Tail light application

[Shimee]

Hi, I'm a new guy. :wave:
I didn't notice an introduction thread so I'm sorry if I may have overlooked it.

For a few years now I've been tinkering with hobby level electronics. Just recently I've taken a course involving advanced level electrics (chemical engineers don't really learn much about electricity). So I've picked up a good deal about resistance, calculating current load, differences between parallel and series, finding voltage drop in wire lengths, etc, but most importantly how to solder :candle:. I've used this to help me create custom wiring harnesses for aftermarket horns, fog lights, wire up foot well lights, make myself a custom glove box light, and so on. As the title suggests, I want to do something a little bigger; I want to create a series of perfboards and replace the horrible incandescent bulbs in my tail lights.



I've researched some background for my goal and found this forum along the way (I'm surprised at how helpful and not "assholish" the members here are, compared to some forums I visit[ed]). I found this site which converts MCD to Lumens: http://led.linear1.org/lumen.wiz. Which poses my first question:

• How do I compare LED brightness? I want to make a board of 15 leds in 3 rows for the reverse lights. The leds I'm looking at are 140° leds with a rating of either 110,000 mcd or 230,000 mcd. Would 110cd be more than enough?
• Equally so, I'm looking at the same type of led for the brake/tail portion and the turn signal portion. Would 110cd be overkill? Would 50cd be more then enough for brightness?

My original plan was using superflux leds which have a rating of 25000mcd. Many people use these for tail lights but I can get bright leds for a little bit less. It's just that I don't know how much is too bright.


Next is my inquiry about wiring the LEDs. First off:

• Where can I get some circuit boards? Radioshack has some but I think they're trying to ripoff customers. And I've had horrible luck on eBay. I just don't know what to search.

THe thing is, I know what circuit boards look like. Its just I dont know where to GET THEM. Ideally I would like a board that looks like this:

Where the orange portions are copper substrate. So that I can simply solder the leds to the board and not screw around with needed to connect the leds by their feet. The board can either be like that with width-wise rows or height-wise rows. But honestly, just access to a board supplier would be good enough.



That said, I've been using this site to rough up an idea of how I could make my circuit: http://ledcalc.com/

Using the parameters 12v, 3.5v-drop, 100mA current, and 15 leds, you can see it says to do five parallel columns with 3 series-leds in each row. This is what I was planning on for my reverse and turns. However, for my brake/tail portion I want to do 10 columns by 4 rows.

• How can I accomplish this? Can I follow the same pattern but just add on 5 more columns and 1 more row?

If you don't want to read all of that, could some one atleast answer this for now:

• Is 110,000mcd x 15leds plenty bright for reverse lights? And is 230,000mcd overkill as far as brightness goes. How bright IS 230,000mcd?

 

[bshanahan14rulz]

The way that the LED calculator figures how many to put in each series "row" is by looking at the available voltage (12) and subtracting an LED's Vf (3.5) until there isn't enough V for another LED to drop (the last 1.5V). This last 1.5V is dropped by a resistor on that row using ohms law, the leftover voltage, and the current you want to run through that string.

Therefore, in order to have 4 LEDs in each row, you would need a source voltage of 14V. You are probably designing for 12V, so 4 in a row would not work. This says nothing about how you can arrange the LEDs, though. Most people get perf board that instead of having rows, has individual pads for each hole, and make the component-to-component connections with jumper wire or something similar.

CPF is the place to go to learn about LEDs, but HIDPlanet's forums has a few very devoted members with lots of help in retrofitting LEDs into conventional taillamps. It is on my list of things to do as well, just want to get a replacement flashlight first.

Also, look into regulating at least the voltage for your entire circuit, since resistors will regulate the current for each string. Also, a member/moderator at the aforementioned forums sells a controller board for tails that can handle many functions. He calls it the Opti-Drive. I don't remember if he sells kits or just the board with some harder to find components, but I think they aren't quite full kits... Either way, that should get you started.

 

[-Virgil-]

Welcome to the board. You need to be careful doing what you have in mind; there are a lot more safety-critical and regulated aspects of the performance of a car's lights than might seem apparent on first thought.

