Headlights are still in the Dark Ages

[Bob Snow]

My 1958 Porsche has better headlights than my 2007 Accord. In Europe the 356 came with a separate lens, reflector and bulb but in the US, Porsche was forced to use sealed beam assemblies with separate clear outer aerodynamic lenses. This was a horrible compromise and the law for many years in the US. I replaced those 7" sealed beams with Hella replacement assemblies that used quartz H4 bulbs back in the 80's and this combination is subjectively better than the lights that came with my new Honda Accord. The Hella lamps have very good optics with a sharp beam cutoff. That said, I did replace the 9005 and 9006 bulbs in the Accord with HIR bulbs from Toshiba with excellent results. No change in the beam pattern, but more light. Not technically legal, but my car still produces less glare to oncoming traffic than all those trucks with higher mounted lights or any car with HID lights.

I don't like HID for a number of reasons.
#1 - Legislated lumen inequity! These lights simply throw more lumens down the road than any DOT legal incandescent and the inequity is a problem for all the drivers with weak lights. Glare is a bigger problem for drivers with poor illumination. If we all had HID lamps, the problem would go away. The haves and the have nots.
#2 - The color. Not just the temperature, but the color fringing. They do not have the best color rendering index.
#3 - Cost and Complexity. Ballasts and a price tag of around $1,000.
#4 - HID Retrofits. Lots of glare from these.

I like HIR bulbs because they produce a whiter light than regular quartz halogen with great color rendering. They are more efficient, requiring no more current than 9006/9005 bulbs to produce about 15% more lumens. Tinted bulbs are the worst. These reduce the light output for an arbitrarily different color. Yellow is fine for fog lamps, but the blue purple and green bulbs out there are an abomination.

I am hopeful that we are five to ten years away from a complete break from these technologies and the adoption of LED as a primary source for headlight illumination as well as all of the other ancillary lights. My Accord has some nice red LED brake and tail lights. Strangely, the rear turn signals are red incandescent, not amber like my previous Honda. I prefer amber turn signals because they are not confused for brake lights when there is a bulb out in the rear.

I firmly believe that the DOT needs to significantly limit the maximum height of headlamps to reduce glare to oncoming motorists.

My two cents.

Bob

 

[Diesel_Bomber]

Read some of the other posts in the automotive section; you'll find you're preaching to the choir here. Amen Reverend!

:buddies:

 

[-Virgil-]

It's good that you're honest and thoughtful enough to acknowledge that you're going by subjective impressions. In actual (objective) fact, Hella H4 7" round lights are poor performers, grossly inferior in almost every way to the stock headlamps with stock bulbs in your '07 Accord. Low beam peak intensity with a 9006 bulb in your '07 Accord is roughly triple that of the Hella H4 with a 60/55w bulb, and the peak intensity is much more optimally placed for effective distance seeing with the Accord lamp. Putting in a 9012 increases the advantage of the Accord headlamp considerably. The only thing the Hella 7" H4 does well when aimed correctly is keep glare out of other drivers' eyes.

HIR bulbs produce about 50% (9011 vs. 9005) and 90% (9012 vs. 9006) more lumens, not 15%. Not because of what somebody's sales lit says, but because that's what they're required to produce by American and international technical regulations.

If everyone in America had HIDs (as currently implemented in America) the glare problem would definitely not go away, it would get much worse. The glare problems are not due to a novelty effect, which is the implication of what you're saying. The difference generally isn't as large as you think between down-the-road illumination between DOT halogen and DOT HID headlamps. The main difference is in width of significant illumination, and that combined with the lack of a requirement for headlamp inspection and autolevelling headlamps in America creates most of the photometric glare problem. Other factors (SPD/light color) also contribute to some degree.

Color fringe/beam color are relatively minor issues, but they're issues nevertheless and they're being addressed at a moderate speed by regulators in Europe, not at all by American regulators, as usual.

You're correct that HID kits are a serious safety problem. Headlamp mounting height could be reduced somewhat, but autolevelling, height-dependent aim (and stringent periodic headlamp aim inspection and correction) would probably be a faster, more effective way to tackle the glare caused by height-related factors.

LED headlamps are about 18 years away from truly widespread commercialization. When that time comes, they will push HIDs out of new high-end and high-efficiency vehicle designs. Halogens will retain the largest market share worldwide. LEDs are definitely not a panacea to solve all headlight problems! They bring their own issues and problems and technologically-inherent strengths and drawbacks.

The red rear turn signal is indeed a "dark ages" item only allowed in America; pretty much everywhere else in the world they have to be amber. The newest research shows cars in America with amber rear signals are about 28% less likely to be hit in collisions relevant to turn signal usage, versus comparable vehicles with red rear signals (brief PDF overview or full research report). If your Accord's a sedan, you can swap to the rear lamp lenses for the hybrid model, which have a clear-lens signal section for use with an amber bulb. It's not even very costly to do so, because the LED matrix for the brake/tail light stays put and you replace just the lens assembly; you can generally find them used.

 

[Diesel_Bomber]

What's turn signal color matter when no one USES them anyway?

(Yes yes, tongue in cheek. Mostly.)

:buddies:

 

[irsa76]

I've always maintained it's not how much (light) you've got, it's how you use it!
The bitter irony is of the 3 cars I've got, the 2 Australian made, Japanese based cars both suffer from the JDOT standard lights, think DOT but slightly biased to LH traffic. Even though Australian designed lights are considered to be the best around, the beater has JDOT because they are replacement lights while my wagon has them because it was the top of the line model and they "look" better. My Chrysler minivan uses genuine Ecode, Austrian built vehicle, and even with standard bulbs, the other 2 cars use premium 55w H3 bulbs, it's much better. Although the beater would be better if I actually aimed the damm things!

