Introduction: DIY LED Car Headlights!
At first phase completion, my first 78 LEDs are dissipating 5.8 watts while my resistors only dissipate 0.85 watts, and the array is already powerful enough to dwarf the two front marker lights, which draw 3 watts each. My final objective is installing, at the very least, 300 LEDs at a total energy expenditure of 21 watts at the LEDs and 3.2 watts at the resistors.
All comments welcome.
I Can pick the laser cutter up either in Florida or in a couple of addresses in Texas. I'm still on the run!
Here's an index:
Step One: Justification and Objectives
Step Two: Exploration and drawing the assembly goals. I decide here to go for two construction phases.
Step Three: I procure the stuff I'll need for phase one.
Steps Four, Five and Six: I build the LED modules, prepare the area to receive them and install them.
Step Seven: I put the first 78 LEDs to a road test.
Step Eight: I put finishing touches to Phase One and start drawing objectives for Phase Two.
Step Nine: Minor improvements.
Step Ten: Brake Lights
Step Eleven: What not to do
Step Twelve: How to do it right.
Step 1: Justification and Objectives
It seems silly to me to start working on something before knowing why.
Lets explore why I decided my Atos required more light fixtures:
A) The low beam setting is dim. Not in the sense of 'Wow: that top-end race motorbike emits more light than my car'; no, it was in the sense of 'Man, I'm driving behind a scooter that sounds like a never-ending fart because its light lets me see the potholes 20m ahead.' Matter of fact, some older fellows have told me they can't even tell from the inside if the lights are on or off.
B) The 'high' beam setting legally is a normal low setting, as these are defined by Mexico's Federal Road Legislation: it allows me to see the road just over twenty meters ahead and doesn't bother drivers moving the opposite way.
So yeah, I do feel justified to add more lights to my car.
Now, let's draw some objectives for this project:
A) The new lights should improve my overall visibility; supplementing and eventually leaving my headlights as 'secondary lights' for those times I enter the somewhat more dangerous sectors of Cancun.
B) The new lights should not add further strain to my car's tiny battery, tiny engine and tiny alternator.
C) The new headlights should not draw criminal attention to themselves. This should be achieved through either making them invisible at a cursory glance or by making them obvious and ugly. Personally, I'd prefer not making my car ugly on purpose, so my primary focus should be stealthy design.
D) The lights need a separate control from the default headlights, thus deserving the name of fog lights. This keeps me from being caught with my pants down in case a police officer pulls me over and asks why I'm driving with my headlights off.
E) Whatever the light needs, I cannot modify the aerodynamic profile of the front end. This includes symmetrical work and not blocking enough of the front grid to make the radiator work more than it already does.
F) The arrays themselves have to be designed so they cool themselves on the air. Most people design LED arrays like they were light bulbs, forgetting that LED's are heat-sensitive microelectronics that have to stay below 85Celsius to work their best.
Step 2: Exploration
The attached photographs include their own explanations, but I do need to make caption on the drawing: the image editor doesn't seem to like adding comments to GIF images.
Anyway, from a top view, it explores how could I place LED modules made from either 5mm white LED's (the yellow cones) or 3mm white LED's (the blue, more concentrated cones). The 3mm ones emit a tad less light per unit than the others, but by being their emissions more focused, they have a better throwing distance.
By the way, before somebody asks, I am not using either blue or yellow LEDs.I'm using those colours in my drawing purely for graphic reference.Schemes A, B C and D are schemes exploring designs using a single kind of LED's each. I wouldn't use A or B because I would need to make extensive arrays to get a good throwing distance; I wouldn't use C because it leaves me with no near field; however, I became interested in D because, although it still has no near field, it gives me the best far field of all the designs, as long as I make the beams cross at a distance between 15 and 20 meters.
Scheme E seems ideal: Good near field, good far field. Still, I kept brainstorming.
Scheme F seems an ideal first phase for my project: it supplements my existing headlights without too many LED's.
Scheme G will be my ideal to achieve: on the second phase of my project, and once I have enough money for all the LED's and don't have my final exams hanging over my head anymore, I should equal the light emission of my headlights on high setting.
Step 3: Procurement for Phase One
By the way, I hope I don't, but I may forget to mention something as it might be too obvious for me. Comment on it and I'll either make a correction or say I didn't use that.
I pulled MXN$500.00 (around USD$45.00) from my budget and bought the following:
A) A hundred 3mm white LED's. Fifty were for the project; fifty were for my stockpile
B) Thirty-five 82ohm resistors (at a quarter watt)
C) Bakelite wafers.
