Bike Forums

Bike Forums (https://www.bikeforums.net/forum.php)
-   Electronics, Lighting, & Gadgets (https://www.bikeforums.net/electronics-lighting-gadgets/)
-   -   Power LED Question (https://www.bikeforums.net/electronics-lighting-gadgets/788183-power-led-question.html)

gna 12-22-11 12:15 PM

Power LED Question
 
Hello--

I built a LED light for my winter bike, using diagram #2 at pilom.com . I used a bridge rectifier I had laying around with a 1000 uf cap. The power source is a Sturmey-Archer X-FDD (dynamo Drum) hub. It produces 6v, 3w. The Power LED is a Cree XR-E-R2 LED from Deal Extreme. I used JB weld to attach the star to the head of a bolt, with a large brass nut for heat sinking. It's very bright and seems to work fine.

I have two questions:

My first question is whether or not I am damaging the LED. I showed the setup to someone at work, and he said the 6V from the dynohub will burn up my LED. I thought I read somewhere on here that these LED can handle the current and the voltage isn't a problem.

The second question is about a tail light. While Pilom has some circuits for tail lights, could I just add a Red LED in series to the white front LED? That would be easier to wire.

roashru 12-22-11 03:38 PM

yes, you will fry it. go to the cree web site look for the product page and read the "data sheet". white for that model says 3.2 volts.

christ0ph 12-22-11 04:27 PM

Dynamos are a constant current source, so paradoxically, adding another LED in series will in many cases get you more power without the penalty you would expect, or so I just read. That said, I think you would be wise to invest in a constant current driver, given that they are so cheap. You'll end up with a more efficient setup that way. One thing you really need to worry about is the hub putting out a much higher peak voltage which because of the rectifier, wont have anywhere to go except charging that cap to a high voltage.. (BEFORE you connect the LED, why not perform an experiment to see how high it will go with your bike on a stand or upside down..with you turning the pedals by hand.. that would be interesting.) In other words, nomatter what the RMS voltage of the generator, I think that with no load, the cap will charge to the PEAK voltage... So then, unless you leave both lights on all the time (which makes some sense if the drag isn't high, but which will slow you down if it isn't)
How old is the hub, (many newer hubs have bult in zener diodes but older ones dont, Ive read, and they can put out very high peak voltages which will then get stored by that cap) and what are the current ratings on the two LEDs?

In the meantime, why not pick the lower of the two LED's maximum current value and then, using Ohms law, pick a low value resistor to throw in series, also put an extremely high value - in (CRT) TVs, bleeder caps are usually megohms- but that might be too high, use a multimeter if you have one to see how long the cap holds a charge under different "road conditions" (you spinning the wheel by hand) ) to put across the cap to bleed off the high voltage should your light get disconnected at some point, to prevent it from being burned out by the charge on the cap.. which could "potentially" get high.. For the series resistor (MAKE SURE THATS ITS *BETWEEN* THE CAP and the LEDS...not before it.. I'd also start with a value that would limit the voltage if the LED was not there to the peak current for the smaller of the two LEDs at some arbitrarily higher voltage- the value I would pick is 20volts.. and then work down from there (because they are both probably way too high)

The thing you dont want is the cap charging up high while riding with the headlight not connected, i.e. no load, and then, connecting it only to have it burn out because unknowingly you gave it a 30 or 40 volt zap there..which ended up pushing much more current through the LED than the peak design value . Even a good heat sink wouldnt help you in that situation.

In other words, smooth out the potential for voltage and current peaks..

I'm feeling a bit frazzled today so maybe I am not making sense.. Does that make sense?


Originally Posted by gna (Post 13629247)
Hello--

I built a LED light for my winter bike, using diagram #2 at pilom.com . I used a bridge rectifier I had laying around with a 1000 uf cap. The power source is a Sturmey-Archer X-FDD (dynamo Drum) hub. It produces 6v, 3w. The Power LED is a Cree XR-E-R2 LED from Deal Extreme. I used JB weld to attach the star to the head of a bolt, with a large brass nut for heat sinking. It's very bright and seems to work fine.

I have two questions:

My first question is whether or not I am damaging the LED. I showed the setup to someone at work, and he said the 6V from the dynohub will burn up my LED. I thought I read somewhere on here that these LED can handle the current and the voltage isn't a problem.

The second question is about a tail light. While Pilom has some circuits for tail lights, could I just add a Red LED in series to the white front LED? That would be easier to wire.


gna 12-24-11 02:23 PM


Originally Posted by roashru (Post 13630066)
yes, you will fry it. go to the cree web site look for the product page and read the "data sheet". white for that model says 3.2 volts.

Correct me if I'm wrong, but I thought current was the prime factor with LEDs, not voltage. The LED can take the 500-600 mA my hub can put out, so that was the big concern. Normally if the voltage increases, the current will as well, but the Dynohub can only do so much.


Originally Posted by christ0ph (Post 13630216)
Dynamos are a constant current source, so paradoxically, adding another LED in series will in many cases get you more power without the penalty you would expect, or so I just read. That said, I think you would be wise to invest in a constant current driver, given that they are so cheap. You'll end up with a more efficient setup that way. One thing you really need to worry about is the hub putting out a much higher peak voltage which because of the rectifier, wont have anywhere to go except charging that cap to a high voltage.. (BEFORE you connect the LED, why not perform an experiment to see how high it will go with your bike on a stand or upside down..with you turning the pedals by hand.. that would be interesting.) In other words, nomatter what the RMS voltage of the generator, I think that with no load, the cap will charge to the PEAK voltage... So then, unless you leave both lights on all the time (which makes some sense if the drag isn't high, but which will slow you down if it isn't)
How old is the hub, (many newer hubs have bult in zener diodes but older ones dont, Ive read, and they can put out very high peak voltages which will then get stored by that cap) and what are the current ratings on the two LEDs?