 

[Bob Snow]

I aimed the Hella lights myself optically, not mechanically, using a diagram that came with them back in the 80's. They have a sharp cutoff and clear point at which the beam kicks up to the right. The car was on level ground and a specific distance from the wall and I just aimed them according to the chart. They do not shine into oncoming traffic unless I am cresting a hill and in normal driving they seem to illuminate the road ahead farther than the Accord, but maybe it is the great view they give along the side of the road that makes them seem subjectively better. I had Cibie Z beams in my VW Rabbit back then and they did not throw light up so high and to the right.

I wonder if the Accord lights are aimed too low? No way to check without mechanical aiming equipment. They do not have such a sharp cutoff and I think this has to do with the difference between European and DOT regulations. The HIR bulbs helped and the Honda is subjectively better than the Porsche now. I did not realize HIR were a full 50% brighter. By reflecting back the IR output which would otherwise be lost exiting the bulb, the filament gets hotter with no additional current draw. I wish these were available in more than just PAR reflectors for general home use.

Both the H4 and 9006 low beams are 55 watt. Is there that much of an advantage to having separate bulbs and reflectors for high and low? Is there a significant advantage having the reflector faceted and the outer lens clear? The Hella has a parabolic type reflector with a fluted lens and is an obvious compromise with a dual filament bulb.

My thought about HID lamps and glare was simply this: The glare from other drivers HID lamps would be less of a handicap if I was driving a car with HID and had better illumination down the road. I was unaware of the difference in side illumination allowed with HID headlamps. I can see how that would contribute to glare. Are they more like fog lights that throw light farther to the side? Plenty of people drive with their fog lamps on in clear weather and that certainly seems to cause glare. Just more bad driving habits like failing to signal or driving in rain with no lights at all. And all those blue tinted headlights!

I still think that headlights mounted high in a truck present more glare to oncoming traffic at close range even when properly aimed. Specifically, as a car with high mounted headlights crests a hill and you are passing, the glare can be worse. Obviously as the distance increases, the problem is mitigated.

Bob

 

[-Virgil-]

I aimed the Hella lights myself optically, not mechanically, using a diagram that came with them back in the 80's.

They come with a leaflet that describes this procedure.

in normal driving they seem to illuminate the road ahead farther than the Accord

No way...they're outclassed even by better-engineered 7" H4s. Stern's got interesting beam comparos here. Presented in descending-performance order. I'd be happy with the first two. I'd accept the third. The fourth (Hella)? Really pretty narrow and weak, no thanks.

I wonder if the Accord lights are aimed too low? No way to check without mechanical aiming equipment.

That's not correct. Your Accord's headlights have no provisions for mechanical aim. Almost all American-market headlights have been visual/optical aim, not mechanical, since about 1999. There are multiple different visual/optical aim procedures for different American lights (see the aim link above, it's not like the Europe/elsewhere situation where all lights are aimed using the same procedure) but your Accord lights are the "VOL" variety, which means they get aimed according to the same procedure as the Hella lights. Go scrutinize the Accord lenses and you'll find the "VOL" marking.

By reflecting back the IR output which would otherwise be lost exiting the bulb, the filament gets hotter with no additional current draw. I wish these were available in more than just PAR reflectors for general home use.

They were available in linear tubular form for torchieres and such, but GE discontinued 'em. Can still get 'em as MR-16s that work well. Don't fret, there are a bunch of really interesting high-efficacy incandescent technologies in the pipeline.

Both the H4 and 9006 low beams are 55 watt.

Yeah, but remember, we don't use the watt as a measure of light. We use it as a measure of power input. There are different auto halogen bulbs with filaments nominally rated 55w at 12v, producing from 890 to 1820 lumens. That's almost a 2:1 ratio. There's also a huge variety of design lifespan, design envelope wall temperature, and other parameters.

Is there that much of an advantage to having separate bulbs and reflectors for high and low?

That totally depends. If you have a sufficiently large reflector, then H4 can work well, because in effect you do have separate optics for low and for high beam (the lower 45% of the reflector/lens is used only for high beam, so can be optimized for high beam). The flip side of that coin is that only 55% of the reflector/lens is used for low beam. Okay, so we can use a double-filament bulb with an unshielded low beam filament (9004, 9007, H13), now we're using the whole optic for low beam and for high beam, but we cannot optimize for either beam, we have to compromise for both. So unless you have a giant optic to work with (which you usually don't), then yes, all in all, there is a large potential performance advantage to having separate optics and bulbs for low and for high beam. "Potential" because it's possible to make good combination high/low beam headlights, and it's possible to make lousy separate high and low beam headlights.

Is there a significant advantage having the reflector faceted and the outer lens clear?

This is another potential advantage. Reflector optics can be made to give greater efficiency than parabolic reflectors with lens optics, but it's equally possible to create bad/inefficient reflector-optic headlights.

My thought about HID lamps and glare was simply this: The glare from other drivers HID lamps would be less of a handicap if I was driving a car with HID and had better illumination down the road.

Somewhat, but not totally. This kind of endless ratcheting upward of glare levels is what the American DOT has advocated for years. "Oh, people will get used to the glare, and once everyone has bright headlights nobody will care." That philosophy is without basis.

I was unaware of the difference in side illumination allowed with HID headlamps.

It's not so much a matter of what's allowed, it's what can be attained because there's so much more light for the optic engineer to work with.

Are they more like fog lights that throw light farther to the side?

Mmm...not really, no. Fog lamp glare is another whole issue with many causative factors.

Plenty of people drive with their fog lamps on in clear weather and that certainly seems to cause glare.

