Custom LED tailights

[FrankLS]

So I started doing some LED testing last night. I have both red and amber LED's in different mcd. Im trying to figure out which ones will be just bright enough.

The lowest mcd im testing is 35k and highest being 160k. I know I wont be able to get a good grasp on how bright they are until I place the exact amount i will be using on the breadboard and behind the lens.

 

[Kumba]

See, I want to do something like this, but using a simple micro-controller so I can get rolling turn-signals and a quick brake light fade on/off/dim. Then I always wondered if I could get a body-color one-way tint or paint over the lens so that the light comes through, but when off it blends together.

if only ideas were like money and time

 

[FrankLS]

I woke up very early this moring and began cuting and shaping the breadboard. I made myself some guides out of some foam craft sheets and used those to cut the bread board.

- Heat gun, clamp and pot = perfect shape. Just giving the panels the same shape of the lens.

- It took a bit of heat but finally got them to the desired size and shape. They became a bit darker due to the heat, but nothing a can of silver paint cant cover.

So I ended up cutting the "ridges" a bit, to allow the breadboard to sit further inside the housing. Fortunatley the cuts didnt go through to the other side so will not have to worry about weatherproofing the backside.

-I placed a dab of hot glue in some places. Just trying to align the panels.

-Heres how the panels sit inside the tailights. I still need to do a bit more of aligning, but im very close to done with the panels. Surprisingly the hot glue is holding very well against the plastic housing. Im debating if I should just use hot glue to hold the panel in place when im all done.

-The small black dots on the outer edges of the panels are not cutting errors or holes in the panel, just sharpie markings when i was sanding the panel down.

Another problem I ran into. The height of a "standard" LED will be cutting it close. It will almost be touching the lens and I dont plan on cuting down the breadboard anymore. I think im going to have to purchase low profile LED like, 5 chip superflux, which are half the height of standard 5 mm.

-Heres a picture showing the clearance issue.

 

[FrankLS]

See, I want to do something like this, but using a simple micro-controller so I can get rolling turn-signals and a quick brake light fade on/off/dim. Then I always wondered if I could get a body-color one-way tint or paint over the lens so that the light comes through, but when off it blends together.

if only ideas were like money and time

body colored one way tint, LOL. If i had the money, I would buy 100 of every LED in ascending MCD and different sizes.

As far as the wiring. Im trying to go with simplicity as im no electrical guru. Just led's, resistors, and maybe a voltage regulator to make sure only 12v reach the led's.

 

[joegr]

So I started doing some LED testing last night. I have both red and amber LED's in different mcd. Im trying to figure out which ones will be just bright enough.

The lowest mcd im testing is 35k and highest being 160k. I know I wont be able to get a good grasp on how bright they are until I place the exact amount i will be using on the breadboard and behind the lens.

I suggest that you pick the brighter ones and then increase the size of the current limit resistor if needed to reduce the brightness. Ideally, you should make a current regulator circuit so that brightness won't vary with electrical system voltage. Even better would be to have a temperature compensation circuit that would reduce the current as the temperature gets lower. (LEDs get brighter when they get colder.)

 

[joegr]

...
Another problem I ran into. The height of a "standard" LED will be cutting it close. It will almost be touching the lens and I dont plan on cuting down the breadboard anymore. I think im going to have to purchase low profile LED like, 5 chip superflux, which are half the height of standard 5 mm.
...[/IMG]

Alternately, you could use SMT LEDs on peel-n-stick flex strips. (I assume that you don't want to do your own PCB layout.)

http://www.superbrightleds.com/cgi-...=DispPage&Page2Disp=/light_bars-flexible.html

 

[joegr]

body colored one way tint, LOL. If i had the money, I would buy 100 of every LED in ascending MCD and different sizes.

As far as the wiring. Im trying to go with simplicity as im no electrical guru. Just led's, resistors, and maybe a voltage regulator to make sure only 12v reach the led's.

No need to buy different brightnesses. Just get the bright ones and vary the current going to them to control brightness.

You want current regulation with LEDs, not voltage regulation. Current regulators are easy to rig up.

 

[FrankLS]

Those are good ideas, ill be doing some research. I know LED's dont produce heat but can the resisors produce heat?

