Putting LEDs In Motorcyle Tail Light Shows How Trivial Becomes Tough

[Maarten Tromp]’s replacement of his motorcycle’s tail light with LED equivalents is a great example of something that every hacker learns sooner or later: interfacing to and working around existing parts can turn a trivial-seeming task into a much bigger job than expected. The more one has to work within the constraints of an existing system, the more opportunities there are for roadblocks and surprise issues to stall progress, and this project is a great example of that.

[Maarten]’s 1999 Honda ST1100 Pan European motorcycle had no aftermarket options for an LED rear light assembly, and he wasn’t too keen on just installing a generic module to replace the original. Instead, he resolved to purchase and disassemble a used factory assembly, and replace the incandescent lamps with some equivalent LEDs. Replacing bulbs with LEDs sounds easy, but doing the job right took [Maarten] almost two weeks in the end.

Problems started early with simple things like how to open up the light assembly itself. The unit isn’t user-serviceable and isn’t intended to be opened, and the parts are sealed shut with a waxy substance. Fortunately, heat does the trick. Another early hitch was the curved base of the light assembly, which made mounting flat perfboard or veroboard a challenge. In the end, [Maarten] settled on a triangular grid of high-brightness LEDs,  driven with LM317 regulators configured as constant-current supplies, mounted on some protoboard cut to fit the unique curve of the assembly. The result accepts the wide voltage range of the motorcycle’s battery (from 10.5 V to 14.5 V) and can still function even if some individual LEDs stop working.

The project has one more example of how working around existing hardware can be a pain. [Maarten] had originally intended to swap out the turn signal lamps for LEDs as well, but there is a glitch. The motorcycle’s turn signal relay will do a fast blink pattern if burnt-out turn signal lamps are detected. Since LEDs consume considerably less current than the original bulbs, the relay will remain stuck in the fault condition. There are a few different ways around this, but it’s a problem for another day. For now, the tail light LED replacement is a success.

Working around existing hardware frequently brings unexpected challenges, but when safety systems (such as lights on a vehicle) are involved, it’s extra-important to make sure things are done right.

55 thoughts on “Putting LEDs In Motorcyle Tail Light Shows How Trivial Becomes Tough

  1. Appreciate the effort he put in, but honestly, I would have just gotten the aftermarket LED bulbs. If his concern is really reliability, this homebrew solution seems far more likely to fail in the middle of a rainy night than commercially manufactured LED bulbs that can be easily swapped out if need be.

    1. I’ve had a lot more drop in LED bulbs fail than I have custom tail light board assemblies. They tend to overdrive the LEDs and undersize the resistors and traces leading mostly to heat death

      1. I’ve experienced this, but the primary failure cause of most of my failed tail led attempts has been death by vibration (see:smallbore honda twin). The bases of every purchase so far of the <$8 shipped options that are drop in replacements are just dimple crimped with the resistor lead wedged between the mating walls. They are all heavier than regular filament bulb as well, which seems to speed up destruction. This is all with the stock vibration dampening rubber in place.

        1. Yeah, this has been my experience: I tried multiple times to make replacement trailer lights using LED’s, and the trailer vibrations just demolished my pcb’s: traces up to components tore off the board, vias tore out, parts cracked. Building vibration-rated pcb’s is _hard_. (Conformal coating helped a lot. But ugh.)

          By the way, the vintage car community has a lot of solutions for ballast loads to keep the blinkers working right, from replacing the blinker unit with one that doesn’t need a load (doesn’t use a bimetallic strip heating and deforming to perform the blinking functionality) to separate discrete ballast resistors to LED lights that are designed to draw enough current to make the bimetal type flashers work.

          1. 2 oz copper; big, fat traces; generous gussets and ‘potting in’ the backside and Vaseline-over-diffuser-painted-fronts with flat black hasn’t failed me yet for trailers

          2. Same failure here except it was in my SUV. 2006 Honda Pilot. Bought the tail light bulbs from local auto parts store and they both failed within a month.

      2. I’ve had the same led bulbs in the tail lights of my truck for around 7 years now. Putco plasma led bulbs, they were probably the brightest led bulb you could get back then. Couple years ago the one bulb came undone from the base and was hanging by the wires, I just super glued it back together. I have a cheap Amazon led bulb in the tail light of my Yamaha V-star 950, it’s a couple years old. I’ve seen more led bulbs fail from moisture getting into the bulb housing than anything.

