How I Built My Own MTB Light

[92se-r]

You can't step up current to meet the needs of your LED. Only voltage.

I believe most LED drivers use PWM(pulse width modulation) to control dimming. Most do not have adjustable current.

You don't use buck drivers for LED's. You do not have to step down your battery voltage for LED's. LED's are like a faucet. At a certain voltage, they turn on. That's it. The LED voltage is fixed and not determined by the battery.

Linear drivers do not mean that they use PWM. Linear dc-dc converters are just a different topology than a switcher. They usually have more power loss through the converter.

PWM is Pulse Width Modulation. You do not short out the LED to do this. PWM basically outputs a square wave, where depending on the PWM duty cycle, there is a ratio between on time and off time. So if you want to dim the LED, you adjust the PWM down so that the pulse width of the output of the driver is less than the time that the output is off. Basically the driver turns on and off really fast and the pulse width(how long the output is high) is determined by what you set the PWM at.

When you choose a LED driver, you need to look at the minimum voltage headroom it requires. Add up your LED voltages, then add your minimum voltage headroom for the driver, and you want to get a battery voltage that is higher than this. Ideally, you want to be closer to the LED voltage + headroom because this means less heat dissipated through the driver. Not really an issue on a bike light with the readily available batteries out there.

Originally Posted by Flat Broke

The drivers are either boost, buck, or linear. Boost drivers take the input voltage from your batteries and step up the current to meet the needs of the LED. If you had a really power hungry emitter or series of emitters that draws more voltage or current than your batteries have in their natural state, you'd go this route. The different brightness levels are achieved by the driver increasing/decreasing the current (amperage) to the LED.

Buck drivers effecitively lower the input voltage from your batteries to ensure that you don't burn out the LED. So if you had a couple LEDs that only drew a few volts, but ran an 18 volt battery pack, you'd want to go this route. The different brightness levels are achieved by the driver increasing/decreasing the current (amperage) to the LED

Linear drivers don't regulate the current like boost/buck type drivers. They dim the light source through PWM dimming or Pulse Width Modulation. As I understand it, the current going to the LED is constant, but the PWM circuitry effectively shorts the LED out an different intervals to reduce the light output. Thinking of light output as energy expended over a given period of time may help to explain it. If you switch the LED on and off for say a millesecond at a time over the course of a minute, it will have put out less light than if you had left it on for the entire minute. Because the LED is turned on and off so fast, the human eye can't see the event take place, but interpolates the data from when the LED is on and off effectively creating the perception that the LED is dimmer. For the split second it was on, the LED is just as bright as it would be under normal operation; but because it was off, for a portion of that minute, it used less power and created the illusion of being dimmer. Maybe someone can give an easier to understand explanation, but hopefully that will help.

If you want to order up parts, you should figure out how much light you want and how long you want it to burn for. There are different drivers available to suit most combinations. Sort out your battery and LED choices first to get the lumens and run time that you want; then pick the appropriate driver to make the power source and LEDs compatible.

Chris

[92se-r]

Brewmaster, I was going over some of your pictures, I think you got some cold solder joints going on. I'd reheat those connections again and make sure the solder flows into the strands of wire or the connection will most likely break off.

Schecky, what kind of light did you have? Mind if I take a look at it? The one you are planning on scrapping.

[Flat Broke]

Originally Posted by 92se-r

You can't step up current to meet the needs of your LED. Only voltage.

I believe most LED drivers use PWM(pulse width modulation) to control dimming. Most do not have adjustable current.

You don't use buck drivers for LED's. You do not have to step down your battery voltage for LED's. LED's are like a faucet. At a certain voltage, they turn on. That's it. The LED voltage is fixed and not determined by the battery.

Linear drivers do not mean that they use PWM. Linear dc-dc converters are just a different topology than a switcher. They usually have more power loss through the converter.

PWM is Pulse Width Modulation. You do not short out the LED to do this. PWM basically outputs a square wave, where depending on the PWM duty cycle, there is a ratio between on time and off time. So if you want to dim the LED, you adjust the PWM down so that the pulse width of the output of the driver is less than the time that the output is off. Basically the driver turns on and off really fast and the pulse width(how long the output is high) is determined by what you set the PWM at.

When you choose a LED driver, you need to look at the minimum voltage headroom it requires. Add up your LED voltages, then add your minimum voltage headroom for the driver, and you want to get a battery voltage that is higher than this. Ideally, you want to be closer to the LED voltage + headroom because this means less heat dissipated through the driver. Not really an issue on a bike light with the readily available batteries out there.

