Forums » General Electronics Discussion

Hi everybody, I'm new in the world of electronics and I just discover this site. It's been a few month since I had this silly idea of using arduino to control the light of my bike. The goal is to learn electronics of course, a basic light will do the job just fine. So the basic idea was like the steady light on some new lights, who allow to keep the light on while the biker stop, to use the dynamo to charge up some super capacitor so the µC stay active. Then adding some capacitive sensor to control the lights trough the µC. I'm thinking maybe brake light, turn lights; ON/OFF main light. Every thing will be fitted inside the handle bar and nearly invisible.

Ok, So my first idea was like that https://dl.dropbox.com/u/5410763/Sketch_schem.jpg

Then I manage to try step by step, this circuit: https://dl.dropbox.com/u/5410763/Sketch_Step2-03.jpg This time, I tried with a chip Attiny85. It's probably obvious that this circuit is not very effective at all. Well, I'm learning... :p

So, now I'm thinking about a new circuit. I'm back with the basic and there is my architecture : https://www.circuitlab.com/circuit/k9fjjq/architecture-circuit-led/

I'm open to any idea and comment. I'm willing to learn but alone is really difficult to find my way.

In advance, Thank you all for reading me. By the way, I'm speaking french so I hope I was clear enough...

Hi, I don't want to go with batteries because they need quite a long time to charge and I just need power for about 5min max, when I stop at a traffic light. So I avoid the problem of incomplete charge/discharge cycle,

for the µC, I want it because I want to try to use capacitive sensors, hidden under the handle tape. There is a library for that within arduino. So I can turn on a blinking led when I turn for example, just by touching somewhere on the handle bar. Then I read about another design where they use an accelerometer to turn on a brake light when they slow down rapidly. It is probably possible to do that without µC but since I know how to program with arduino, I wanted to try with. (powering everything is complicated enough for me :p ) I was thinking using a µC could save some space on the pcb board. And when after powering this circuit, I can simply reprogram the chip for any functionality I choose. Well, ok, it's not that simple but let say the number of leds will stay constant so the first step is powering the system with everything turned on and put a manual switch instead of a µC

"energy harvesting from low power sources" cool, good idea. I just buy a thermo-ellectric cell, I wanted to try something with it. (not with the bike)

I found the LTC31** yes, it seems good but it is a really small package. I will search at other vendors...

So, if I may, can we go back to the sketch and my question now is: - the rectifying part, It's ok, I get it. then - how can I split the juice? :) If I put 2 capacitors for V backup, I can step up the voltage of them to 5,4V but then, do I have to stabilize the voltage at 5,4V for V direct? probably. So can I put the switching regulator at 5,4V before both? Someone told me to focus about current control. So I probably need something to regulate how much current is going to the superCap bank and leave as much as possible current for V direct.

Ok, I still have to read about harvesting from low power sources and how to control current. But I tried something like that. https://www.circuitlab.com/circuit/k9fjjq/architecture-circuit-led/ Let say U1 is a switching regulator 5V.
I'm using diodes schottky to reduce voltage drop. Then, how can I invert the transistor? I know, like that, both V direct and Backup power the system.

Sorry, I write a lot but I'm happy to have feed back from someone who know about that. Thank you.

"Someone told me to focus about current control."

That may not mean:

"... I probably need something to regulate how much current is going to the superCap bank and leave as much as possible current for V direct."

I think what your someone meant was;

(i) use a current mode SMPS rather than a voltage mode and

(ii) use a constant current SMPS to drive the LEDs.

That's why I suggested you read up about driving LEDs from the chip vendors. TI, National Semiconductors (now part of TI), LT and Diodes Inc all have some good articles and apps notes about that.

Think about what you are trying to do in terms of a highly variable low power source (dynamo) supplying a rectifier and a boost (step-up) SMPS which then charges a battery (battery and/or supercaps) to supply regulated power via an SMPS buck (step-down) regulator for the uC, sensors etc., and a third SMPS to drive the LEDs.

You probably don't need to think about switching between battery/supercap/direct. If you design the power train properly, it will take care of that for you.

BTW: you probably will have to use batteries because a supercap is not really suitable for long term high current demands like LEDs.

https://en.wikipedia.org/wiki/Electric_double-layer_capacitor

Why do you care about battery charging time? Once the batteries are charged to start with then if you design your system correctly (match allowable discharge to available charge!) then the batteries should never discharge enough to need long charging times. If you really do need more power than your dynamo + supercap can deliver because you are stuck out in the dark in heavy traffic so you are standing still or moving slowly for a long time then the batteries keep the lights on (a safety requirement) and so what? you can charge the batteries when you get home because you are not under a lorry on a dark road.

A minor inconvenience can save a major one.

:)