There are specifications for minimum and maximum intensity, for each different function, through a large range of horizontal and vertical angles. That's to make sure that not only can the guy directly behind you see and recognise your brake lamps as brake lamps (tail lamps as tail lamps, etc.) when he's sitting at about the same height as you, but so can the guy in the next lane over to the left, sitting down low in his Corvette...and so can the guy in the next lane over to the right, sitting way up high in his semi truck...and so can the guy on the on-ramp in his SUV. The optics are crucial for spreading the right amounts of light through the correct range of angles, and it is extremely unlikely you'll achieve an acceptable result by just sticking a board full of LEDs behind lenses meant to work with filament bulbs, so you'll need to pay some attention to retrofitting optics, not just light sources. This PDF has a great deal of info on how to design and implement LED vehicle lights, including information on intensity requirements.

There are also specifications for the minimum intensity ratio, again through a large range of H and V angles, between functions that share a lit compartment (brake/tail, for instance, or park/turn). That's so that your taillamps can't be mistaken for brake lamps, or your brake lamps for taillamps. The principal test points within the beam have to have at least 5:1 ratio, and other points have to have at least 3:1.

There are specifications for minimum projected active illuminated surface area, to make sure that the lamps, when lit, are "big" enough to do a reliable job of grabbing attention and quickly and accurately conveying the intended message. For LED lamps, there are lumen- and color-maintenance requirements to make sure the heat buildup caused by prolonged operation doesn't drop the lamps' output below the required minimum or skew its colorimetry outside the regulated boundaries -- even high-end name-brand LED lamp units (Peterson, Truck-Lite, etc.) lose significant light, and especially the amber ones shift color, with prolonged operation. And that's with rather sophisticated thermal management and control circuitry to compensate for emitter junction temperature.

By all means experiment and homebuild -- it can be done; I have seen some top-flight LED retrofit work out of this guy, for example -- but please keep in mind that you're working with crucial crash-avoidance devices, not toys. They have to work correctly in every respect to provide adequate safety performance, not just light up and "look good enough" in your subjective opinion.

 

[Shimee]

First off, thanks for the replies. This is great info.

The way that the LED calculator figures how many to put in each series "row" is by looking at the available voltage (12) and subtracting an LED's Vf (3.5) until there isn't enough V for another LED to drop (the last 1.5V). This last 1.5V is dropped by a resistor on that row using ohms law, the leftover voltage, and the current you want to run through that string.

Yup, I'm familiar with this. Learned about forward voltage and reacquainted myself with voltage consumption and additive voltage (actually additive resistance by Ohms law) "using up" the voltage source. I feel kind of stupid not thinking of this to explain why that calculator was doing what it was doing.

Therefore, in order to have 4 LEDs in each row, you would need a source voltage of 14V. You are probably designing for 12V, so 4 in a row would not work. This says nothing about how you can arrange the LEDs, though. Most people get perf board that instead of having rows, has individual pads for each hole, and make the component-to-component connections with jumper wire or something similar.

Now that I think about it I realize a perfboard with pads would be a better choice. It would allow for more flexibility.

Would you know of a good place to get this kind of perfboard?

Also, look into regulating at least the voltage for your entire circuit, since resistors will regulate the current for each string. Also, a member/moderator at the aforementioned forums sells a controller board for tails that can handle many functions. He calls it the Opti-Drive. I don't remember if he sells kits or just the board with some harder to find components, but I think they aren't quite full kits... Either way, that should get you started.

Thanks for pointing this out, I've heard of a PWM before. Although I looked it up on eBay and it seems like he is only selling kits. I wish he still sold a fully made board.

While trying to search for it I stumbled across this thread: http://spyderchat.com/forums/showthread.php?t=26031. It gave me a few pointers on what to do. It atleast showed me what the PWM looks like and who the user is that makes them. Maybe I'll try getting in touch with him.

There are specifications for minimum and maximum intensity, for each different function, through a large range of horizontal and vertical angles. That's to make sure that not only can the guy directly behind you see and recognise your brake lamps as brake lamps (tail lamps as tail lamps, etc.) when he's sitting at about the same height as you, but so can the guy in the next lane over to the left, sitting down low in his Corvette...and so can the guy in the next lane over to the right, sitting way up high in his semi truck...and so can the guy on the on-ramp in his SUV. The optics are crucial for spreading the right amounts of light through the correct range of angles, and it is extremely unlikely you'll achieve an acceptable result by just sticking a board full of LEDs behind lenses meant to work with filament bulbs, so you'll need to pay some attention to retrofitting optics, not just light sources. This PDF has a great deal of info on how to design and implement LED vehicle lights, including information on intensity requirements.