    1. I’m not going to claim to know the answer to that. But if the correct light with the correct brightness and pattern are coming out of it then what’s going to happen? Are there police that go around pulling people over and disassembling their tail lights to see if they are stock or some illegal homebrew solution?

      1. Probably not, but if you’re in a wreck, at least in the US, some slimy lawyer will seize upon your non-street-legal rear lights as grounds to sue you for everything you own.

      2. Nobody will pull you over and inspect, but if you get rear ended and end up paralyzed for life, you will likely not have any financial recourse if it is discovered that you have a modified taillight. I would still do it though

  2. The quick and dirty fix for the fast flasher issue is to place a resistor in parallel with the LED ‘bulb’.
    Or replace the flasher relay with an electronic one

    1. A replacement flasher is just a NE555 and a MOSFET away.

      Been using one on my bike for close to 10 years now based on the 50% duty cycle example circuit in the National Semiconductor LM555 datasheet and a IRF540N MOSFET. This is on a bicycle, but there’s no reason why the same circuit couldn’t be adapted for a motorcycle.

      1. Now this is an example of a really good hack. Outstanding!


        [I’ll make no comment regarding the missed opportunity of using a Raspberry Pi–and four weeks of your life–to replace the (very) pedestrian 555 timer (which deign you probably completed in one hour, total). See? I didn’t say anything]

        1. I did look at using a microcontroller instead… started with a ATtiny85… threw in the necessary level-shifting logic to detect the state of the two indicator switches, output transistor driver, voltage regulator.

          The circuit wound up being no smaller, but the circuit was a lot more complex… and for what? A souped up blink sketch basically. The MCU didn’t add anything feature-wise, just made the circuit more complex and error-prone. Making the switches switch +5V instead would remove the need for level shifting on the inputs, but smoke would be released if that +5V rail ever contacted +12V by accident.

          Diode-OR and plain dumb electronics won out since the NE555 is perfectly fine running at 4.5-15V.

          Even SMD vs through-hole didn’t work, I made a SMD version, then realised that in the space where the SOIC-8 was, I could fit the original PDIP-8… in the space where the 0805 resistors, diodes and capacitors were, I could fit a through-hole resistor/diode/capacitor end-on. The PCB with through-hole components is basically ~20mm square.

          1. With a microcontroller you could have made a current measurement to detect broken LEDs, if you wanted. But if you don’t do this, it is not better than the NE555 solution.
            For level shifting of the switches a simple resistor is enough, togehter with the internal protection diodes of the µC. Or if you do not want to use them, a voltage divider, something like 47k and 22k, both in 0402 size would be enough.

      2. Issue generally is that the flasher is a sealed unit with a couple of prongs at weird angles to each other that plug into a socket on the wiring harness, so it’s a pain to build a replacement unit that fits the footprint. I’ve had luck with cutting open an existing flasher unit, gutting it, and putting a replacement in, but as above comment details, my experience with putting DIY electronics in high vibration automotive environments is that they last months at best and when they fail and I open them up there are just random broken pieces inside because the solder joints have failed, the traces tore off the pcb, it looks like someone took a power chisel to the board.

        1. So cut the proprietary connector off and solder in something more common… or hack apart the flasher to separate connector from circuit.

          I wouldn’t expect warranty to cover things like incandescent→LED replacements so why would the flasher be any different?

    2. I have this problem with my wife’s car. I replaced the blinker bulbs with LEDs a couple years ago now. I am aware of the parallel resistor solution but am too stubborn to do that. I replaced them in the first place because the fixtures were poorly designed and the bulbs were cooking the plastic. I didn’t remove a wasteful heat producing incandescent light bulb to replace it with a wasteful heat producing resistor!

      I thought I could just pop out the blinker module and replace it with one that is made for LEDs. I couldn’t find it though. A little research gave me conflicting information, some stating that there should be a module somewhere, others stating that for this model it is built into the car’s main computer.

      I have intended ever since to look a bit harder, then if there really is no way to replace the blinker or adjust the computer go ahead and resort to the resistor method. Maybe I can mount it to the metal body of the car, away from the plastic somewhere. But time keeps passing and with a couple years having driven it as-is with no consequences it’s kind of hard to get motivated to do anything.

    3. Electronic flashers already exist for LED systems. Easily found, at least in the US, at any auto parts store.

      There’s the old typical bi metallic ones like the bike has, and there’s what I think they refer to as an “electronic flasher”.