I bought everything on STEREN. STEREN has presence in Mexico and the States, and you can shop online or locate the store nearest you through this website: http://www.steren.com
I have a bit of a stockpile on anything else I could need:
D) Around a dozen different types of glue (from hot glue to epoxy)
E) Around fifty 5mm white LED's
F) Assorted cables (different calibers of sound and microelectronics' cable). For best results, I prefer using thermocontractible insulation or glue-as-insulation ratter than electrical tape.
G) Clear acrylic paint. You might recognize it more easily as 'transparent graffiti'.
A) A Dremel with enough accessories
B) A hacksaw blade
C) All kinds of electrical pliers: precision-cutting pliers, needle-nose pliers, round-bending pliers, wire-stripping pliers, et cetera.
D) A soldering iron
E) A mirror, a flashlight and my cell-phone's camera. At times, those were my only tools for looking at the work area.
F) Basic safety equipment: security glasses, an air extractor, et cetera.
A) In my case, an acid to pull off the copper layer from the Bakelite wafers. Try and get uncoated Bakelite, unless you actually know how to etch it.
I used FeCl3 on a 15% aqueous solution starting at 40 Celsius and kept at 33 Celsius (the ambient temperature) and stirred every now and then. The copper was gone within a half hour.
Step 4: First Phase: Component Construction
The following images should all be properly captioned, but before getting there, I should tell that all my white LED's are 3.5 V, 20mA. Thus, I solder them in series of three with an 82 ohms resistor for a twelve volts power source (the car's battery). I don't do LED circuits without resistors because:
A) By getting the LED's over nominal voltage you make them change colour.
B) By getting the LED's over nominal current you make them overheat and rob them their extraordinarily long lives. A 20mA LED working at 50mA will be giving out slightly below double its normal light emission (the overheating makes it far less efficient), but it will only live between 100 and 500 hours (4 to 20 days), contrary to its theoretical mean life of 50,000 hours (5.7years).
C) You have to be out of your mind to try and get a no-resistance circuit connected to a no-resistance, very high current power source, such as a lead-acid battery. Let's put it this way: if I ran our of matches and my lighter ran out of gas, I would be able to turn on my gas stove using an LED, jumper cables and a motorcycle battery.
Something else not to forget: I didn't do all the component construction before opening up the hood for the next step: preparing the area. Before that, I had only made the 2 arrays of 24 5mm LED's.
Step 5: First Phase: Area Preparation
Simply put, I hacked, drilled, unscrewed and sanded, not necessarily on a given order. I also unglued the Hyundai emblem, sanded its area and painted it matte black again.
Step 6: Phase One: Physical Installation
Simply put, I installed the LED arrays on their resting places. To make things worse, I had to wait till nightfall before fixing most of the stuff in place.
Most of the stuff I said because there was just no-way I would attempt to install the free-in-the-air LEDs inside the Hyundai emblem after dark: I would already need to install them by touch, and I would be out of my mind if I gave up my option to use a mirror to guide myself.
I'm sorry not to have too many images on this step, but my hands were sticky most of the time and my mobile phone , and LG MG800c "Chocolate" pretty much has to stay clean to 'want to' work properly.
Step 7: Phase One: Test Drive
Simply put, I let my dying test battery rest in piece (connected to its charger), connected the arrays directly to the car's battery and drove to the nearest Oxxo (like a Mexican Seven-Eleven) for a soft drink. The complete array isn't yet too powerful, but it already shines enough for me to turn off the high beams. I'm glad with the results so far. I kept testing the thing for a couple of days.
Step 8: Closing Up Phase One
A) I made a button console somewhat inspired by the Batmobile from "Batman Returns".
If I was ashamed to be geeky, I wouldn't have started this project.Problem is, the console ended up with so many buttons that I would be ashamed to install it until I figure out what to do with all of them.
B) I decided to change my interior lights with a bunch of red LEDs I had around. Although they are different sizes, they are all identical as far as electricity goes: 2.1V, 20mA. I set them up in two series of five with an 82ohm resistor for each series, and only removed the original light bulb to replace it with them.
I like how my car's interior looks now, but my friends don't. I say it looks like the interior of a military aircraft (as seen on the movies), but my friends insist it looks like the entrance to a cheap whorehouse. To each their own, I say.
C) As I didn't install the button console, I have temporarily soldered the frontal LED array to feed on the marker lights. It was a simple operation of pulling out a headlight, peeling the correct cables, doing a quick solder joint, using some electrical tape for a temporary fix and bolting back in place the headlight assembly. Undoing it will take me like ten minutes once I have more LED modules ready to integrate, makes my array work even if it isn't completely finished and adding a half amp doesn't put undue pressure on the fuse or switch working the marker lights.