 

[FrankLS]

I didnt know there were stick on LED's. I was actually planning on placing them individualy and wiring them in groups of 3 with one resistor per group. Once i purchase the LED's ill be able to figure out the exact ohm resistor and how many per group.

 

[joegr]

I didnt know there were stick on LED's. I was actually planning on placing them individualy and wiring them in groups of 3 with one resistor per group. Once i purchase the LED's ill be able to figure out the exact ohm resistor and how many per group.

LEDs do produce heat. The size/brightness that you are using don't generally dissipate enough to worry about, but the high brightness ones certainly do. They require heat sinks. Yes, the resistors will dissipate some heat as well.

Red LEDs usually have a voltage drop of 1.8 to 2.2 volts. That means you could probably put four or five in series.

The important figure is the current rating of the LED. You want to size the current limit resistor to result in that current or less. Current depends of the LED voltage drop (varies), the electrical system voltage (varies) and the resistor value. You can use a constant current regulator instead of a resistor and not have to worry about the variable voltage drop or the varying electrical system voltage. (Of course, often a fixed resistor is good enough.)

 

[Kumba]

Are the tail lights PWM or Linear driven? I know that there is always a slight voltage drain across them which means LED's need a resistor to actually go dark. Just wondering how they are driven otherwise.

 

[joegr]

Are the tail lights PWM or Linear driven? I know that there is always a slight voltage drain across them which means LED's need a resistor to actually go dark. Just wondering how they are driven otherwise.

I think (but haven't verified) that it's neither. I think that there is a relay to switch the bulbs on and off. Parallel to the relay is a high resistance resistor that leaks some current out to the bulb. If the bulb is there, the current is so low that the voltage across the bulb is very tiny, almost zero. If the bulb is not there (burned out) then there no current flow so the voltage goes to 12V or so. A voltage detector on the line sees this and uses it to know that the bulb is burned out. (Voltage while off = bulb out.) I think there is also a current detector so that burned out bulbs can also be detected when the bulb is being driven. (No current while on = bulb out.)

This is how they probably do it, but it is not the only way they could be doing it. I'd have to pull apart a REM or FEM to know for sure.

 

[Kumba]

I think (but haven't verified) that it's neither. I think that there is a relay to switch the bulbs on and off. Parallel to the relay is a high resistance resistor that leaks some current out to the bulb. If the bulb is there, the current is so low that the voltage across the bulb is very tiny, almost zero. If the bulb is not there (burned out) then there no current flow so the voltage goes to 12V or so. A voltage detector on the line sees this and uses it to know that the bulb is burned out. (Voltage while off = bulb out.) I think there is also a current detector so that burned out bulbs can also be detected when the bulb is being driven. (No current while on = bulb out.)

This is how they probably do it, but it is not the only way they could be doing it. I'd have to pull apart a REM or FEM to know for sure.

If that is the case, then something like this would work great: http://www.maxim-ic.com/app-notes/index.mvp/id/4316

You would probably want to put a FET across the LED output to get more current capability, as well as divide up your array, but it would certainly simplify the set-up. I am not sure if the voltage drain for the bulb-out function would dissipate through the chip without hurting it or if you'd still need the resistor. To be on the safe side I would probably go with the resistor.

Wonder who sells those evaluation boards. Really not in the market to have a run of PCB's made. I guess if we got enough made, they could be done cheaply.

 

[joegr]

...
Wonder who sells those evaluation boards. Really not in the market to have a run of PCB's made. I guess if we got enough made, they could be done cheaply.

You'd have to check with your local Maxim sales rep to see if he would sample you one. (They probably don't sell them, but they will give them away - if you are in the right industry.)

As for having some of your own boards fabricated, if you don't mind long lead-time (a month or so), there are some foreign board houses that have amazingly low prices for runs of one to five boards. (Well under $100 for five with shipping.)

 

[Kumba]

If that is the case, then something like this would work great: http://www.maxim-ic.com/app-notes/index.mvp/id/4316

You would probably want to put a FET across the LED output to get more current capability, as well as divide up your array, but it would certainly simplify the set-up. I am not sure if the voltage drain for the bulb-out function would dissipate through the chip without hurting it or if you'd still need the resistor. To be on the safe side I would probably go with the resistor.