Yup, that's one big factor. Also poor-quality fog lamps, insufficient (or zero) aimability of fog lamps, and others. And then we could talk about improper use of rear fog lights! "My fabulous expensive Jaguar came with this little extra switch on the dash. I don't know what it does, but I obviously paid for it, and I want the whole world to know I paid for it, so I'm gonna leave it on!" These putzes drive around with their front and rear fog light switches in the "on" position all the time. What a joke.

Just more bad driving habits like failing to signal or driving in rain with no lights at all.

Yeah..."Oh, I don't have to pay attention to my car, I can just keep talking on my cell phone, because my car has automatic daytime running lights so I'm good." Damn clueless idiots shouldn't be able to have a license.

I still think that headlights mounted high in a truck present more glare to oncoming traffic at close range even when properly aimed.

I think you might be missing the point at least partially. According to the American DOT, "properly aimed" is the same angle for the headlights 16 inches off the ground on a Corvette as for the headlights 5 feet off the ground on a lifted Jeep.

 

[irsa76]

That totally depends. If you have a sufficiently large reflector, then H4 can work well, because in effect you do have separate optics for low and for high beam (the lower 45% of the reflector/lens is used only for high beam, so can be optimized for high beam). The flip side of that coin is that only 55% of the reflector/lens is used for low beam. Okay, so we can use a double-filament bulb with an unshielded low beam filament (9004, 9007, H13), now we're using the whole optic for low beam and for high beam, but we cannot optimize for either beam, we have to compromise for both. So unless you have a giant optic to work with (which you usually don't), then yes, all in all, there is a large potential performance advantage to having separate optics and bulbs for low and for high beam. "Potential" because it's possible to make good combination high/low beam headlights, and it's possible to make lousy separate high and low beam headlights.
This is another potential advantage. Reflector optics can be made to give greater efficiency than parabolic reflectors with lens optics, but it's equally possible to create bad/inefficient reflector-optic headlights.

If you have alook at the 2 photos in my last post, all 3 cars have seperate high and low beam lights. The nissan uses traditional parabolic reflector with lens optic, while the Mitubishi and Chrysler both use reflector optic lights, although Mitsubishi calls them multiparabolic lights. Of the 3, the Chrysler's lights are miles, literally!, in front.

 

[-Virgil-]

Yeah, the Austrian-built Chryslers got really good lights made by ZKW (also of Austria). H7/H7 on the minivans, with well-engineered optics. Somehow ZKW seems to make good use of a particular plastics material for their reflectors that other companies don't have much success with. You should see (no you shouldn't!) the pathetic junk the American-market version of your same minivan got. Low-bid, no-name trash with a single #9007 bulb and sloppy optics.

The old Japanese headlamp specs are now progressively being replaced in Japan by European ECE specs. Wouldn't mind seeing some clear, sharp, close-up photos of the optics in your "J-code" headlights on your Mitsu and Nissan, and maybe some beam pattern shots (projected on a wall from 5 or 6 meters away). The two "J-code" cars use H3 in the low beams, too, or just the high beams?

 

[2xTrinity]

It's good that you're honest and thoughtful enough to acknowledge that you're going by subjective impressions. In actual (objective) fact, Hella H4 7" round lights are poor performers, grossly inferior in almost every way to the stock headlamps with stock bulbs in your '07 Accord. Low beam peak intensity with a 9006 bulb in your '07 Accord is roughly triple that of the Hella H4 with a 60/55w bulb, and the peak intensity is much more optimally placed for effective distance seeing with the Accord lamp. Putting in a 9012 increases the advantage of the Accord headlamp considerably. The only thing the Hella 7" H4 does well when aimed correctly is keep glare out of other drivers' eyes.

HIR bulbs produce about 50% (9011 vs. 9005) and 90% (9012 vs. 9006) more lumens, not 15%. Not because of what somebody's sales lit says, but because that's what they're required to produce by American and international technical regulations.

If everyone in America had HIDs (as currently implemented in America) the glare problem would definitely not go away, it would get much worse. The glare problems are not due to a novelty effect, which is the implication of what you're saying. The difference generally isn't as large as you think between down-the-road illumination between DOT halogen and DOT HID headlamps. The main difference is in width of significant illumination, and that combined with the lack of a requirement for headlamp inspection and autolevelling headlamps in America creates most of the photometric glare problem. Other factors (SPD/light color) also contribute to some degree.

The issue with HIDs is pretty ridiculous. It amount to the fact that the DOT has mandated a maximum beam intensity from headlights (in candela). Rather than reduce the lumen ouptut of teh HIDs (which is about 3x higher than halogens) to comply, they simply project the light into a much wider angle so the max intensity into any particular angle is the same, but the total angle the light is projected into is much wider so as to "get rid" of the extra lumens legally...

The red rear turn signal is indeed a "dark ages" item only allowed in America; pretty much everywhere else in the world they have to be amber. The newest research shows cars in America with amber rear signals are about 28% less likely to be hit in collisions relevant to turn signal usage, versus comparable vehicles with red rear signals (brief PDF overview or full research report). If your Accord's a sedan, you can swap to the rear lamp lenses for the hybrid model, which have a clear-lens signal section for use with an amber bulb. It's not even very costly to do so, because the LED matrix for the brake/tail light stays put and you replace just the lens assembly; you can generally find them used.

I wonder if this study distinguishes betewen cars that flash one of the regular taillights to signal, or that have separate, dedicated turn signal lights but that happen to be red. Obviously all amber lights would have to fall into the category of "dedicated turn signal light". So it could be that the cutting corners and using a single red light for multiple purposes is a bigger problem than the color red. Even taking that into account however, I still woudlnt' be surprised if there's a significant advantage to the amber.