That PDF was a great read. I've looked at the philips site before but never noticed it. I read on some of the points detailing the importance of angle intesity of LEDs and how this is probably the most important factor in rating LEDs.

And I don't mean to come off as ignorant, but my goals are simply hobbyistic. Obviously I'll never achieve industry standard. I just don't have the equipment nor the mechanical/optical engineering background to achieve this. But I have to try to get as close to that threshold as possible.

There are also specifications for the minimum intensity ratio, again through a large range of H and V angles, between functions that share a lit compartment (brake/tail, for instance, or park/turn). That's so that your taillamps can't be mistaken for brake lamps, or your brake lamps for taillamps. The principal test points within the beam have to have at least 5:1 ratio, and other points have to have at least 3:1.

Looking around for LEDs I've found that its very easy to find some that are intensely bright, enough to be more than suitable for the application. And this is where the PWM unit comes into play. Without it the LEDs would be constantly flared up, blinding drivers behind me.

Although the most important thing to consider is the radiation angle of the LEDs. Obviously the typical dome-cap LEDs would be ill suited with their pathetic viewing angle of 35° or so. This is why I'm looking at LEDs that are rated at no less than 140°.


There are specifications for minimum projected active illuminated surface area, to make sure that the lamps, when lit, are "big" enough to do a reliable job of grabbing attention and quickly and accurately conveying the intended message. For LED lamps, there are lumen- and color-maintenance requirements to make sure the heat buildup caused by prolonged operation doesn't drop the lamps' output below the required minimum or skew its colorimetry outside the regulated boundaries -- even high-end name-brand LED lamp units (Peterson, Truck-Lite, etc.) lose significant light, and especially the amber ones shift color, with prolonged operation. And that's with rather sophisticated thermal management and control circuitry to compensate for emitter junction temperature.

Hmm, I'll have to keep this in mind. I didn't consider this until you mentioned it and when I came across it in that article. I'm hoping this doesnt affect my attempts too much. My tail lights are rather spacious inside so the LEDs wouldn't be too confined causing them to heat up.

By all means experiment and homebuild -- it can be done; I have seen some top-flight LED retrofit work out of this guy, for example -- but please keep in mind that you're working with crucial crash-avoidance devices, not toys. They have to work correctly in every respect to provide adequate safety performance, not just light up and "look good enough" in your subjective opinion.

The funny thing is that even the shoddiest LED retrofit would be better than a filament bulb setup. LEDs vastly supercede filament bulbs in terms of activation time and brightness. But radiation angle is still the biggest factor here, like you mentioned above.

 

[bshanahan14rulz]

I think I've seen the guy who's behind the work at clearcorners, something about finding a good way to dim the lights, since they were too bright or something.

Scheinwerfermann, check out beginning of this video by AZDave from HIDP. All his lights he made himself, along with the controllers etc.. For show only, of course ;-) they are dimmed down to more legal levels for on-road use.

http://www.youtube.com/watch?v=K6sZg3iL5EY
http://www.youtube.com/watch?v=pgexwYy8H28 <<long video, only first part has lighting, then credits roll and he plays some music or something and shows a buncha pics.

 

[-Virgil-]

even the shoddiest LED retrofit would be better than a filament bulb setup. LEDs vastly supercede filament bulbs in terms of activation time and brightness.

Sorry, no. Good lights are good, bad lights are bad, no matter what light source or optical technique is used. There are good filament-type lamps and bad ones, and there are good LED lamps and bad ones. A good filament-type lamp gives better safety performance than a bad LED lamp. "LEDs supercede filament bulbs in terms of brightness" is not even a cogent sentence; I think you mean "exceed" rather than "supercede", and even if that's what you meant, it's still not true. LEDs do have a nearly instantaneous rise time, which is definitely beneficial in brake lamp applications, but a rise time about 80% as good as LEDs can be had from filament bulbs with a simple preheat circuit and instantaneous rise is not necessarily beneficial in turn signal applications (and may in fact be a disadvantage).

 

[Shimee]

http://www.youtube.com/watch?v=K6sZg3iL5EY
http://www.youtube.com/watch?v=pgexwYy8H28 <<long video, only first part has lighting, then credits roll and he plays some music or something and shows a buncha pics.

Hell I'd use that on the road. I saw that video before and it was the most awesome thing I've ever seen.

