NOT IR green plus you can adjust exposure from -2 stops to + 2 from there.

Here is a MTB video using the camera with flashlights? Riding in the tunnel effect. Looks like he needs more flood and more throw, but the camera shows what is lit and not, very well.

Another but of a motorcyle on streets, night mode + 1 exposure and you can see headlight beams on the street sometimes in spite of the glare of street lights. My dark unlit street will be a lot like the first video but to come close to human perception of them but on screen, I'm guessing I will need +2 stops of exposure (signal multiplication, in actuality, I suspect). It also has a remote tag mode so I can video traffic on an 'as needed' basis for 10 minutes blocks with enough resolution for plates, so the camera won't go unused after. I will need to not cuss at idiots when I yell at tehm to get over, though.

Thoughts or ideas on this camera or another that might suit?

Thoughts or ideas on comparing to vehicle headlights?

Thoughts or ideas on determining the spill intensity?

I have a light meter in lux which I can set on a tripod to get some comparable readings from bike, car, and light truck for some numbers to back the video and my observations.

It's good to get the lights ready for testing as the weather turns nice.

BrianMc said:

Thoughts or ideas on determining the spill intensity?

I have a light meter in lux which I can set on a tripod to get some comparable readings from bike, car, and light truck for some numbers to back the video and my observations.

Click to expand...

Is there not a light measuring setup for the StVZO in the "Lets design..." thread? I would suggest this test method (without applying the 'rules')

By the way: Nice camera, that Drift HD170, but a bit too pricey for me (~€300 over here)

The Drift 170 Stealth is a bit beyond what I was planning, too. But I recently received a gift that will cover it and the accessories (and maybe fund a support level here?). I have a bad feeling the file editing grind might force a new computer, though. I have asked a vendor if they have video showing its night use on streets not full of neon and overhead lights, maybe showing a car's headlight beams. As far as I can see other than being one unit and maybe a bit heavy on a helmet, it all the video camera I can imagine needing.

It would be interesting to know how car lights in NA come out on the StVZO standard, and whether it is fair or it not, how unfair it is in that regard. The DX meter will have to do, though. I will look. Good we can google the thread.

SvSVO allows up to 5 watts from a 12 volt system. So at 700 mA My current 'low' setting, the twin lights and driver consume about 4.35 W. More than twice that at 1.4 A, the currrent 'high' so 'beyond' the German standards. At 3 A. over 21 watts and over 300 % over the 4 watts of the standard. In use, the 1.4 A setting is brighter on the road with these lenses than my car's headlights on high beam.

From Bandgap post1:



Chart minus German explanation, translation below. Check original if German is better for you.

From Bandgap post2:

"All measured at 10m distance
Further than 3.4 deg above the horizontal, at any lateral angle, less than 2 lux
Horizontal straight ahead intensity (called intensity 'HV') - at least 10 lux
all the way between 4 deg either side of horizontal straight ahead - at least half of HV.
from horizontal straight ahead to 1.5 deg below - at least half of HV
From HV all the way down to 5 deg below - greater then 1.5 lux
all the rest of the box including 4 deg left and right, and from horizontal down to 5 deg below - at least 1 lux"


The max of 2 lux above 3.4 degrees at any angle is preventing stong spill well above the horizon.
The straight ahead minimum of 10 lux insures you have a light and not a firefly.
The beam must not drop below 50% of the straight ahead value for at least 4 degrees either side. This disallows an overly narrow beam.
Same from 1.5 degrees below straight ahead, making a beam of at least some depth.
Minimum 1.5 lux from straight ahead to 5 degrees below straight ahead = some near light.
Minimum 1 lux in the rest of the rectangle from 4 degrees left to 4 degrees right, some spill.

Other than the limits for light level above 3.5 degrees above the horizontal, there do not seem to be any other maxima limits. Though that is the very one I suspect the 1.4 A level has problems with. The question is does it? And if so, by how much?

So as I understand it, I convert angular measurements into vertical and horizontal distances at 10 m and create my target to match the figure above, preferably on a nice blank wall or I'll need to fabricate my own wall. I then place my lux meter at the various points and record what it reads. I will need to do this at various light outputs, and at maybe three angles: level and 5 and 10 degrees down angle.

The advantage is that the Edelux has been mapped out already and so we have easy comparisons and the other is I can see how much I have strayed beyond StVZO. The other aspect will then be to see how it rides as a cyclist or appears as a driver and there a decent video camera with decent night capabilities that even if less sensitive thatn a person can be compared from screen to reality to determine by how much the camera differs from what a person sees.

Comments?

Just one. You don't need a white wall to measure Light Intensity (Lux). The sensor will be pointed toward your light, not toward the wall. In fact, you don't even need a wall, except for the ease of positioning your sensor. A mesh wire fence would do nicely, so you can be behind the fence and not be in between your sensor and the light.

Another good reference point; Car headlights. Found some generic graphics, but not the measurement method.



Obviously (?), bicycle light do not have to produce a 30m wide beam Determining the 5 Lux line on the ground would give a good idea of the area that is illuminated effectively.

Meanwhile, in Another thread:

swhs said:

6.25 lux at 10 m (actually, StVZO does car headlamp measurements at 25m, so 1.0 lux measured at 25 m) compared to 2.0 lux for bicycle headlamps.

This makes it clear why one is getting blinded by car headlamps on parallel roads. See for more my StVZO page: http://www.xs4all.nl/~swhs/fiets/tests/verlichting/stvzo_en.html

Click to expand...

So use the StVZO way of measuring the bicycle headlight, and use the 'car' value of 6.25 Lux at 10m to judge the blinding effect of your lamp - either better or worse than a car headlamp.

Thanks. I was thinking 'wall' to lay out blue painters masking tape for sensor positioning and save me making something to place the sensor on. I can't get 10 meters from my garage door, which would have been perfect. Maybe I can borrow a neighbor's garage door and driveway where I can get that far away. I can't think of a suitable piece of fencing I can aim at with impunity. I can just make a 1" x 2" lumber frame for my 'wall' if no garage is suitable.

I read SWHS's analysis of StVZO (thanks BTW, if you should read this, SHWS). (Aside: very pleased that he confirms that the MJ-808 lights fail in road friendliness.) These halogen to LED conversions are verboten, it seems. Beyond StVZO by definition, if not intent. I knew that with lesser bike lights I was being blinded by car headlights. My night vision was a big part of being able to see the road with them. I also had a gut feeling that bicycle lights were not allowed as much above the horizon light as other vehicles in those standards judging by the German dynamo lights I was seeing road beamshots of (Peter White Cycles, eg). It is timely to have that 6.25 lux above 3.5 degrees at at 10 M figure for cars as an upper limit to work with! If I am under that, good enough and likely more visible which is what I am really after. Same sauce for the goose and gander.

I also agree we don't need to see deer at the side of the road as early as a faster moving car driver does, so 30 m width isn't needed. 10 m would be ample to pick up skunks at the side of the road (not pleasant to surprise one and be sprayed). Actually, a contour map of lux lines on a road suraface would be interesting and a means of optimizing the lights' aiming. I'd like to have an unlit large parking lot...hmmm. This is the work of a few nights, it looks like.