IMO a bigger problem is that not enough people with red or amber turn signal lights actually use them... Here in southern California I'd say it's about 60/40 ratio between people who usually signal and those who never signal on the roads. A significant fraction of those who don't signal also don't look before changing lanes, either. Just about every day I drive on the freeway I see some guy completely shift lanes without noticing the guy in his blind spot first THEN look and realize "oh crap I just forced that guy onto the shoulder!" and quickly swerve back...

It's not just lane changes either. Lots of people don't signal when they turn off of roads into parking lots, or turning at intersections, either..

 

[jtr1962]

LED headlamps are about 18 years away from truly widespread commercialization. When that time comes, they will push HIDs out of new high-end and high-efficiency vehicle designs. Halogens will retain the largest market share worldwide. LEDs are definitely not a panacea to solve all headlight problems! They bring their own issues and problems and technologically-inherent strengths and drawbacks.

18 years? I'd be surprised if LEDs don't dominate the market by 2015. They've already pretty much taken over illumination applications of 200 lumens or less. The beams can be shaped much more easily via TIR optics to whatever shape is mandated by the DOT. They never need replacing, hence the beam pattern can't be inadvertently screwed up by mounting a replacement lamp a few degrees off center. Cost? That's dropping and should be competitive within a few years. Design freedom is another advantage. With LEDs you lose the need for bulky headlamp assemblies, and can distribute the light across the entire front of the vehicle instead of concentrating it in two sources. This means less glare even when putting the same amount of light on the road. Seriously, I don't see halogen remaining at all in perhaps 10 years, much less retaining the largest market share. Just the fact that they need to be periodically replaced is a huge drawback and a major safety concern.

Of course, all this goes on the assumption you'll still need projector lamps in 10 or 20 years time. With all the talk of autos driving themselves lighting might simply need to be adequate enough to allow pedestrians to see the vehicle, with no need at all to light the road ahead.

This kind of endless ratcheting upward of glare levels is what the American DOT has advocated for years. "Oh, people will get used to the glare, and once everyone has bright headlights nobody will care." That philosophy is without basis.

Glare is far less of an issue if roads are well lit with street lamps as your eye isn't as dark adjusted. Really, more localities should install adequate street lighting. Actually, they all should. The reduction in accidents would more than pay for it. Even with great headlights it's far more dangerous driving on unlit roads.

 

[2xTrinity]

Glare is far less of an issue if roads are well lit with street lamps as your eye isn't as dark adjusted. Really, more localities should install adequate street lighting. Actually, they all should. The reduction in accidents would more than pay for it. Even with great headlights it's far more dangerous driving on unlit roads.

Glare isn't a huge problem in cities at all, in my expereince. It's a problem on two-lane mountain roads where NO street lights are present (or really even feasible at all). HIDs are horrible in those cases for the following reasons:

• They throw light to the sides. This means the lane of oncoming traffic in a 2 lane road
• They have abrupt cutoffs. This means a sudden transition from darkness to extremely bright light in your eyes as vehicles crest hills etc. Even halogen projectors are bad in this respect.
• Jackasses (usually in raised trucks) like to run with their high beams on continuously, because there are no street light...

That said, I agree with you on street lighting. I'd like to see a shift away from high- and low-pressure sodium, to a sensibly-aimed white light source. IMO a better way to deal with light pollution than forcing low-pressure sodium to be used, woudl b to use a white-light source that is sensibly aimed (ie, shine light at the ROAD, not on random ditches or up in the air), and that shuts itself off when no drivers are around. This isn't feasible with HID streetlights that require minutes to start up, but with LEDs, they might be triggered using the vehicle sensors in the road, or even motion detectors.

 

[jtr1962]

That said, I agree with you on street lighting. I'd like to see a shift away from high- and low-pressure sodium, to a sensibly-aimed white light source. IMO a better way to deal with light pollution than forcing low-pressure sodium to be used, woudl b to use a white-light source that is sensibly aimed (ie, shine light at the ROAD, not on random ditches or up in there), and that shuts itself off when no drivers are around. This isn't feasible with HID streetlights that require minutes to start up, but with LEDs, they might be triggered using the vehicle sensors in the road, or even motion detectors.

Great idea, and this would make street lighting feasible in places where it currently isn't (i.e. two lane mountain roads with no cheap way to connect to the grid). You can have a streetlight with a solar panel which charges by day. Since the light will only be on for a fraction of the time at night, the energy collected would be more than adequate, especially with LED efficiency continuing to rise.

And yes, I'm all for LED streetlights which can be made to actually aim their light on the road where it's wanted, not in the sky or a ditch.

 

[irsa76]

Yeah, the Austrian-built Chryslers got really good lights made by ZKW (also of Austria). H7/H7 on the minivans, with well-engineered optics. Somehow ZKW seems to make good use of a particular plastics material for their reflectors that other companies don't have much success with. You should see (no you shouldn't!) the pathetic junk the American-market version of your same minivan got. Low-bid, no-name trash with a single #9007 bulb and sloppy optics.

The old Japanese headlamp specs are now progressively being replaced in Japan by European ECE specs. Wouldn't mind seeing some clear, sharp, close-up photos of the optics in your "J-code" headlights on your Mitsu and Nissan, and maybe some beam pattern shots (projected on a wall from 5 or 6 meters away). The two "J-code" cars use H3 in the low beams, too, or just the high beams?

I found out this morning the Nissan uses H4 and H3, I think. The van and the Mitsu use H7 for high and low beam, the Mitsu uses H3 for the fogs. I should have remember the H7 since spent ages getting some good bulbs for it!
Will try and get some reflector and beam shots, all 3 cars have lens in pretty average condition though.
As for the US market Chryco minivan lights, as it happened I saw not one but 3 Dodge Grand Caravans the same shape as mine today, US embassy staff. One was parked so I had a quick look at the reflector, "looks" ok to the eye. But the other one was driving and had the lights on, I do the same thing, and my god did they glare. The fact the reflector is so big seems to make it worse.