Sorry, no. Good lights are good, bad lights are bad, no matter what light source or optical technique is used. There are good filament-type lamps and bad ones, and there are good LED lamps and bad ones. A good filament-type lamp gives better safety performance than a bad LED lamp. "LEDs supercede filament bulbs in terms of brightness" is not even a cogent sentence; I think you mean "exceed" rather than "supercede", and even if that's what you meant, it's still not true. LEDs do have a nearly instantaneous rise time, which is definitely beneficial in brake lamp applications, but a rise time about 80% as good as LEDs can be had from filament bulbs with a simple preheat circuit and instantaneous rise is not necessarily beneficial in turn signal applications (and may in fact be a disadvantage).

I meant to say "supersede" and spelled it wrong. Although use of that word on this topic is a matter of opinion. Many people such as optics engineers, Philips, and including myself believe LEDs should replace (ie: supersede) filament bulbs. No amount of arguing will change my viewpoint as filament bulbs in my opinion are inefficient, outdated technology. You on the other hand think LEDs are better but not necessarily a replacement (ie: exceed) for filament bulbs. To each his own.

As far as my original argument, I was basing it solely upon performance between LEDs and filament bulbs. When you factor in real world uses there are a slew of nuances that come in play. One being the housings used (such as HIDs in halogen reflector bowls) and another being human-eye perception. Obviously fluted lenses are made to diffuse light across its entirety, which is ideal for filament bulbs. LEDs would suffer from this since their strong point is directed radiation (I luckily have clear lenses for the most part). The other concern is brightness. I personally hate cars with LED tail lights where the LEDs are so &#@^ing bright that it hurts my eyes. This is why I'm trying to figure out what mcd brightness I should look for.

 

[Neondiod]

Hello!

Firtly, looking at how many different requirements a cars light has to meet, it's quite surprising that they look so different from car to car. Rear lights on some cars for example:

* Some are very big, some has no reflector just the bulb give the light.
* Some is strong, some is weak.
* Some has a narrow beam, some is almost omnidirectional.
* Some is red, some is pink, some is violetish red, some is orange red some are deep red, some are clear glass, some is black glass.
* Some has one bulb, som has 9 bulbs.
* Some has smooth glass, some has fasetted glass, some has diffused glass.
* Some is filament bulbs, some is leds (is there a car with CCFL-tubes?)
* Some leds is shining stable, some leds is flickering (very anoying IMO).

And all the above is originaly equipped standard cars with fine new lights.


Second, I can't help with any mcd numbers for a rear light. But a 21W filament bulb (break, turn signal, back up-light) is about 300 lumen. A 5W filament bulb (position light) is about 50 lumen.


BR

 

[-Virgil-]

Many people such as optics engineers, Philips, and including myself believe LEDs should replace (ie: supersede) filament bulbs.

Yes, of course they should - in lighting devices designed and engineered to use LEDs as light sources. That does not mean homebuilt "gee I guess that's prolly bright enough" driveway retrofits make the lights work better than bulbs. Mostly they don't. This isn't an exciting duel of opinions, it's a boring ol' matter of physics, regulations, and safety performance. You are looking at modifying devices other drivers directly rely on to determine what you're doing...in a fraction of a second...at high speed. It would not be wise to dismiss the consequences of getting it wrong.

I'm trying to figure out what mcd brightness I should look for.

Well, brake lamps and red rear turn signals have to emit between 80 and 300 candela on axis, for example, and there are 1,000,000 mcd in one candela. Take it from there.

 

[Neondiod]

Well, brake lamps and red rear turn signals have to emit between 80 and 300 candela on axis, for example, and there are 1,000,000 mcd in one candela. Take it from there.

Isn't it 1000 mcd = 1 candela ?

The latest Toyota RAW4, has 9 superflux leds in the combined tail/brake-light.

 

[-Virgil-]

Bah, watch me get carried away and type too many 0's! One millicandela (mcd) is one one-thousandth of a candela; one candela is indeed 1,000 mcd.

The 2nd-generation Prius (just replaced by the new one) uses six emitters in its brake lamp. Truck-Lite makes good single-emitter 4-inch round brake lamps and turn signals. And then there's the Osram Joule lambertian LED light source, used by itself in a growing number of Ford-product (and other) brake lights. It obviously can be done; it doesn't require a huge emitter count, but it does require sophisticated, highly efficient optics.