I appreciate the clarification on the PWM functionality. Whether it shorts it out or turns it off, at least I got the part about it basically manipulating the time the led is on in a given period right. I agree that "short" is an incorrect term and I apologize for using it.

However, with regard to the Bflex, Nflex, etc drivers, I'm pretty sure they regulate the current. You can check out the spec at http://www.taskled.com/bflex.html for an example. The only one of their drivers that I see that lists PWM is the D2DIM.

This is something I've been trying to get my head around, and maybe you can help clarify it. I've always understood electricity as a balancing act using ohms law. Basically you could manipulate the current, voltage, or power of a circuit by changing the other two variables. As a result, couldn't you trade run time for voltage in the case of power hungry LEDs like the Ostar. Obviously I'm missing something, and I'd love to put all the pieces of this puzzle together.

As I understand what you're saying about LEDs; once you meet their voltage requirement, you can drive above that voltage to an infinite amount so long as you keep the current within spec? Is there some point where excessive voltage turns a LED into a DED?

I thought it was possible to do something like use a "boost" driver to not only maintain the current, but also increase the voltage like with a traditional step up transformer. If you used the 6 die Ostar as an example, running at .7A, it utilizes 22v. Could you use a step up/boost driver like this http://www.taskled.com/maxflex.html to drive the emitter with say a 7.4 volt battery? In this scenario, does the driver effectively consume more amps to generate the required voltage for the emmitter? If so, what does the equation look like for that? I'm going to guess that if I need to take 7.4 volts to 22volts, I'm going to multiply the .7A current of the driver by 2.97 (22v divided by 7.4v=2.97). So this would equal 2.079 Amps. From there, you would just divide the mAh rating of the battery by the amperage consumed correct? I'm guessing I've got something roally hosed up because that would mean that a 4800mAh 7.4v battery would run for 2.3hours powering a 22V Ostar at 700ma as described above.

I'm not trying to be a smart ass, I'm genuinely interested in this stuff and would like to get the fundamentals down solid for any future projects I attempt. Thank you for taking the time with your previous posts, and I hope that you can further explain this not only for myself but for others that may be interested.

Chris

[92se-r]

YOu are currect. They actually do have adjustable current.

LED's are diodes. So if you look at a VI curve of a diode, it will help explain what I'm going to say better. A Diode consists of a PN junction. In laymans terms, if the PN junction is positively biased, the anode voltage higher than the cathode voltage, then the diode will allow current through. The thing you need to remember with diodes is that their voltage drop across the diode is always fixed. However, the amount of current they allow, is basically what you will provide it. So for example, a battery can source a ton of instantaneous current. If you just hookup a battery straight to a LED, it will blow up the LED because the LED allows as much current as the battery will give it. It'll heat up and blow up. That is why you either use a resistor in series with the LED to current limit it, because think of a resistor as a current limiter. Or you use a constant current driver like all these LED drivers. That way the LED does not blow up. You can overvoltage a diode by hooking it up backwards and going over its max rated reverse voltage.

The OSTAR LED's will affect battery life but you want to use a higher battery voltage to keep input current closer to the output current. Where you will also see the difference is the extra power(heat) you need to dissipate now off the LED's. So yes, you will lose some battery life, so use a higher battery voltage. If you use a really low one like the 7.4V, your battery life will really suffer. A 14.8V 4000mAh battery, you would lose around an hour, so actually, that is significant.

Originally Posted by Flat Broke

However, with regard to the Bflex, Nflex, etc drivers, I'm pretty sure they regulate the current. You can check out the spec at http://www.taskled.com/bflex.html for an example. The only one of their drivers that I see that lists PWM is the D2DIM.

This is something I've been trying to get my head around, and maybe you can help clarify it. I've always understood electricity as a balancing act using ohms law. Basically you could manipulate the current, voltage, or power of a circuit by changing the other two variables. As a result, couldn't you trade run time for voltage in the case of power hungry LEDs like the Ostar. Obviously I'm missing something, and I'd love to put all the pieces of this puzzle together.

As I understand what you're saying about LEDs; once you meet their voltage requirement, you can drive above that voltage to an infinite amount so long as you keep the current within spec? Is there some point where excessive voltage turns a LED into a DED?