 

[Bob Snow]

OK. I pulled my Accord up to the garage door and made some tape marks with a level and rolled back 25 feet for a quick check. Two things. The beams converge horizontally. If I had another 75 feet, they would cross! The lights are aimed much too low when compared to the VOL chart. My driveway is relatively flat but I need to find a dead flat location to adjust them accurately. This car was purchased new, so no front end damage to account for this, just badly aimed from the factory. The effect of the poor aiming seems to throw all of the light onto the ground by about a hundred feet.

Thanks for the heads up on visual aiming for newer cars. I will get this done soon.

As far as the tail lights are concerned, mine is a coupe, so no interchange with the hybrid Accord. Hadn't thought about it, but I think that the combined tail and turn signal bulbs are in fact the worst. At least my red turn signal is a separate bulb.

 

[270winchester]

out of curiosity I understand that Ford has a big operation in Australia. Have you had much encounter with Fords and if so, any difference between AUstralia market Fords' headlamps and say NA market Fords?

 

[-Virgil-]

The issue with HIDs is pretty ridiculous. It amount to the fact that the DOT has mandated a maximum beam intensity from headlights (in candela). Rather than reduce the lumen ouptut of teh HIDs (which is about 3x higher than halogens) to comply, they simply project the light into a much wider angle so the max intensity into any particular angle is the same, but the total angle the light is projected into is much wider so as to "get rid" of the extra lumens legally...

Much of that "extra" light usefully and significantly improves the safety performance of the headlamps. There is no single DOT maximum beam intensity. Rather, as in all other beam pattern standards, there is a matrix of "test points" at various locations throughout the beam. Some of them have a specified minimum intensity, some have a specified maximum intensity, and some have both a minimum and a maximum. In the American beam standard, the test points are all pretty much alike regardless of light source - tungsten, tungsten-halogen, HID, or LED. There are differences according to whether the headlamp produces beams intended for old-fashioned mechanical-aim unit or for visual/optical aim, and there are a couple of differences in the high beam peak intensity spec for certain dual-filament tungsten-halogen light sources (9004/HB1 is allowed to emit much less light on high beam, and 9003/HB2 is required to emit much more light on high beam). But there are no photometric requirements specific to HID headlamps. The ECE regulations used in Europe (and almost everywhere else), on the other hand, have completely separate photometric regulations for HID headlamps vs. tungsten or tungsten-halogen headlamps vs. LED headlamps. The differences take account of the performance characteristics peculiar to each kind of light source. It's a philosophical difference; the Americans say "HID glare complaints are baseless; all U.S. headlamps have to meet the same glare limits…why do the Europeans insist on messing around with multiple different headlight regulations?", while the Europeans say "It's just not reasonable or realistic to lump all headlamp light sources together and pretend there are no differences".

I wonder if this study distinguishes betewen cars that flash one of the regular taillights to signal, or that have separate, dedicated turn signal lights but that happen to be red.

You'll know the answer is "yes" as soon as you…um…like…read the study! ;-)

it could be that the cutting corners and using a single red light for multiple purposes is a bigger problem than the color red.

Maybe. But there are issues with separate red brake/red turn lights, too. If the (red) rear turn signal is too close to the (red) brake light, it's almost impossible to see the turn signal until you're too close to the car for it to make any difference. ECE regulations don't allow red rear turn signals, but do require a mandatory separation of at least 100mm between closest adjacent lit edges of the (red) brake lamp and the (red) rear fog lamp. This provision does a reasonable job of preventing the rear fog from making the brake lamp invisible. An identical provision for separate red brake and red rear turn signals in American regulations would likely help avoid the invisibility of red turn signals too close to the brake lights (many present VWs, '96-up Ford Taurus/ Mercury Sable wagons, the previous Nissan Sentra, a fair number of Audis, lots of Hondas and Toyotas, '96-'04ish Chrysler minivans, etc.). The VWs are some of the worst, right now. You have a circle of about 5" overall diameter that's all dim red for the taillamp. That's fine, but the inner 2-1/2" spot lights up bright red for the brake function, and the outer ring lights up bright red for the turn signal. Seeing the turn signal is all but impossible until the VW's tailpipe is embedded in your car's grille.

The only models they don't do this on are the Tiguan (amber central spot, red outer ring, but well enough balanced that there's no conspicuity or discernibility problem), the Tuareg and the Jetta wagon (conventional non-circular arrangement with big red brake/tail and
smaller amber turn underneath that).

Oh, they also don't do that inner/outer circular thing on the current Jetta. They do a different but equally pathetic thing: Circular red brake lamp with a left bulb and a right bulb. Step on the brake and signal for a turn, and the outer half of the applicable brake light flashes while the inner half remains steady. Again, very difficult to discern the turn signal til you're on top of tthe car.

Another problem is that the standards (both U.S. and ECE) stipulate minimum and maximum allowable intensities, but make no reference to adjacent lamps. So a minimum-allowable-intensity turn signal can be right next to (or in the middle of) a maximum-allowable-intensity brake lamp, and the setup is still legal. Shouldn't be, but is. Back to complaining about VWs: Outside North America, the Passat has an opposite and equal(?)
problem to the red/red circle/spot issue described above: the whole 5" circle still serves as dim red tail, and the inner 2-1/2" spot still serves as bright red brake, but the outer ring
lights up bright(!) amber for a nice, conspicuous turn signal...which makes the inner red spot brake light impossible to see until you're on top of the car. So there are lots of ways to make a car's brake lights and turn signals legal, but only some of those implementations make the brake lights and turn signals good.

IMO a bigger problem is that not enough people with red or amber turn signal lights actually use them.

