Arduino powered taillight project.

[bttws]

Hey all, first the pictures.... then I'll make another post.

The cut lens (I'm guessing most people have not seen the inside of these) There is a layer of glue connecting the 2 halves of the lens. I tried baking it in the oven to warm the glue, but I was afraid of melting it and Dremeled it instead. I am guessing that I can find a sweet spot with temp and time to get them apart without melting.


The LED array, I used hotglue and lexan sheets to make the separators. The used black paper keep the light from leaking (very basic for now)


A closeup of the LEDs, They are high flux LEDs, the are sometime called automotive grade because the can handle higher voltages, are brighter, are more rugged, and most important, they have 2 extra leads per side. This leaves a backup in case of a solder joint failure.


The back, not the best solder job, but it works. that was before I got a better soldering iron. (yes those are plastic knifes)


The arduino uno and circuits running it. Right now very simple, just 3 transistors and a few resisters.

[bttws]

For starters, some light reading material, the us DOT regs on lighting....
http://www.law.cornell.edu/cfr/text/49/571.108

Thanks for all the suggestions.....

I have been thinking about this for a long time. I drive a truck so I have plenty of time to think about Ideas, I also look at car taillights all day. I hope to have a car next year. When I do, I plan on doing extensive modification to the interior. I want to add a Tesla S style full touch screen control and digital dash. I will be adding a ton of sensors, so some of the functions of the taillights would be specific to my car only. I would like to produce a product, so I would keep that in mind for people who do not wish to permanently cut/modify their cars. I would most likely buy new lenses form DMC Houston and hand build the electronics and hardware. With all the work involved, I would imagine they would not be cheap to make. How much would you pay for a set?

my full Ideas

Software
The goal is to right a easy to use customizer to compose a set of instructions to taillights can read. When done the program would send the files to a SD card. You would then put the card in the taillgihts and reset.
A few things I have in mind....
I want the program to be smart enough to know how bright different lights need to be, so it wouldn't let you make a file with no brake lights or something. I might be able to have the ardiuno check voltage outputs on the leds and revert to a some sort of default if something is wrong.

I want to have a set actions that you create sequences for. The program should also prevent you from using white lights for turn signals, etc.
Each action would have a -Warm up animation -Steady animation and -Cool down animation. I say animation with a grain of salt. An animation may be one frame long and loops as long as the action is active.
Brake animations might be unsafe and may be limited to a 1/2 sec or less warm-up time.

example, lets say parking lamps. When the action is activated ( eg. turn on the parking lamps) the warm up animation would play. This could be a fade up or some sort of sweep animation. After the warm up the steady would play as long as the lights are on. This would most likely be a one frame looped. lets say all 9 red squares. When the lights are tuned off the cool down animation once would play like the warm up animation.
This format would work for all the actions

I want 4 modes on the tailights,
1. Stock.... this is preprogramed in the code and will activate when there is no SD card or there is some kind of error with the sd card file. This would be a basic animation set, that emulates the stock look.
2. Custom... this is what is on the SD card that you make with the computer
3. Parade/carshow.... This mode would play a playlist of animations from the SD card that you make on the computer. This mode will be set back to stock when the car is restarted or a preset speed is reached, to prevent you leaving the mode on.
4 Audio mode (for car shows) Similar to the 3rd mode but this would turn the taillight into a audio visualizer, hooking up to a mp3 player or a car stereo audio line.

Safety
Of course I hope this could improve the safety of the car, I have heard that most people think the DMC taillights could be better and brighter. Being able to utilize all 9 red sections for parking and braking lights would be a huge plus. You also get more space for turn signals. Depending on how well the LEDs test out in the open, I may install a light sensor to change brightness levels depending on time of day. That way lights could be plenty bright at day, but not blinding at night.

My ideas on speed proportional brake lights.
I was thinking of 4 different braking modes. These will be achieved by using gps and accelerometer data.
1. Normal - This just everyday average braking. But could include harder braking under 10 mph. If you slam on the brakes in a parking lot because you see a space, I don't want the system acting like you are skidding to a halt from 55mph
2. Hard - This would be with a moderate deceleration, or brake pressure. This would play a different braking sequence. Maybe one or two flashes then steady.
3 Full/Panic - This be a signaled by a rapid deceleration or very high pressure (at speed). This would play the panic sequence which could be bright flashing brake lights until the brakes are released. Speed is important here because I down want the lights flashing like crazy at a stop light if the brake pressure is high. The taillights would optionally be able to activate a relay to turn on the hazard flashers until stopped then off again when the brakes are released.
4. Skid - This one would be only be possible with a sensor detection wheel spin. This could play a different sequence.