  3. A bit of spray-on conformal coating wouldn’t go a miss. You can buy it in aerosol cans.
    No matter how much you try to keep water out of something, it will find a way and corrode your PCB.
    Heat cycling will cause condensation and pressure build up if you try and completely seal the unit.

  4. I built some LED taillight assemblies for my 1980 scirocco with 5W LEDs underdriven for reliability. They worked great. Cycles are tougher though— much more jiggly than a car.

      1. That definition depends on the country. In Germany, such a aftermarket-DIY-modification of the lights would simply be illegal on public roads. One would have to reckon with a fine and shutdown of the vehicle in the event of an inspection.

        There is the possibility of allowing something like that – but the path is long, rocky and expensive.

        What surprises me: the article is apparently from the Netherlands. Is that thing really legal in the Netherlands (neighboring country of Germany and also part of the EU)? Or is it according to the motto “where there is no plaintiff, there is no judge”? ;-)

          1. I think that leaving the original bulb in and adding LEDs is the way to do it. At 70 MPH 200ms = 6 metres (I think, it is early). That’s over three times the length of my old BMW R75 and for sure I would rather someone stopped an inch behind my number plate than 4 metres in front of it.

          2. The last time I was rear-ended (stopped at a pedestrian crossing in broad daylight) it was by a BMW. In that case my brake lights had been on for quite some time, it was more down to driver observation.

            Luckily, they paid in full.

      2. Definition:

        At least in Germany (and I assume the entire EU) some vehicle parts have to have a type approval application from the KBA (§ 22a StVZO). The application has to include an expert’s report by a certain part depending inspection authority (§ 5 Absatz 1 Satz 2 Nr. 1 FzTV) . Those parts have an approval mark.

        It is possible to get an application for a self-made unique item (§ 21 StVZO).

        Without that or with modifications the vehicle looses the type approval and thus the liability insurance. This is a misdemeanor (Ok, I couldn’t care less) and a huge financial risk.

        Brake light have to have a type approval application.

    1. Why is that good? How is it the only or even best way to protect people from poorly done lights?

      The law could easily specify that any light must be the correct color, blink in the correct pattern and be sufficiently visible from the correct angles up to some specified distance and not blind other drivers without ever invoking type acceptance or any sort of anti-DIY law.

      If someone DIYs their bulbs it should be on them to meet those requirements. If they don’t then sure, consequences. But if they do then why do you think it is a good thing to punish them?

      Or are you thinking of reliability? Stock incandescent bulbs are still legal. Those are guaranteed to burn out eventually. Unless you check your turn signals every time before you drive away and you happen to get lucky, the filament breaks while parked it’s almost a sure thing that one day you will be driving with a burnt out bulb, at least for a moment until you notice it and do something about it. Why would someone whose DIY signals fail not also notice it and do something about it just as quickly?

      1. Stock lights do burn out but it takes years. I have a 23 year old Buick whose brake lights had just burnt out. The turn signals are still going strong after this time,. And when they do burn out I just go down to the local auto parts store and get a replacement that takes about 15 minutes to replace. Instead of spending 15 hours making a homebrew solution and that makes the car un-sellable when a buyer inspects the car and sees the nasty little mod that was done. Not to mention invalidates my insurance and leaves me wide open for damages.

        I can understand the hack if it’s a classic car where there are no original parts available. But that’s about it. Otherwise it’s not worth the effort.

        1. Experience will vary. Between my car and my wife’s I was typically replacing about 2 bulbs per year until I switched hers to LED. In the past I have had vehicles which seemed to just burn through bulbs every couple of months. I suspect poor voltage regulators.

          When I was fresh out of college and didn’t yet have a lot of money to spend on car repairs I had a stock tail light assembly go out. It was a poor design, the PCB was too exposed to the elements and tended to corrode. It was an old truck, not worth very much but I needed transportation. The money they wanted for a replacement assembly (same bad design) was almost half what the truck was worth! And I didn’t have it!

          I eventually ended up replacing part of the PCB with perf board and covered it all in epoxy so it wouldn’t happen again.

          I never even thought that it might not be legal to do that. I’m guessing that where I lived it was legal but I don’t know, never checked. Either way, nobody could tell the difference and the thing was more reliable, and thus safer then the stock parts ever could have been.