That is indeed a big problem. Some of it is due to just plain laziness, and some of it's selfishness ("I can't be bothered using my turn signals, I'm on the phone!"), and some of it is idiotic rationalization ("If I use my turn signals, nobody will let me turn or change lanes!"). But I have to wonder if some of it might be due to long-running, deep conditioning of turn signal uselessness and confusion due to the mishmash of combined/separate/red/amber turn signals in North America.

Since we've veered a little off topic, here's a (completely obvious, leading) question: should significant, uniform, wide-angle side-on visibility of a car's turn signals be optional (as in America) or should it be mandatory as in much of the rest of the world (side turn signal repeaters on the fenders or sideview mirrors)?

out of curiosity I understand that Ford has a big operation in Australia. Have you had much encounter with Fords and if so, any difference between AUstralia market Fords' headlamps and say NA market Fords?

Totally, completely different. American headlamps are not legal in Australia, because they are for the wrong (i.e., not the left) side of the road.

OK. I pulled my Accord up to the garage door and made some tape marks with a level and rolled back 25 feet for a quick check. Two things. The beams converge horizontally. If I had another 75 feet, they would cross!

Are you absolutely sure? It's possible, but less likely than it used to be. They're not horizontally adjustable…this is a requirement of U.S. law: visual/optical aim is allowed only in the vertical direction. You can either provide for mechanical horizontal aim checking, or provide an onboard calibrated-scale horizontal aim indicator, or you can make the headlamps horizontally fixed. The given rationale is that horizontal visual/optical aim isn't possible (ahem...Europe has successfully been doing it since the 1950s) and that horizontal aim isn't important or necessary because today's headlight beams are wider than they used to be (I guess this means today's cars don't get knocked around in service). If you look carefully, you may find capped or otherwise blocked/disabled horizontal aim adjustment provisions on your headlamps; outside America horizontal aimability is still required, and those headlamp assemblies were used with different optics in countries that apply ECE regulations.

The lights are aimed much too low when compared to the VOL chart. My driveway is relatively flat but I need to find a dead flat location to adjust them accurately.

Naw, shining them on a wall from (what you think is) 25 feet away on (what you hope is) level ground is nowhere near as precise or accurate as the correct use of an optical beam checker. It can be difficult to find these in America (most garages still guess at it if they even bother checking at all) but it's getting easier. Call around, be persistent. The device looks like a big TV camera that gets wheeled in front of each headlamp, one at a time.

As far as the tail lights are concerned, mine is a coupe, so no interchange with the hybrid Accord. Hadn't thought about it, but I think that the combined tail and turn signal bulbs are in fact the worst. At least my red turn signal is a separate bulb.

Not necessarily better. See above. They sold the (American) Accord sedan in other countries around the world, not sure about the coupe.

18 years? I'd be surprised if LEDs don't dominate the market by 2015.

Whoops, I mistyped — not 18 years, but the year 2018 is the automotive lighting industry consensus, give or take a year or so, as to when we'll start to see mainstream volume presence of LED headlamps. And they will not dominate the market for a long, long time well after that date.

They never need replacing

If I could do a Rocky the Flying Squirrel voice, I would, but I can't so you'll have to do it in your head: "That trick never works!" Automotive service is extremely tough, and there have been many technologies in the lighting field alone touted as "life of car" items, and it has never worked out that way. Remember, HID bulbs and ballasts were going to be "life of car" items, and yet somehow there's a brisk market in replacements for failed bulbs and ballasts! Fact is, LED emitters do and will degrade with use and age, and sometimes they fail well before their statistical rated lifespan. The Europeans are grappling with the issue of how to standardize and regulate LED light sources so that the owner of a 6-year-old car does not find himself faced with the choice of buying a new $5,000 headlamp assembly or parking the car forever, but without retarding LED evolution. The Americans are pretending there's no such problem.

Cost? That's dropping and should be competitive within a few years.

No, cost is not dropping. We're still in the area of the curve where cost remains the same (and very high) but lumen availability increases. Eventually, once emitters are available with luminous efficacy levels sufficient to make more than just boutique headlamps for special low-production models, cost will begin to drop…slowly. LED headlamps will not be cost-competitive with any other headlamps within a few years, or even a couple-few years. The 2018 timeframe is when they'll begin to be cost-competitive with HIDs. In the long-range future, this will accelerate because HIDs will increase in cost due to volume reduction as LED headlamps replace HID headlamps in new high-end and eco-focused car designs.

It's useful to keep in mind the predictions for HID headlamps (~50% market share by 1996, near-total market domination by 2002…here it is 2009, and we have not seen anything even close to that level of HID market share.)

Design freedom is another advantage. With LEDs you lose the need for bulky headlamp assemblies

Not so. You still need very bulky headlamp assemblies — you don't need as much room for optical packaging as with present point-source headlamps, but you need a lot more room for thermal and power management.

can distribute the light across the entire front of the vehicle instead of concentrating it in two sources

This is still deep in the realm of concept-car fantasies. Not only is such a solution vastly more costly than even the hideously expensive existing LED headlamp assemblies (Audi R8, Lexus LS600h, Cadillac Escalade), but it's also forbidden by regulations worldwide. It will be quite awhile before we see solutions like this on actual, real production cars.

Seriously, I don't see halogen remaining at all in perhaps 10 years, much less retaining the largest market share

Halogen will still have at least a 70% market share worldwide in twenty years, solely because of cost and complexity. The halogen market share will of course be lower in Europe, Japan, and to a lesser extent North America, but it will not resemble anything like LED headlamp market domination. Your concept of LED headlamp cost (static and over time) appears to be several orders of magnitude too optimistic. You might find a great deal of interesting reading over here, though a subscription is required to access most of the content.