here is how BMW's system works. youtube

A few optional ideas.
Brake light hold - If you have a manual you could have the brakes stay active at a light until you start moving again. so if you sit with clutch in and no brakes on, the car behind will still see you are stopped and have brake lights on.
Automatic deceleration brake lights- If the accelerometer detects and decrease in speed over a certain amount it could send a brake light even if you did not touch the brake.
Probity tailgate warning - Add a proximity sensor and if it detects at close following car at speed, the system could flash the both amber sections once to alert the driver behind ( not sure about this one )
collision alert. My car is going to have a aftermarket lane departure/front collision avoidance system. I am going to tie into that so if it shows impending collision or hard braking it would start a brake action before I press the pedal, maybe saving me from being rear ended.
Automatic Parking lights - If the light values are low enough, the rear parking lights could come on even if the parking lights are not on.

Design

Not much to say yet. I will be using some 3d printing make the enclosure on the back of the lens. It will have to be water tight and sealed with a few selector buttons and sd card slot behind a locked door.
There would be a connector to use stock wiring and connector to go the other side of the car for the other taillight.

Expansion

On my car I will be putting LED arrays behind the side markers and front turn signals. The system can work together and animation sequences would include these sections too. I guess if there is demand, I could make a kit to add these too.


well those are my ideas, if you read it all sorry it was so long. Let me know what you think. I really think it would be fun to build these and see them installed in other people's cars. We'll just have to see how this turns out.
BTW I will be happy to accept any lenses of any type anyone might have laying around I'll put them to good use.

-Kevin

[Tomcio]

That's great information about the light intensity, wavelength and panic light blink speed -- thanks! Weird that they aren't very specific about the rear light intensity, though. I should really hunt down the DOT specs sometime...

No problem
I found all this info while doing research for my LED side markers and rear lights. I like to share
For the actual intensity I decided to not go by the numbers alone and I got a lux meter. They are cheap but work great actually. I think mine was in the range of $10 to $15.
I don't know how accurate this thing is but I don't care. I just need to compare the readings. So, I took some measurements from my other two cars and from some of other cars (I got some strange looks from my friends ). I got an average of all those readings for the front turn signals, side markers and taillights (position, brake, turn, reverse and fog). Then I built a circuit with LEDs and checked the result. This allowed me to choose correct LEDs and use a correct number of them to get the correct light output. This way if someone ever questions my lights they can test them and they will confirm that the light output is correct.
BTW, for the front turn and rear combination lights I ended up choosing 4 chip super flux LEDs. Single chip units were not bright enough even when mounted very close together.
For reverse and daytime running lights I went with 6 chip cool white super flux.

I also didn't think about brake pressure vs. pedal travel vs temperature/moisture being an issue for detecting panic stops. My thought was more that if you were panic-braking, you probably had your foot on the floor, at which point the specific pressure wasn't very important.

You don't even realize how different the brakes are depending on the temperature alone. Our brains are really great at adapting to changing conditions and we don't even notice it. At below freezing the pedal will require more effort than when it's a nice hot summer day. When the brakes are up to temperature they require the least amount of force and travel. Then when they start to overheat you'll need even more effort and the pedal will travel further in.
Then there are things like vacuum assist that changes over time, moisture in the brake fluid, moisture on the rotors/pads. How about a situation where one of the hydraulic systems develops a leak? If one of the calipers fails you'll only have the brakes work on the other system (the hydraulic system is divided into two) and the brake pedal will actually travel almost to the floor without a panic stop.
It would be impossible to adjust switches to work correctly.

I was wary of using an accelerometer because it was a secondary input (the behavior of the car braking) instead of a primary (telling the car to brake), but perhaps I'm overly concerned about that. Your mention of the 6 m/s^2 spec suggests that I am that an accelerometer is a way to go.

I believe that this is the best way to go. The panic function would activate regardless of the driver behavior. So, as you step on the brake pedal the regular lights come on. Then when the deceleration is large enough the panic function is activated. In my opinion it adds to the safety because if you decelerate rapidly (even when not intended) the panic flash will activate. I can imagine a few unfortunate situations where this might be helpful.

Also, it would ensure the lights flash only when you're actually panic-stopping, and not just sitting at a stop light with the brake pedal down hard.

