In-depth circuitry for LED taillight from scratch

First I have to commend you for attempting to construct something you want but cannot find. Those two grey cylinders are indeed diodes. The stripe is the negative side. An interesting read on wikipedia explains some of the functions of diodes, to limit current to one direction only and to provide a voltage reference.
Here the diodes likely isolate the brake circuit from the running light circuit, so that the brake lights do not backfeed the running lights elsewhere. Without those diodes, when you engage the brakes all the other running lights would turn on too.
Measured resistance of the diodes would only be accurate with them removed from the circuit, not attached to the LED load. Note that multiple LEDs will draw more power and produce more heat than single LED, and might not be any brighter than incandescent. If you find a lux meter you can determine brightness of a typical 1157 versus your LED version. Some cheap LEDs are actually not brighter than incandescent.
My question to you is: why construct an 1157 bulb when reliable bulbs exist at low cost that also meet DOT requirements for brightness? See the Sylvania 1157 for $13 each that meets DOT specs and draw only 2 watts. Some regions require vehicle inspections and homemade lighting could be a problem.

Thanks for the reply! The answer to your question is that my design is a bit extreme. It will use low profile red LEDs arranged serially in a grid-like layout, and then covered with a thin layer of light-diffusing red acrylic normally used in backlit signage. The intended effect will be like a monitor/screen backlight, but functioning at the brightness of a taillight/brakelight... a very thin panel of totally even, bright, diffuse lighting with no hotspots. Glowing. The entire assembly should be very thin--- probably 3/16" thick, and maybe 4x6" l x w. It will sit on an aluminum license plate holder that the PO replaced the OEM fender with. In the pic below, I'm going to remove the existing taillight and the exposed aluminum will be covered by this very thin design, which also will have the red acrylic diffusion layer wrap around the sides of said aluminum. I hope this explains it adequately, but I may just have to revisit this with images when I'm done. I have a CNC machine and have used them professionally for years, so having this all fit precisely and be very thin is more of a design challenge than a fabrication one. I think the idea is that it will look like an uncanny, impossible object. (cont'd below)



Luckily, my state doesn't require inspections for my bike and I've never been bothered by police despite several obvious modifications and (sometimes) extreme noise. And if I do get in trouble I'll just bolt back on the taillight in the above pic.

I still don't understand how the taillight in the above pic limits the brightness for the running light-only circuit, if it's not done through the diodes. The running light circuit AND taillight circuit both run through the same resistors. Regardless, it sounds like what I'd need to do for my design with 12v serial LEDs is have a 1N4007 diode + resistor (ohms TBD) feed the half-bright running light circuit, and then the brake light circuit bypasses that diode/resistor to go directly through the LEDs, and of course it has its own diode to prevent backflow. Then the LEDs have a shared ground wire going out. Does this sound right? Will I need some kind of voltage regulator or voltage stepdown to keep it at 12v exactly, or should I be fine without?

Thanks again for the input.

For anyone interested, I've finished my taillight/mini fender design and most of the materials have arrived. As I said above, the PO replaced the OEM fender with a bent piece of sheet aluminum with a cheap chinese light and the license plate bolted on. It was always meant to be temporary but I've grown to like how minimal it is, so I've decided to try to design a "nice" version of this concept.

Here is the design:






For clarity, the black L-shaped piece is the hoop at the rear of the motorcycle frame that the rearmost seat rubber rests on. The construction of the fender/bracket is from 1/8" 6061 Aluminum sheet, and the light lens is 3/16" red acrylic. I am making a dense LED array that sits right behind the red acrylic, made from 5730 red LEDs that are EXTREMELY bright at full brightness. The red acrylic is mostly opaque but slightly translucent, and the idea is that when backlit by this LED array it will glow with this very uncanny, uniform light that will look really nice. I have some photos where I was testing that effect further down in this post. I think these 3D models of this design make everything here look deceptively thick and large, but this lens face is only about 3"x3", and when it's attached to the sheet aluminum body, the whole sandwiched assembly is only about 3/8" thick. So I'm hoping that the design looks somewhat like an impossible object--- too thin to contain electronics, let alone a very bright, even light.

To fit the LED array into this, the aluminum is pocketed about 1/16" deep to house a circuit board with the LEDs. Here's an image of the design with the red acrylic lens removed to show the recessed LED array:



Wiring is difficult to fit in all this, so I took a page out of honda's book and designed this array so the whole circuit grounds directly into the aluminum sheet metal. To make this work, the entire assembly is rubber mounted to the motorcycle frame, which insulates it not only electrically from the motorcycle but also from vibration and shaking, which was another goal I had here. The fender/bracket I'm replacing is pretty rigid but it does shake at higher RPM and I don't want to destroy the delicate electronics here. The only wire in here is the positive part of the circuit, and it is concealed inside the metal which is difficult to explain but I'll just have to show you when I'm done. The Aluminum also acts as a heat sink for the LED array, which will be glued onto it with special thermally-conductive silicone. The array will produce a lot of heat and the aluminum well suited to conduct all that heat away.

