Na FreeHostingu Endora běžà desÃtky tisÃc webů. PÅ™idejte se jeÅ¡tÄ› dnes!
VytvoÅ™it web zdarmaNa FreeHostingu Endora běžà desÃtky tisÃc webů. PÅ™idejte se jeÅ¡tÄ› dnes!
Vytvořit web zdarma
After some experience with a trike I can say the wider a bike looks, the more space it gets from overtaking cars and the fewer dangerous situations occur. So I tried to put two taillights on a two-wheeler. And to make it less boring, I added blinking daytime running lights. Despite trying to simplify the circuit as much as possible, my record in amount of wires on a bike was broken again.
XL-FDD 3 W dynohub in 26" rear wheel and two Li-ion cells in series. A little boring, but proven. The only novelty is added filtering capacitor which allows blinkers to stably run on dynamo alone, and maybe it also helps the batteries a bit.
Headlight is placed at the very front to avoid illuminating anything in my field of view. Inside, there are two high-power LEDs for high beams and eight 5 mm ones for running light. High beams have two levels of intensity, running light can shine continuously or blink. At the rear end there are 12 red LEDs on a 40 cm wide bar, in two groups as far apart as possible. These can shine or blink, have three levels of intensity, and automatic dimming for indicators. The indicators are two on the rear bar and two on handlebars, each made of three yellow LEDs. Brake light consists of six reds. To make the bike visible from the sides, there are two yellow lights (2+2 LEDs) near the wheel hubs. No more circus effects - the frame must stay uncluttered to attach cargo and to provide lifting points, and the tailbox floor is consumable (unless I learn to never crash, it will get destroyed eventually). One red LED illuminates the speedometer.
Total consumption: daytime running lights in continuous mode 150 mA, about half that in blinking mode. Night lights (continuous running lights + level 1 high beams + side yellows + speedo) 320 mA, full night lights (level 2 high beams + the same) 430 mA. The dynamo produces about 390 mA at 20 km/h and 410 mA at 25 km/h. It looks like my power balance will be slightly negative at night, but significantly positive during day, so it should work.
Subcircuits in dashed frames will be described below. All voltages and currents mentioned below are actual measured values. LED datasheets usually indicate nominal current of 20 mA and some nominal voltage that doesn't match it.
Generated AC current is rectified by a diode bridge D1, switch S1 disconnects it. The diodes are soldered together directly, buried in hot glue and hung on rear fork next to the generator. DC power is stored in filtering capacitor C1, voltage is regulated by two shunt limiters VL (see below).
Two-pole switch S2 disconnects the battery, only the common ground remains connected. And also the voltmeter behind S3 button switch. Normally both S1 and S2 are on while riding, current flows from generator to main power bus, from the bus to lights, and the battery absorbs excesses or covers shortages. If needed, only one of the sources can be used.
Switch S4 powers step-down converter for USB output.
Two normally closed buttons S5 are glued to brake levers and close when the lever is pulled and releases the button. The current powers brake light (BR) directly. S6 powers a set of night lights (headlight, yellow side lights, dashboard light).
Running lights and indicators are interconnected:
Three-position switch S11 controls running lights: left position blinking, right steady, middle off. Front white light in 2×2S2P configuration draws the same current as taillights and they blink alternately, so their combined consumption is more or less constant. When blinker (BL at lower left) outputs high, first T1 transistor opens, front light shines, second T1 is closed and taillights are off. When the blinker outputs low, first T1 closes, front light turns off and voltage gets to second T1's base, opening it and turning taillights on. So much for theory. In practice, base current of the second transistor is enough make the front light shine: very weakly, but visibly. So I declared it intentional and added a second resistor between taillights and ground to make it the same on both ends.
Three-position switch S10 sets taillight intensity. Right position is maximum for daylight, fog or against setting sun. Middle position is minimum when I don't want to glare people behind me. And left position is middle mode for normal nocturnal operation among cars.
Indicators are controlled by two-position switches S8 and S9. In right position an indicator is off and corresponding taillight gets full power. In left position blinker circuit (BL at middle left) and given indicator powers on and resistor R1 is engaged, so that corresponding taillight dims to make the indicator visible. The switches are independent, both indicators can be on if needed. The combination of indicators with blinking running lights looks chaotic and unreadable (two different frequencies), so it shouldn't be used. But there's no need: I indicate turns by hands during day, and run steady taillights at night.
I also considered processor control. It would make possible to replace multipole switches with single pole, automatically change running lights to steady mode when indicators are on, control intensity with lossless PWM or maybe switch something based on riding speed. But it would be more expensive, would require programming in addition to soldering, and the processor would consume additional power. Would it simplify the circuit? With one central processor, there would be one less wire for rear lights, but otherwise next to nothing. With multiple serially communicating processors, only a three-wire bus would have to run across the bike, but the mass of added electric boxes would probably even up the mass of saved wires. So I stayed with an analog solution.
I used two-position buttons for indicator switches (first press on, second press off) to avoid levers sticking out below handlebars, but it was not a good idea after all: the small button hidden under a little roof is hard to find by touch, especially with gloves. And it's hard to check if pressing it had any effect. As soon as I run out of patience, I'll replace them with levers.
Another problem that was not obvious during the build was placing the three main switches behind the shifting grip. Normally OK, but with gloves it's easy to flip them while shifting. Gloves also make it hard to check their position by touch. If I unknowingly switched of my taillights this way, I'd only know it after someone runs me over from behind. So I'll have to add running light indicator, illuminate the switches for visual checking, or both.