Great point!
Just make sure you mount the sensor in the correct orientation! It needs to read the actual deceleration and that means you will probably need a second board with the sensor pointing forward. Since those sensors are tiny this shouldn't be a problem. A little stand-off board will do the trick.

[jangell]

For starters, some light reading material, the us DOT regs on lighting....
http://www.law.cornell.edu/cfr/text/49/571.108

Thanks!

I have been thinking about this for a long time. I drive a truck so I have plenty of time to think about Ideas, I also look at car taillights all day. I hope to have a car next year. When I do, I plan on doing extensive modification to the interior. I want to add a Tesla S style full touch screen control and digital dash. I will be adding a ton of sensors...

I have very similar plans. When I get to it is another story...

With all the work involved, I would imagine they would not be cheap to make. How much would you pay for a set?

I personally am interested in building my own as a learning experience, but I'll be following your project closely. Good luck!

Safety
Of course I hope this could improve the safety of the car, I have heard that most people think the DMC taillights could be better and brighter. Being able to utilize all 9 red sections for parking and braking lights would be a huge plus. You also get more space for turn signals.

I'm a really big fan of actually having separate lights for taillights, turn signals and brake lights (which makes it clear to other drivers where to look for those indicators even when they're off), and especially of having the turn signals be yellow or clear with yellow bulbs. I don't like how many modern cars seem to be combining the brake lights and tail lights, and red turn signals are not as recognizable as yellow (I found a study once saying that clear lenses with yellow bulbs were most visible for turn signals, actually -- you get the change in color as well as the change in brightness that way). I have no idea how many people share this view, though.

Depending on how well the LEDs test out in the open, I may install a light sensor to change brightness levels depending on time of day. That way lights could be plenty bright at day, but not blinding at night.

That's an interesting idea. I agree that the DMC's taillights are a bit dim in normal use, so that could help.

My ideas on speed proportional brake lights.
I was thinking of 4 different braking modes. These will be achieved by using gps and accelerometer data.
1. Normal - This just everyday average braking. But could include harder braking under 10 mph. If you slam on the brakes in a parking lot because you see a space, I don't want the system acting like you are skidding to a halt from 55mph
2. Hard - This would be with a moderate deceleration, or brake pressure. This would play a different braking sequence. Maybe one or two flashes then steady.
3 Full/Panic - This be a signaled by a rapid deceleration or very high pressure (at speed). This would play the panic sequence which could be bright flashing brake lights until the brakes are released. Speed is important here because I down want the lights flashing like crazy at a stop light if the brake pressure is high. The taillights would optionally be able to activate a relay to turn on the hazard flashers until stopped then off again when the brakes are released.
4. Skid - This one would be only be possible with a sensor detection wheel spin. This could play a different sequence.

here is how BMW's system works. youtube

Skid is an interesting one. I'm very curious about what you'll use to detect wheel spin, since I'd much rather buy an off-the-shelf unit for my own projects than build one if something compatible exists already. It would also make my speedometer conversion a bit easier, if it returned how fast the wheel was turning.

I'm curious about the value of Hard vs. Full/Panic, not because I'm saying it's a bad idea but because I genuinely don't know how useful it would be in a general sense. I see that in the BMW video that they do make the lights brighter, and I also wonder how visible that is in real world use. That said, I'll assume a company like BMW has done their research and know what they're doing.

A trick with flashing all the brake lights too slowly is that people could think you're pumping the brake instead of holding it down, and thus that you're not braking as hard as you really are, but I guess BMW doesn't think hat's a problem either. I'm guessing that 4 Hz rate mentioned elsewhere is fast enough to eliminate that, but the BMW one seemed to blink slower than that (( think).

I've added a high-center stop light (third brake light) to my car, and was considering just having that flash on panic stops.

Automatic deceleration brake lights- If the accelerometer detects and decrease in speed over a certain amount it could send a brake light even if you did not touch the brake.

That's interesting; at least it might get people to back off. Or to think that you're one of those guys who rides the brakes.

Probity tailgate warning - Add a proximity sensor and if it detects at close following car at speed, the system could flash the both amber sections once to alert the driver behind ( not sure about this one )

Might work. It would look like you put on your hazards for a moment, which might make them back off a bit.

collision alert. My car is going to have a aftermarket lane departure/front collision avoidance system. I am going to tie into that so if it shows impending collision or hard braking it would start a brake action before I press the pedal, maybe saving me from being rear ended.