 

[MichaelW]

NHTSA should change their slogan to 'People wasting other people's money [on worthless makework studies]'
Do these people have a basic understanding of human anatomy/physiology?

No sh*! that red turn signals suck, even positive signalling red turn signals still stink (marginally less)

Page 2 is such a cop out. Here is the action to be taken now!
Set Jan 1 2010 as the deadline; after set date no new vehicles may be introduced [to be sold] with red turn signals.

Table 12 headline-funny, now if we could separate tail lights from brake lights, then we would have something.
and quoting a 1974 study where in non driving situations the red turn signal is more conspicuous. (No, nothing has improved since then

How the f^%k is the new Ford Mustang allowed to have an caricature of a turn signal, instead of a real turn signal.

 

[jtr1962]

Whoops, I mistyped — not 18 years, but the year 2018 is the automotive lighting industry consensus, give or take a year or so, as to when we'll start to see mainstream volume presence of LED headlamps. And they will not dominate the market for a long, long time well after that date.

You do know that future predictions are fraught with peril? Remember the famous "640K ought to be enough for anybody"? And I remember well when LCD screens first started appearing in the early 1990s that many "in the know" were saying even best case it'll be 25 years before the venerable CRT is no longer the majority of the market. Well, here we are about 15 years later and the CRT is for all intents and purposes dead. Anyway, if indeed the auto industry consensus is that LED won't have a mainstream volume presence for another 9 years, and won't dominate even in 20, it's not because LEDs aren't up to the task. Rather, it's because this is what the auto industry wants. I see a parallel here with HID which was predicted to dominate the market by 2002 but didn't. Here's why in my opinion-it's more profitable for the automakers to keep HID and LED as expensive luxury options than to aim for making them standard. By keeping HID as an option, the automakers in turn kept the price higher as less volume was made. This was in keeping with their desire to make it an expensive, luxury item for which they could justify a hefty markup. An HID system which might have cost $300 if made in volumes sufficient to make standard equipment on all cars ended up costing $1500 instead. And this $1500 cost might have ended up being a $3000 or $4000 markup in the sticker price. However, if HID was standard equipment the automakers couldn't charge extra for it. Granted, they may have raised prices enough to recoup their extra costs above and beyond a halogen setup, but that's about it. And now sadly it looks like they want to do the same thing with LED, which incidentally in volume production could end up costing not much more than halogen in ten years time (although it's admittedly very expensive now).

I'll deal with your other arguments next:

If I could do a Rocky the Flying Squirrel voice, I would, but I can't so you'll have to do it in your head: "That trick never works!" Automotive service is extremely tough, and there have been many technologies in the lighting field alone touted as "life of car" items, and it has never worked out that way. Remember, HID bulbs and ballasts were going to be "life of car" items, and yet somehow there's a brisk market in replacements for failed bulbs and ballasts! Fact is, LED emitters do and will degrade with use and age, and sometimes they fail well before their statistical rated lifespan. The Europeans are grappling with the issue of how to standardize and regulate LED light sources so that the owner of a 6-year-old car does not find himself faced with the choice of buying a new $5,000 headlamp assembly or parking the car forever, but without retarding LED evolution. The Americans are pretending there's no such problem.

I'll grant you that LEDs will fail in use, some long before their touted 50,000 hour lifespan. Every part of a car fails in time, some prematurely. That's nothing new. Compared to everything else, a well-designed LED system on average will last longer than HID and certainly longer than halogen. It can certainly be pushed as a "life of car" item if 99% of units last that long. The automaker can then afford to fix the 1% that don't under warranty. And IMO LED can be that reliable. There are three main problems in automotive use-power spikes, vibration, and heat. A properly designed driver board takes care of power spikes. Indeed, this is the easiest part to do, and the longevity of far more complex electronic devices in cars like navigation systems tells me this isn't a problem at all. Next is vibration. If anything, this is probably what killed most HID systems but LEDs are inherently very immune to vibration. The main reason for failure due to vibration might be overheating if the LED becomes unglued from the heat sink. But proper design can make this exceedingly unlikely. Last, we have heat. An LED headlamp assembly is by definition exposed to cooling airflow. Not hard to run the heatsink surfaces around the headlamp, exposing them directly to this airflow. Besides that, there isn't as much heat as you would think (more on that later).

No, cost is not dropping. We're still in the area of the curve where cost remains the same (and very high) but lumen availability increases. Eventually, once emitters are available with luminous efficacy levels sufficient to make more than just boutique headlamps for special low-production models, cost will begin to drop…slowly. LED headlamps will not be cost-competitive with any other headlamps within a few years, or even a couple-few years. The 2018 timeframe is when they'll begin to be cost-competitive with HIDs. In the long-range future, this will accelerate because HIDs will increase in cost due to volume reduction as LED headlamps replace HID headlamps in new high-end and eco-focused car designs.

The primary driver in cost reduction will not be automotive but general lighting. And by most accounts mass adoption of LEDs for general lighting is less than 5 years away. A P7 or MC-E emitter which might sell for $20 now and emit 800 lumens may well sell for $5 and emit 2000 lumens in 5 years. Part of the lumen increase will be due to efficiency, the rest due to being able to run at higher drive levels.

Not so. You still need very bulky headlamp assemblies — you don't need as much room for optical packaging as with present point-source headlamps, but you need a lot more room for thermal and power management.