Here's a view of the rubber mounts, low profile bolts, and a little weatherproof box to contain the electronics for the LED array.




The little power box for the array uses a buck converter to output a regulated, even 12v. I've found that LEDs have a way of flickering/strobing at low RPMs, and the capacitors in this converter should give these LEDs a smoother, consistent light. It also regulates the output, which is adjustable with a little on-board potentiometer, so the light won't "swell" with brightness as the RPMs increase. I can also use the potentiometer to set the brightness to an appropriate level. The brake light and taillight switch circuits now trigger MOSFET transistors that switch the light between a dimmer running light and a full-brightness brake light. Hard to explain, but this simplified the circuit a bit.

Circuit diagram:


They're pretty small, so the power box I'm making should be easy to conceal.

Yes, that is possible, although I'm banking on a few different things to minimize that. The 5630 LEDs I'm using are EXTREMELY bright and they are the ones that you often find in LED headlights. I created a test series of 6 of them and at 12v they are blindingly bright and pull modest current. To create the uniform lighting effect in this taillight, the LED array I'm making will be using 96 of them in 16 series of 6. At 12v, this light can be set up to be too bright for anyone to miss--- in fact I am expecting that I will have to set the buck converter to feed the taillight something much more dialed back, like 11v or 10v to get an appropriate brightness, but I have some headroom if I need more. Also, I mocked up the design in cardboard and the fitment has some of the light resting in the shade of the seat, so at least part of it will be shielded from the sun.

The acrylic will also be media blasted to give it a matte finish, so hopefully the glare will not be too extreme. I'm wrapping up the design and about the start toolpaths for CNC milling:

Made the PCB for the LED array tonight. It's only 15 thousands of an inch thick but still quite strong because the base is fiberglass. Cool how transparent it is at this thickness though. Also made a jig out of black acrylic that will align the LEDs into their intended placements with little LED-sized pockets and hold them in place so I can solder them without them sliding around. There is another shallower pocket that will hold the PCB--- it's visible in the pics if you look closely. I didn't notice that the CNC mill didn't cut the jig pockets all the way through the material like it should've, so most of the pockets have an acrylic "skin" I will have to cut out manually. Tomorrow I'm going to start on the toolpaths and milling for the red lens. With all my efforts to miniaturize the components in this design, I've managed to get the overall thickness of the aluminum-LED-acrylic sandwich down to just over 1/4"--- so it's very small! Will keep adding pics as I go.

So I have a few updates on this process. I soldered up the LED array seen above and milled the lens:



It looks pretty good but in person there are still "hotspots" of light created by the LEDs, so I went back to the drawing board and designed a denser LED array. This was also pretty much my first time soldering, so I made some critical mistakes. For one, I dwelled too long with the iron and the PCB material got really hot and warped. When I tried to flatten it while giving it heat from a heat gun, the LEDs started cracking. Additionally when I was testing the first array I stress tested it at full brightness for a few minutes and it started to smoke. My laser thermometer clocked it at 400 F, so I won't be running it at full brightness. I tried running it at full brightness again while pressed flat against some 6061 Aluminum sheet, which is basically how it will be in the final light design, and interesting the aluminum acted as an amazing heat sink and the temperature of the light didn't get over 120 F while against the aluminum.

Here is the denser array and the LED mounting jig I created:


The denser array is slightly larger so I also had to remake the red acrylic lens. I decided to mill a few different options, the main differences being the subtle amount of curvature on their faces.





And here they are with the machining marks sanded off and brushed with some scotchbrite:



I am going to test some different finishes including media blasting, brushed satin, and a high gloss.

Next I made the electronics box for the DC-DC buck converter and the transistors that will switch between brightnesses for the taillight and brake light. I even managed to get a decent fly cut on the black Delrin, which is sometimes a bit finicky with machining. The lid has a rubber gasket that will weatherproof it. I still need to figure out how the wires will enter and exit the box.







Then I machined the aluminum body for the light/bracket:



The ring around the edge of the pocket has a milled texture because that's where I will be gluing the red acrylic lens and I wanted as much roughage and surface area as possible for adhesion.

Bezel for the lens, also textured for gluing:




Test fitting the LED array:




And testing the fitment of the bezel and lens:

Following up the post above, a final mockup of the whole thing. The slim piece at the top there is a brace that will go on the underside and conceal a wire that will power the LED array. It will also give the milled pocket some strength, which is important because the material at the bottom of the pocket is pretty thin.