I'm hugely interested in this. What aftermarket system were you thinking of using? I figured I was going to have to roll my own (and thus spend a lot of time learning machine vision and image processing stuff to detect road lines). I've always wanted to see car manufacturers have systems to detect the lines and make it hard to change lanes unless your turn signal was on, and nudge you into he lane if you start to drift out of it. I think some are actually starting to do this, too.

While I would like to build an emergency braking system into my car (my idea being that I'm going to hit something anyway, so the system would slam on the brakes faster than i could so I'd hit it slower; it's something I've been thinking about for more than a decade now, and some manufacturers are now doing something like it to avoid collisions entirely), I don't trust my programming skills enough to use them on critical systems that can get me killed. I don't want a leaf to fall onto a sensor and cause the car to slam on the brakes, for example, so it would require a huge amount of testing. I'd like ABS too, but I won't do it for the same reason, unless I can get a complete aftermarket kit.

Automatic Parking lights - If the light values are low enough, the rear parking lights could come on even if the parking lights are not on.

I just leave my lights on all the time anyway -- I modified the headlight switch wiring to turn the lights off with the ignition. Not that I haven't thought about something similar, though. For example, it's not clear to me if headlights during daytime actually reduce your visibility to other oncoming drivers in a manner similar to some World War II projects where they experimented with hiding tanks and planes by covering them with banks of lights to reduce the contrast between them and the background (which worked, until radar obsoleted it), in which case just running with marker lights would indeed be safer.

Design
Not much to say yet. I will be using some 3d printing make the enclosure on the back of the lens. It will have to be water tight and sealed with a few selector buttons and sd card slot behind a locked door.
There would be a connector to use stock wiring and connector to go the other side of the car for the other taillight.

I also plan on using 3D printing for a lot of my stuff, although I was hoping to reuse the cut-off back of the taillight assembly. I hadn't yet decided if I was going to build a new wiring harness or mate it to the stock one. I was trying to think of a way to seal the assembly with a gasket of some kind to weatherproof it, but still be able to open it again if I needed to fix somebody soldering or something. Hadn't quite worked out the details yet.

Expansion
On my car I will be putting LED arrays behind the side markers and front turn signals. The system can work together and animation sequences would include these sections too. I guess if there is demand, I could make a kit to add these too.

Also something i want to do. I might just buy the LED arrays from you so I can wire it into my own project, just so I don't have to build them too. I want the LEDs to run the entire length of the markers, which may prove slightly tricky as there is only really room to put anything where the existing bulb is (although it's been a few years since I took those out; I may be misremembering).

well those are my ideas, if you read it all sorry it was so long. Let me know what you think. I really think it would be fun to build these and see them installed in other people's cars. We'll just have to see how this turns out.

Sounds great -- good luck!

-- Joe

[jangell]

I found all this info while doing research for my LED side markers and rear lights. I like to share
For the actual intensity I decided to not go by the numbers alone and I got a lux meter. They are cheap but work great actually. I think mine was in the range of $10 to $15.

I like the idea of a lux meter -- I was just going to eyeball it, but using real values makes a lot more sense. Presumably you have to measure the brightness at the same distance to get an result to compare with, much like measuring sound levels with a decibel meter.

BTW, for the front turn and rear combination lights I ended up choosing 4 chip super flux LEDs. Single chip units were not bright enough even when mounted very close together.
For reverse and daytime running lights I went with 6 chip cool white super flux.

Nice -- thanks for the info. I had overly ambitious plans involving quite a lot of LEDs (I think I was up to 450 per assembly -- this is clearly madness, and I really don't want to solder all those pins).

You don't even realize how different the brakes are depending on the temperature alone. Our brains are really great at adapting to changing conditions and we don't even notice it. At below freezing the pedal will require more effort than when it's a nice hot summer day. When the brakes are up to temperature they require the least amount of force and travel. Then when they start to overheat you'll need even more effort and the pedal will travel further in.

Speaking of which a brake upgrade is on my list as well. For some reason I have to pump my brake twice for it to feel good: the first time it dips quite a lot, and the second time it feels firm. No one has been able to figure out what's going on, so I was just going to give in and get the expensive UK braking kit. Well, after I put the engine back in the car, anyway...

(accelerometer I believe that this is the best way to go.

I think you're right -- it will ensure the panic mode triggers at the right time, and only at the time.

Just make sure you mount the sensor in the correct orientation! It needs to read the actual deceleration and that means you will probably need a second board with the sensor pointing forward. Since those sensors are tiny this shouldn't be a problem. A little stand-off board will do the trick.