Wonder who sells those evaluation boards. Really not in the market to have a run of PCB's made. I guess if we got enough made, they could be done cheaply.

Actually, I wouldn't mind using these on some angel eyes to make them "wink" with the turn signal. Don't know how cheesy it would be, but might be neat. Just put a small nc relay in front of it hooked up to the turn signal that supplies voltage to the stop side. The dim side would be hooked up to voltage all the time. When the turn signal blinks, the stop pin goes low and the angel eye dims or "winks" with the turn signal. Not sure if there's an issue with the umnitza eyes or not though.

I really like this design. No external PWM, no external voltage regulation, and a pretty simple self-contained circuit. Just might need a higher-current driver on the output. Incase you were trying to power 200 LEDs or something. Maybe include a pot so you can adjust the level of dimming that you get via R2/R3. Don't suppose anyone is good with PCB layouts and want to put together something for a quote?

 

[joegr]

... Don't suppose anyone is good with PCB layouts and want to put together something for a quote?

It's a lot of what I do, but you wouldn't want to pay my price and I don't have time for freebees right now. Maybe someone else?

 

[Kumba]

If I had these made, would anyone else be interested? Say they were $50/ea or so (rough ballpark).

 

[2000sport]

I may be able to find time to draw something up and get some quotes through some of our partners at work (pentalogix, screaming circuits, etc)

My company owns Cadsoft Eagle, which is a PCB design software.

I just emailed myself a reminder. If I have time tomorrow, I will mess around in Eagle and draw something up, and make some calls.

That application note includes full schematics for the board, so this shouldnt take long.

 

[Alax7]

Actually, I wouldn't mind using these on some angel eyes to make them "wink" with the turn signal. Don't know how cheesy it would be, but might be neat. Just put a small nc relay in front of it hooked up to the turn signal that supplies voltage to the stop side. The dim side would be hooked up to voltage all the time. When the turn signal blinks, the stop pin goes low and the angel eye dims or "winks" with the turn signal. Not sure if there's an issue with the umnitza eyes or not though.

If you wire up angel eyes to the turn bulb, they "wink" or " blink" when the turn bulb goes on and off when turning. This only works if you keep the amber filament bulbs, for some reason, it stops working this way when you install LED turb bulbs.

 

[Kumba]

I may be able to find time to draw something up and get some quotes through some of our partners at work (pentalogix, screaming circuits, etc)

My company owns Cadsoft Eagle, which is a PCB design software.

I just emailed myself a reminder. If I have time tomorrow, I will mess around in Eagle and draw something up, and make some calls.

That application note includes full schematics for the board, so this shouldnt take long.

I am willing to do a limited production run depending on the final cost of PCB and assembly. Being 0.35 TQFN I wouldn't really expect anyone at home to be able to solder it up. Looks like a simple enough circuit though. Not sure if you should stick with 0.35 SMT or use 0.5 for everything but the chip. I guess stick with 0.35 and we can just get a quote on a complete assembly.

About the only thing I might want to change is putting a pot in place of R1 and R2 so you could adjust the "stop" and "tail" brightness, but I haven't a clue what values you would want to throw in there that would be meaningful or how it would change C3. Guess it's best to leave it alone.

 

[2000sport]

I will review the schematic and determine some options.

One of my co-workers has a ton of experience with LED driver circuits, so I will bounce some questions off of him as well.

 

[owlman]

I know someone who can help, he works for cheap, though his quality control is lacking.

 

[joegr]

... Being 0.35 TQFN I wouldn't really expect anyone at home to be able to solder it up. ....

Actually, you can get really good results if you use solder paste and a toaster oven. The key is to use the right temperature profile. (You have to put the board in at a low temperature for a while to preheat all the parts, then a high temp for a bit to flow the solder, and so on.)
When the paste melts, the solder will draw the part leads to the centers of the pads, so the parts placement doesn't have to be exact, just very close. This assumes that the pads on the PCB are well designed.

 

[Kumba]

I know someone who can help, he works for cheap, though his quality control is lacking.

He looks way to over-stressed. But I do here he likes cheezburger.

 

[mlara1129]

I know someone who can help, he works for cheap, though his quality control is lacking.

haha i think i know his cousin...he prefers to sleep tho...