I anticipated that the thin material at the bottom of the pocketed recess holding the LED panel would be a bit of an issue when I put the aluminum in a bending brake to make two bends, and I was right. The bends happen very close to the pocketed recess, and I worried that the very thin material at the pocket floor would want to bend instead of the intended bending zones. So I tried to put as much distance between the bends and the pocket as the design would allow, but it still wasn't enough:





The sharpie lines are where the brake should've made the bends, but as you can probably see, the bends actually happened well outside those lines. On the reverse side, the bend is happening right where the LED pocket begins, and it unfortunately deformed the pocket way too much for this blank to be used. The machined tolerances in the pocket are extremely tight to fit the LED and bezel, and now they're completely out of wack. So I'm currently redesigning the aluminum body and the milling process... what I'll do is mill everything except the pocket, make the bends, and then put it back on the CNC to mill the pocket to the tolerances I need.

Thank you both! Always happy to share the process.

I made a few big mistakes when I began this project. First, I bought the cheapest hand-lever-powered bending brake I could find on harbor freight, and the other is deciding to use 6061 T6 Aluminum, which I've since learned ranks among the worst metals to try and bend. The harbor freight brake is made out of 1/4" angle iron and when I try to bend this 1/8" 6061 in it, I swear the 6061 bends the brake more than the other way around. It's insane how strong this aluminum is, but it mills beautifully.

Tonight I milled a replacement blank with some tweaked toolpaths correcting some earlier errors. I can also show how some of the wire concealment works in this design:








So as I mentioned above, this doesn't have the LED pocket yet. I practiced making bends in some scrap 6061 and I feel confident bending this blank the way I need, and then I have a jig that will let me put the whole bent assembly back on the CNC to make the pocket. I think the main issue was simply that the pocket was way too thin to stand up to being in the brake. In the last 3 pics above, the little brace/raceway I made will conceal the single wire in this design and also lend a bit of strength to the milled pocket so it doesn't flex when I hit speedbumps at 40mph. Though now that I've tried bending this aluminum in the brake, I'm convinced it doesn't need the extra strength. All the areas in this design with the milled knurling texture will be glued with a two-component methyl methacrylate glue, which has a similar strength to epoxy but much more flexibility and better adhesion to plastics. \

A few other design notes:

I mentioned earlier in this thread, but the negative half of the LED circuit runs through the aluminum itself, so I only need a wire for the positive side of the circuit. I will actually be using two positive wires for redundancy, because this whole design is so compact and all the parts are so tiny that the wires and solder joints are very fine and I'd rather be safe than sorry. I also don't trust my soldering... To weatherproof the electronics and give everything a bit more strength, I'm gonna fill the LED pocket with optically clear potting compound once I have everything assembled. The potting compound will help manage the heat, but I also used a thermal silicone cement to attach the LEDs to the PCB, and then the PCB will be mounted to the aluminum using a thermal epoxy. The aluminum should act like a heat sink to keep the LEDs cool.

Once I've finished/sanded/filed away all the marks on the aluminum, I'll have it type III clear anodized, which should make it tough as nails. I still haven't figured out what kind of finish to give the red lens, but I will probably media blast it and then hit it with a matte 2K clear coat. I have a feeling that it will look really good with a kind of even sandblasted texture.

It's coming together!

Today I finally figured out how to make my flimsy harbor freight brake bend the 6061 T6 the way I wanted, and it came out GREAT. The brake came with a thick metal bar that you clamp onto the workpiece to secure it and form the "setback", but I found that it was causing problems for a few reasons: first, it was thick but not thick enough to avoid flexing, which was causing the workpiece to rise up off the brake and made the bend unpredictable. Hard to describe without images of it. Secondly, this metal bar was digging into the aluminum piece when I clamped it, creating big ugly marring marks in the aluminum which didn't work for me. So I planed a piece of hardwood and filed a large radius onto it, and used it instead of the metal bar. It was much stiffer and didn't leave ANY marks in the aluminum, which was great. I was thrilled because I had expected warping and deformation of the metal when it got bent, especially visible along the edges where there is very small chamfering/bevelling, but the metal remained LASER straight and flat on the unbent regions and at the bent edges.

As I described above, I made the jigs that allowed me to mount my bent aluminum blank back onto the CNC to do a few other cuts:



The final result:



I don't mean to toot my own horn, but the small details came out phenomenal. I'm extremely pleased with how it materialized. In the area that holds the license plate, I milled a recess that's about a millimeter deep, which is also the thickness of the metal that my license plate is. When it's mounted, it makes the license plate lay flush with the aluminum and gives the visual impression that the plate is much thinner than it is. It creates something of an optical illusion in person in that way which is pleasant to look at.

Some various detail shots:












I have a set of fine jewelers files that I'll be taking to this piece to sort out some remaining burred edges, nicks, scratches, and other necessary details.