Yup; I'm not expecting that to be much of a problem. since I plan on having a lot of sensors in my car anyway, I may have a single unit located somewhere in the cabin that tells the brakes when to pulse.

Thanks!

-- Joe

[Shep]

Can I just get a set for my desk??? Honestly, all this talk of patterns and legality and brightness, and I just want to get one for my cubicle...



To be frank, I'm loving this project. And it gives me an idea on what to do with my old taillight housings. Maybe even turn them into a display or something!

(Also, have you tried a heat gun rather than an oven? Might help you melt the glue without destroying the taillights.)

[jangell]

If you're going to have one for your desk, you could write an app for your computer to control the lights based on something going on on your system. I did that for my flux capacitor project, turning it into a CPU activity meter and low memory indicator. The app I wrote is a bit buggy, but it does the job.

-- Joe

[bttws]

I could make a desk version.... Maybe I should do that first. I could have it display network status or something. Maybe a audio visual or some kind of clock....

Bluetooth connection with a smart phone is possible.

[Tomcio]

For some reason I have to pump my brake twice for it to feel good: the first time it dips quite a lot, and the second time it feels firm. No one has been able to figure out what's going on

Joe, this is a little off topic here but I might be able to help you out. First we need to get some facts.
Are you sure that there are no leaks in the system, the fluid is fresh and clean, the calipers are clean and dry? If yes then the main question is if you feel any vibration when you apply the brakes? The best way to check for this is to get the car up to highway speed and step on the brake. Light pedal pressure at first that you can then increase. If you feel any vibration or any pulling to one side and vibration then one or more likely both rotors have high and low spots on them. It is a condition that people call "warped rotors". It's very unusual for rotors to warp but it's fairly common for the brake pads to "burn and melt" onto the rotors when you stop your car with very hot brakes. This causes a high spot on the rotor and that high spot pushes the pads away when you drive. The first push of the pedal needs to get the pads closer to the rotors and the second push is when you get normal operation of the brakes. Turning the rotors on a lathe or replacing them will fix the problem.
So, do you get any vibration when braking at high speed?

[jangell]

Joe, this is a little off topic here but I might be able to help you out.

Thanks. My car hasn't been drivable since last year; I'm in the slow process of doing a 3.0L swap after discovering hole in my 2.8L block. As such, I'll be going from memory.

First we need to get some facts.
Are you sure that there are no leaks in the system, the fluid is fresh and clean, the calipers are clean and dry?

Fairly sure. I had Rob Grady take a look at it at one point, although that was quite a while ago now (2-3 years), and my local shop about a year and a half ago. The system was bled each time, and it didn't appear to be leaking. I'm not sure it's worth checking this out now since the car hasn't run for a year, but I'm guessing I could push the brake pedal and see if anything odd is going on.

The car does have a Saag 50/50 master cylinder in it, if that matters. I admit I was going to use this problem as an excuse to update to the UK performance brake kit, but I don't have the money for that right now either.

If yes then the main question is if you feel any vibration when you apply the brakes?

I never noticed that, only that the pedal was soft on the first pump, and firm on the second. This would happen basically every time I braked, and I got used to double-pumping them. Not super safe, since it increased my braking distance. I never did a test to see if the soft pedal braked as well as the hard pedal, but the soft pedal did still slow the car.

The best way to check for this is to get the car up to highway speed and step on the brake. Light pedal pressure at first that you can then increase. If you feel any vibration or any pulling to one side and vibration then one or more likely both rotors have high and low spots on them. It is a condition that people call "warped rotors".

This is where my memory gets hazy. I remember having my local shop try to diagnose the brakes, reconditioning the rotors (I think) about two years ago. But I can't seem to find the work order for that in my records.

It's very unusual for rotors to warp but it's fairly common for the brake pads to "burn and melt" onto the rotors when you stop your car with very hot brakes. This causes a high spot on the rotor and that high spot pushes the pads away when you drive. The first push of the pedal needs to get the pads closer to the rotors and the second push is when you get normal operation of the brakes. Turning the rotors on a lathe or replacing them will fix the problem.

Ah, I didn't know that -- I'd only heard of warped rotors before, not of high spots from melted pads. I have the rear right tire off the car (need to get access to the parking brake cable to disconnect it when pulling the engine/trans), and I could probably pull the others; would this kind of thing be visible the rotor, or is it subtle enough that to feel for it?

So, do you get any vibration when braking at high speed?

I will be sure to check as soon as the car is roadworthy again. Thanks!

-- Joe