Here's where I have some design experience. Let's say you want to match an HID lamp with LED. A typical HID is 3500 bulb lumens although probably only 60% of this actually gets on the road due to problems focusing a unidirectional emitter. So that's 2100 lumens output. The focusing problem is much easier with LEDs as they emit in a plane. A TIR optic, even a lousy one, has an efficiency of 85%, so you'll need an emitter with about 2500 lumens. Extrapolating out to five years, it's reasonable to assume we'll be at 150 lm/W, so power input to the LED is about 16.7 watts. Note however that a 150 lm/W LED emits about 45% of the input power as light. This means you must get rid of only 9 or so watts of heat. Since you have the headlight right in a place with access to great airflow this is trivial, and most of the heatsink can be right around the headlamp, even painted to match the car if you want to hide it (paint will only make it a little less efficient). As for the heat from the driver, that's probably less than 2 watts using some figures from my own designs. In fact, I can design 97% efficient LED drivers without too much trouble which can drive arrays of 10 or 20 watts, and they're not much bigger than 1.25" square. So much for bulk. And I don't see why on earth a headlight system like this would cost $5K. My boards can be made in quantity for about $10. Even adding in emitters, optics, housings, and perhaps leveling systems I don't see why it should cost over $1K if mass produced, even using today's level of LED development.

This is still deep in the realm of concept-car fantasies. Not only is such a solution vastly more costly than even the hideously expensive existing LED headlamp assemblies (Audi R8, Lexus LS600h, Cadillac Escalade), but it's also forbidden by regulations worldwide. It will be quite awhile before we see solutions like this on actual, real production cars.

I'll grant you this point as it really doesn't matter. Even if we're restricted to a pair of headlights like today LED has no showstoppers in 5-10 years time.

Halogen will still have at least a 70% market share worldwide in twenty years, solely because of cost and complexity. The halogen market share will of course be lower in Europe, Japan, and to a lesser extent North America, but it will not resemble anything like LED headlamp market domination. Your concept of LED headlamp cost (static and over time) appears to be several orders of magnitude too optimistic.

The problem is who's going to be making these halogens in 20 years? With the impending bans on household incandescents the major manufacturers have stated they'll no longer be making incandescent bulbs by about 2015. China has said something similar as a matter of national policy. If indeed automotive remains the last bastion of halogens, they're going to end up costing way more than today given that they'll be made in miniscule numbers compared to today. That will only help the market penetration of LEDs.

Like I said earlier, this seems to be more about what the automakers want as opposed to what is technically feasible. My prediction if your predictions turn out to be true will be that we'll have a huge market in made in China $200 LED halogen retrofits which have the same glare problems as today's HID retrofits. They'll work great from the end user's perspective, of course, since most people couldn't care less about blinding other drivers. But in the end, the automakers will have done their customers and the public another huge disservice but refusing to give the consumer what they want (other than the few willing and able to fork over an exorbitant premium just for the "privilege" of having adequate headlights that don't fail regularly).

 

[-Virgil-]

You do know that future predictions are fraught with peril? Remember the famous "640K ought to be enough for anybody"?

Totally different kind of prediction.

Anyway, if indeed the auto industry consensus is that LED won't have a mainstream volume presence for another 9 years, and won't dominate even in 20, it's not because LEDs aren't up to the task. Rather, it's because this is what the auto industry wants.

I didn't say anything about the auto industry. We're not talking about the automakers' marketers and product planners and stylists and designers here, we're talking about the automotive lighting companies' researchers and scientists and engineers who are actually working to make LED headlamps a reality. It is in their definite interest for every car to have the most advanced lighting technology — if they had a way to make it happen next year, they would definitely do it!

An HID system which might have cost $300 if made in volumes sufficient to make standard equipment on all cars ended up costing $1500 instead.

Sure, but halogens cost a tenth of your lower figure!

I'll grant you that LEDs will fail in use, some long before their touted 50,000 hour lifespan. Every part of a car fails in time, some prematurely. That's nothing new.

What's new is that the evolution of LEDs is such that it's entirely plausible for the emitters Carla CarOwner needs for her headlamp are no longer manufactured when she needs them.

A properly designed driver board takes care of power spikes. Indeed, this is the easiest part to do, and the longevity of far more complex electronic devices in cars like navigation systems tells me this isn't a problem at all.

The interior of the car, where the nav screen lives, is a cushy environment compared to outside out front where the headlamps live. And even the best-designed system or component usually gets "cost-reduced" to the point of less-than-calculated reliability in the automotive sector. I still do not think you have a grasp of exactly how severe the cost strictures are in automotive engineering and build.

Next is vibration. If anything, this is probably what killed most HID systems

Err...no. HID systems are essentially vibrationproof.

An LED headlamp assembly is by definition exposed to cooling airflow. Not hard to run the heatsink surfaces around the headlamp

It is in fact quite difficult to do so effectively and cost-effectively (to do it, you need heat pipes, which are not cheap).

there isn't as much heat as you would think (more on that later)

There may not be as much heat as you think I think, but there is definitely more than I think you think!

I don't see why on earth a headlight system like this would cost $5K.

Your architectural(? Or non-mobile, anyhow) lighting experience does not translate to the automotive realm as directly as you are assuming.

The problem is who's going to be making these halogens in 20 years?

The same people who're making them now.

With the impending bans on household incandescents

Most lighting efficiency regulations stipulate a minimum luminous efficacy. They are not bans on any particular technology. There are advanced tungsten-halogen lamps with efficacy equal to good CFLs in the prototype stage now, and more in development. We have not seen the last of the glowing filament, though we're certainly going to see some novel evolutions of it in the near- and mid-term future.

the major manufacturers have stated they'll no longer be making incandescent bulbs by about 2015.

GE has stated they'll have high-efficacy incandescent bulbs on the market in a couple years, and they're a major maker. Again, you're working from the framework of household and other non-mobile lighting, and while some evolutionary timelines overlap with those in the automotive world, many do not.

we'll have a huge market in made in China $200 LED halogen retrofits

I wouldn't worry too much about it until someone comes up with a way to make a cylindrical lambertian emitter of about 1.5mm diameter and 8mm length. ;-)