Future follow up on my project will turn up the speed and deal, on single stepswith all the design, construction, installation and testing of new modules being added to the array. I will only stop when I find quirks and details worth expanding.
Comments and criticism always appreciated.
Step 9: Minor Improvements
Life's been hectic lately: my finals, finishing up my party hall, my party hall crashing down spectacularly when the police raided my inauguration party and trashed down and looted the place (gotta hate corruption), the insurance company being sympathetic but also a very slow bureaucratic monster, spending eleven hours a day at school (my summer courses started), my car needing service, et cetera. Fortunately, I'll have nearly a whole day for myself on Friday, with the party hall being down right now.
About this Instructable:
1) I repaired my camera. That is a story by itself:
I went to my mobile phone's authorized service, and after a few days without it, they called me and told me that it had humidity damage and, thus, the warranty was void. I celebrated that the warranty was void, picked the phone up, disassembled it at home, saw that the problem was some lime deposits built up around the inside of the multi-connector (power/USB/hands-free), and using a sharp needle and five minutes I got it to work as good as new.
I wish I had taken pictures of all I did to the phone and then made an instructable of it. Thing is, it is ratter hard trying to take pictures of a camera when it is the camera that you are repairing.
2) I finally took those beam shots. They are posted here.
3) I detached the LEDs from the marker lights' circuit, and put a temporary switch for them on the instrument panel. If one of the prime objectives behind my project is to have optional lights that I can use when I want and can turn off when I cross a dodgy zone, I had to have direct control on them even this early on the project.
4) I put five more red LED's on the interior. I'll probably end up adding a whole bunch.
5) I experimented with 10mm acrylic sheet. I could cut it fine, but my polishing techniques need a lot more improvement. I'll probably need a real big buffing wheel or 'outsourcing' those pieces. Don't hold your breath: I'm currently exploring different avenues.
Step 10: LED Brake Lights
Simply speaking, a client canceled a job containing a bunch of economic red LEDs and now those will not go to waste.
Please follow the pictures. Those contain all the necessary information.
Step 11: How Not to Do LED Headlight Modules
This was supposed to be a step on how to do LED modules at top density. It turned into, simply speaking, a lesson on things not to do when designing LED headlight modules:
A) Leaving the resistors on the front of the design.
B) Not drowning your resistors in silicone or some similar kind of liquid-start chemical insulation. I used a thin insulation that had previously worked fine for flashlight design and that did work fine for a week and a few rains. Thing is, it didn't hold that well when I forded a flooded road: half of the lights began flickering before definitely dying a week later.
Ironically enough, the lights riding the lowest, those installed on the lower grill, have never hesitated.
Fortunately, these modules were made for my own use: I would hate my reputation to be tarnished if they had been returned from a client. I did waste a Sunday and more than USD$20.00 on these, though.
Step 12: Continuing on Phase Two...
That's about definition enough: I'm continuing this project.
The first image shows how things look right now: two upper 'edge' arrays, each with 21 5mm LED (pointed forward, 5 degrees down and 5 degrees outwards), a central array of 60 LEDs at top density (pointed straight forward, replacing the twelve LED array around the Hyundai symbol) and two lower 'edge' arrays, each of 24 5mm LEDs (pointed about ten degrees down, five degrees inwards). Not counting the eighteen LEDs that sit under the hole where the Hyundai symbol used to sit, the array is consuming 10.5 Watts at the LEDs, 1.6 Watts at the resistors and draws precisely 1 Amp (1000 mA) from the source.
As of July 1Âº, the last modules have been in place for a week and a half and it's been raining so hard that my car even looks like it had been washed: if these modules where about to fail, they would have already done so.
The Second image is a working image on the construction of the Upper 'edge' LED modules. They are mostly the same as in step four, except that I painted the boards before starting to work.
The Third Image shows my building of the center module.
The fourth image is a close-up on the capacitors I scavenged out of an old computer's power supply and connected to the central array. I used an scavenged bridge diode to isolate these caps and their array from the rest of the car and, of course, I had to tweak the resistor values for the central array.
The fifth image shows a couple of octagonal mirrors which I fitted into the hole for the central array. Not that useful, but they push out an extra little light.
The sixth image is the test of the array under the same conditions as always. I will need to reach the 300 mark before the light is good enough, but I'm glad that with only 102 actually pointed to the front, I do have decent illumination.
I have decided That I will need to have a third phase, composed of Luxeons or such High Intensity LEDs. Don't hold your breath, though: I first need to finish the second phase, them learn how to work Luxeons.