There are two blinker circuits, so multiply the part list by two. Indicator blinker is only used briefly, so it can get away with cheaper SA555 timer consuming something over 10 mA. Running light blinker works almost all the time and deserves a better TLC555 timer drawing less than 5 mA. Resistors R1 and R2 set the ratio between on and off duration (this combination yields about 1:1) and C2 capacitor sets blinking frequency. C1 is recommended for stability, maybe it could be omitted. Supply voltage can be anything between 4.5 and 16 V and it doesn't affect blinking frequency, so the 5 V regulator I used last time was useless and I omitted it now. Output resistor R3 is necessary to limit base current for following transistor.
Each end of rear light bar carries nine LEDs: middle three are indicators, the rest are running lights. If they lighted up at the same time, they would be impossible to tell apart, hence the crazy automatic dimming. Spatial separation would solve it too, but it would need two of the bars.
Having a drawer full of old LEDs, I didn't want to buy new ones. But there were not enough of each colour, so I had to combine leftovers: orange to the front and yellow to the rear, for better contrast with white and red. Resistors are recycled from scrap eletronics, all nominally 100 Ω, but some were actually 2 Ω lower. Rear indicators are soldered on common boards with the running lights, front indicators are crow's nests in hot glue blobs.
Designed purely to be seen, or for walking the bike. To illuminate the road while riding, at least 30 of these diodes would be needed, and that wouldn't fit my fork. The running light is crammed on common board with main headlights, resistors hidden on backside (not that I didn't want to show them, they just wouldn't fit in there). The housing is made of a plastic cup and a piece of plexiglass.
If I had enough LEDs of the same type, this would be the same as one of the taillights. Instead, I combined two types which needed two different resistors. The LEDs form a vertical strip on a strut between light bar and bottom parking feet. At first I planned to put the bottommost diode a bit lower to make it look like an exclamation mark, but simplicity won.
S6 switch turns on three parallel circuits:
LED1 - auxiliary yellow lights near wheel hubs to make me more visible from the sides, because front and rear lights don't shine sideways. Cars coming from a side should illuminate my reflectors, but there are also other things out there and not all of them have headlights.
LED2 - small red light illuminating the speedo. This time without adjustable brightness - in practice, I only adjusted it once and have never touched it since. Only one LED is needed for this, but I wanted the same 6 V of threshold voltage as everywhere else, so I added in series 8 unidentified old diodes from the bottom of my drawer. They ate about 4 volts and a resistor limited the current to 3..4 mA, which is just about right for this application. A 4 V Zener diode would do the same job.
LED3 - long range headlight to illuminate the road ahead. It is on all the time at night, so it needs a low power mode to avoid blinding others, but still just enough to see by: S7 switch off, both R3 resistors engaged, 130 mA of current. When there's no one to blind, S7 is switched on, one resistor is bypassed and current rises to 240 mA. The LEDs are rated for even more, but the generator wouldn't be able to feed them indefinitely. Each R3 is actually four 10 Ω resistors connected in 2S2P, so they still have 10 Ω, but quarter of heat to dissipate. The resistors are not placed inside the headlight, they help heat the battery compartment instead. The LEDs are cooled by an aluminium heat sink which holds the whole circuit board in place.
Two 18650 cells from a discarded e-bike battery, connected in series. Capacity somewhere around 1500 mAh. I learned that I don't need more - charging takes whole day anyway, and in the evening I milk them and make room for the next day. Now it may be different, flashing lights consume less and charging will be quicker. Time will tell.
The cells hang on handlebar tiller by zip ties, inside the main circuit box heated by waste heat from voltage limiters and headlight resistors - theoretically, this might extend their usability in cold weather, but the heating is probably too weak to matter. I didn't include charging switch or service connector this time, they proved to be useless.
Voltmeter is digital again, but this time I managed to find a bare module without housing, which could be embedded in the transparent sidewall of the box. No screws needed, it's held in place by a mess of wires compressed behind it.
The same circuit again. I set threshold voltage to 4.15 V. Indicator LEDs are hidden in transparent sidewall of main box and shine rearward, I can see them after tilting my head to the left. Power transistors are bolted on a common heat sink (insulated by silicone spacers because their conductive backsides are on different potentials). I can touch the bolts on my trike while riding; their temperature doesn't feel too high, so I didn't make the heatsink too big this time. If I'm wrong and the transistors burn out, I'll cut the heatsink in two and bolt it on without spacers to make it suck the heat faster.
Mostly the same as last time, but I omitted the surge filtering coil on the input - the circuit is not switched on frequently, so I hope the high currents caused by battery connected to empty capacitor will not cause much damage.
Lesson learned last time: don't mount switches on the seat. The place receives much water during rain, and gets in the way when the bike is carried or falls. So I now put everything on the handlebars, except S1, S2 and S4 which are on the main box, unreachable while riding. Time will tell if the seals are enough.
This bike's handlebars are troublesome. They can't be anywhere else than they are now, so at night, when I need to look thirty metres ahead instead of three hundred, they are exactly in my line of sight. I'll probably get used to it, but it was quite a nasty surprise.
Another special feature is that the handlebars are used to counteract pedaling torque from the bottom bracket on front fork. The forces are not large, they can be easily managed by one hand, but they interfere with delicate switch manipulation. That probably can't be helped, having to reach for some fixed part of the frame would be even worse. I'd recommend to mount the switches further apart and avoid the sensitive three-position ones.
Too early for that, it will need several hundred kilometres of testing.