      2. It is about the devil you know vs the one you don’t. DIY quality has absolutely no limits – up or down. At least the incandescent bulbs have a documented behavior/life cycle. And, I don’t think it is about reliability alone. DIY might be too dim or too bright or not have proper dispersion angles or what ever. There is a lot of engineering that goes into those parts.

        Perhaps the average HaD reader is capable of thoughtful design and careful engineering across a number of disciplines. But, based on some of the yahoos I’ve seen modding their off-road gear leads me to believe that in the broader world thoughtful and careful are rare commodities. And, of course in the YouTube era, any clown with a camera becomes an instant expert. My favorite one is the guy that built a plasma cutter torch height controller with an arduino, no isolation and friction fit connections. what could possibly go wrong?

        1. Again, that can all be addressed by legislating that the light be the correct color, brightness, angle and blink the correct pattern so that it is visible, understandable and doesn’t blind people. Obviously no police officer is going to be carrying around laboratory calibrated equipment to measure such things but they will know when they see it if it doesn’t look right and might be a safety issue.

          No doubt such rules already do exist. They don’t let people drive around indefinitely with stock lights that have somehow gone out of spec do they?

          So why punish the DIYer?

      3. “Unless you check your turn signals every time before you drive away” . . .

        When I got a moped, that was exactly what I did. The electronics were a mess; it took over 10 years to find one fail-open resistor was preventing the choke from working and caused the headlight to flicker. Since we couldn’t find that, I just learned to always check everything.

        Kept it up when I got a motorcycle. Might only check every light once a month, but everything got checked regularly.

        Do cage drivers not do this?

        1. That’s great that you check that but it missed the point.

          Lights eventually do fail. You can check your turn signal before you pull out of the driveway but that still doesn’t guarantee you will never be on the road with a failed light. The filament can still burn out while you are on the road. If a police officer sees you they will pull you over. Where I live if the officer is nice you will get a verbal warning. But you had better have it fixed the next time they see you driving. I haven’t ever had one but I have heard of officers giving out “fix-it” tickets. It’s a ticket that goes away if you take your vehicle to the police station or courthouse or somewhere and show them that you have fixed the problem in a given time limit. Or, if they don’t like you they might just give you a regular ticket.

          My point is that brake and turn lights failing is a part of driving. It happens. And everywhere with roads and the rule of law must have some sort of system for dealing with this eventuality. So why punish the DIYer? That same system can deal with them if they are building faulty lights as deals with everyone else if they don’t maintain their stock ones.

  5. My experience is not with motorcycles but with LSA, where I have been making avionics, sensors, transducers, and the associated signal conditioning circuits, mostly for use by myself and fellow aviators. While impulse shock loading for land vehicles is more than for small aircraft, thermal and vibration requirements for aircraft are more onerous.

    MilStd810 and DO-160 an MilStd704 have been my practical guidance for both design and test of stuff for LSAs. I would suggest that automotive electronics hobbyists download and study, at a minimum, MS810.

  6. Or you could just buy an LED bulb. Of course then you have to possibly deal with Hyperflash.

    Then you buy resistors and wire them in. Or if you are like me and don’t do splicing, then you buy sockets with pre-wired resistors that are way too big to fit into the space available behind the taillight assembly. So, you buy bulbs that are fairly expensive and put those in in hope that you will NEVER have to replace them.

    Finally, they work without Hyperflash. Of course if you have to sit through more than one signal cycle with your blinker on, they heat up and Hyperflash.

    I have contended for DECADES that there is a special ring in Hell in which automotive engineers are condemned to repair the vehicles they designed for ALL ETERNITY.

  7. To answer the most frequently asked questions:

    This mod is not street legal in the Netherlands. Everything on your bike (or car) should be type approved (DOT). However, the modified tail light surpasses stock in every way (brightness, opening angle, response time) so I’m still confident it’s a good idea. And if I ever want to switch back to stock, I still have the original all-bulb unit and can easily swap them.

    After 4 months, about 1000km and a lot of speed bumps the mod is still going strong. There is no condensation visible on he lens or any other sign of trouble. Hopefully version 2 will be finished before any problems come up.

    Regarding version 2, that one will have a proper PCB. Also it turned out I’m overdrving the leds (datasheet does not specify maximum current), which will also be addressed in the next version. So hold tight (but don’t hold your breath) for the even better version ;) If you happen to own the same bike, send me a message and I’ll hook you up with a PCB.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.