I thought it was possible to do something like use a "boost" driver to not only maintain the current, but also increase the voltage like with a traditional step up transformer. If you used the 6 die Ostar as an example, running at .7A, it utilizes 22v. Could you use a step up/boost driver like this http://www.taskled.com/maxflex.html to drive the emitter with say a 7.4 volt battery? In this scenario, does the driver effectively consume more amps to generate the required voltage for the emmitter? If so, what does the equation look like for that? I'm going to guess that if I need to take 7.4 volts to 22volts, I'm going to multiply the .7A current of the driver by 2.97 (22v divided by 7.4v=2.97). So this would equal 2.079 Amps. From there, you would just divide the mAh rating of the battery by the amperage consumed correct? I'm guessing I've got something roally hosed up because that would mean that a 4800mAh 7.4v battery would run for 2.3hours powering a 22V Ostar at 700ma as described above.

I'm not trying to be a smart ass, I'm genuinely interested in this stuff and would like to get the fundamentals down solid for any future projects I attempt. Thank you for taking the time with your previous posts, and I hope that you can further explain this not only for myself but for others that may be interested.

Chris

[Flat Broke]

Thanks for getting back to me on this. I'm a big fan of the "teach me to fish" aspect of stuff like this. With that in mind, did I actually have the voltage and mAh figgured out correctly for the scenario I had posted with a 22v/.7A emitter?

Oddly enough when I use the calculator at http://www.jtice.com/led_pro/led_pro.htm and their default 83% efficiency, I get a run time of 1.91 hours using the same voltage and amperage constraints as above. If I bump the efficiency up to 100%, I get 2.31hrs which is what I arrived at using the calculation in my previous post. Maybe I had it right the first time around? If someone can show me the correct order of operations to figure out the run times, and characteristics of this circuit inedpendent of a 3rd party calculator, I'd greatly appreciate it. It looks like my first attempt was correct, but I'd like validation on that before I make any purchases based upon that math

Chris

[tkblazer]

i'm starting to think that this thread and the geek thread need to be merged

my avatar sums up my thoughts on a the last few posts

[92se-r]

It's correct based on 100% efficiency, but there are a lot of differnet things that factor into efficiency that I am too lazy to talk about.

Personally, I think the Ostars are way overrated. If you look at their Lumens/Watt, they dont compare to the Cree's and Luxeons. They just package it really small, which is their only advantage. I'll be posting a picture of the light I just finished soon.

[Flat Broke]

I'm with you on the lumen/watt and $$/lumen negatives associated with the Ostar at 500ish lumens at 700ma. Buuut there is a PZ-QZ bin range that runs between 610 and 1120 lumens at 700ma, so that helps out. They're pricey little buggers though at $42 each. The compact form factor is what I'm after, hence my interest in this particular LED over the Crees or Seouls.

If you want the revised data sheet for that emitter, feel free to shoot me an email. Also, if you ever want to expound upon some potential efficiency losses in the driver and LED, I'm all ears.

Since my math seems to jive, I'll be placing my orders for parts on Monday to start an Ostar based light. Feel free to post up specs and pics of your light.

Chris

[uhyeah]

Now that you have used this light for a few months, are there any changes to the basic design that you would recommend? I have everything bought to build one including the cyan colored Luxeon LED's. So far, mine will only differ from your design by my use of a 0~5k ohm variable resistor for dimming, and I'll be gluing/ sealing the whole mess to the heat sink with "potting" epoxy.
Dino B. suggested that I ask these questions, since up until now, I've been a "lurker" without much to say.

[Dino Brown]

Originally Posted by uhyeah

... Dino B. suggested that I ask these questions, since up until now, I've been a "lurker" without much to say.

Dino B suggested you do it ONE WEEK AGO!!!

Lurkers....

[92se-r]

Potting epoxy has a very high thermal resistance if I remember correctly. NOT ideal for transfering heat. Or are you using it with some Arctic Silver and then using the potting epoxy on the OUTSIDE of the mating surface for structural support?

I found that I was able to remove the 7.4W of heat fine with just a square piece of Al.



I had some painting issues though and camera issues.

If you let the thing sit on a table with no airflow over it and run it for about 30 minutes, it settles at around 55-60C surface temp. While riding though, you can't even feel that it is warm.

Originally Posted by uhyeah

Now that you have used this light for a few months, are there any changes to the basic design that you would recommend? I have everything bought to build one including the cyan colored Luxeon LED's. So far, mine will only differ from your design by my use of a 0~5k ohm variable resistor for dimming, and I'll be gluing/ sealing the whole mess to the heat sink with "potting" epoxy.
Dino B. suggested that I ask these questions, since up until now, I've been a "lurker" without much to say.

[jeffj]

I'm pretty sure he's using Arctic Silver (or similar material) for heat transfer and the potting epoxy for structural support.