In the meantime, why not pick the lower of the two LED's maximum current value and then, using Ohms law, pick a low value resistor to throw in series, also put an extremely high value - in (CRT) TVs, bleeder caps are usually megohms- but that might be too high, use a multimeter if you have one to see how long the cap holds a charge under different "road conditions" (you spinning the wheel by hand) ) to put across the cap to bleed off the high voltage should your light get disconnected at some point, to prevent it from being burned out by the charge on the cap.. which could "potentially" get high.. For the series resistor (MAKE SURE THATS ITS *BETWEEN* THE CAP and the LEDS...not before it.. I'd also start with a value that would limit the voltage if the LED was not there to the peak current for the smaller of the two LEDs at some arbitrarily higher voltage- the value I would pick is 20volts.. and then work down from there (because they are both probably way too high)

The thing you dont want is the cap charging up high while riding with the headlight not connected, i.e. no load, and then, connecting it only to have it burn out because unknowingly you gave it a 30 or 40 volt zap there..which ended up pushing much more current through the LED than the peak design value . Even a good heat sink wouldnt help you in that situation.

In other words, smooth out the potential for voltage and current peaks..

I'm feeling a bit frazzled today so maybe I am not making sense.. Does that make sense?

You make some sense, but it's limited by my understanding. I may just throw another LED in series for a rear light and be done with it. I just leave them on all the time, so I'm not worried about the capacitor. The X-FDD is a newer hub, so I'm pretty sure it limits output.

minisystem 12-24-11 08:59 PM


Originally Posted by gna (Post 13629247)

I have two questions:

My first question is whether or not I am damaging the LED. I showed the setup to someone at work, and he said the 6V from the dynohub will burn up my LED. I thought I read somewhere on here that these LED can handle the current and the voltage isn't a problem.


In a word, no, you're not damaging the LED. So long as it isn't overheating (your heat sinking sounds adequate). As mentioned, dynamos operate as constant current devices, meaning they reach a saturation current at some minimum speed. I've run the XFDD on my testing jig and it saturates at around 600 mA. That means that even on a very rapid downhill descent, you can't blow the LED because the hub just won't put out anymore current (I've measured up to 50 km/h and at that point the current has plateaued).

Dynamo hubs cause confusion because their output voltage is determined by the load. The open circuit voltage of a hub can get very high (over 100V say some, although I've never measured more than 30V on my test setup - some modern hubs have their voltage outputs limited by an internal zener diode). But when the load is an LED the output voltage will the the forward voltage of the LED, no matter what the speed is. With a fixed load like the LED, current is a function of speed, not voltage.

In my experience, you do not need to take any extra precautions to protect the LED other than heat sinking. A constant current LED driver seems intuitively like a good idea, but it adds complexity and isn't necessary. The rectified output of the hub dynamo is sufficient.

(I say this after spending months designing a driver circuit that regulated LED current and realizing my efforts were unnecessary!)


Originally Posted by gna (Post 13629247)
The second question is about a tail light. While Pilom has some circuits for tail lights, could I just add a Red LED in series to the white front LED? That would be easier to wire.

Yes, adding this in series should be fine. The front and rear LED will get the same current, so you just need to make sure that both LEDs have a maximum current rating of at least 600 mA.

Disclaimer: I don't like to state anything with too much authority. Most of what I know comes from the wise folks in the bicycle lighting forums at candlepowerforums.

unterhausen 12-25-11 12:08 AM

I don't think you can fry an XRE with a Sturmey dynohub. I'm not sure I believe you have good enough heat sinking though.

znomit 12-25-11 12:43 AM

Yeah the JB Weld heatsinking to the bolt might be a little iffy but the circuits on Pilom.com are solid.
The general idea is a bike dynamo puts out no more than 500mA so modern power LEDs are fine.

Get some thermal glue from DX too.

gna 12-28-11 11:28 AM


Originally Posted by unterhausen (Post 13638261)
I don't think you can fry an XRE with a Sturmey dynohub. I'm not sure I believe you have good enough heat sinking though.

Is it the JB Weld or the bolt and nut?


Originally Posted by znomit (Post 13638311)
Yeah the JB Weld heatsinking to the bolt might be a little iffy but the circuits on Pilom.com are solid.
The general idea is a bike dynamo puts out no more than 500mA so modern power LEDs are fine.

Get some thermal glue from DX too.

Like this?

I have some 1" square aluminum tubing stock left over from rebuilding a trailer. Perhaps I can use that to build a light, and make the whole body the heatsink. Hmmm...

roashru 12-30-11 01:11 PM

heat sinks are a major factor, but anything pass 5 volts damages solid state electronic components (parts). amps can be controlled with a heat sink. volts is what burns them up.

minisystem 12-30-11 06:39 PM


Originally Posted by gna (Post 13647983)
Is it the JB Weld or the bolt and nut?

A quick search seems to suggest a consensus that JB-Weld is not a very good thermal conductor. As suggested,a thermal epoxy would be an improvement.


Originally Posted by gna (Post 13647983)
I have some 1" square aluminum tubing stock left over from rebuilding a trailer. Perhaps I can use that to build a light, and make the whole body the heatsink. Hmmm...

That seems to be a popular method among DIY light builders.

znomit 12-30-11 07:17 PM


Originally Posted by gna (Post 13647983)
Is it the JB Weld or the bolt and nut?



Like this?

I have some 1" square aluminum tubing stock left over from rebuilding a trailer. Perhaps I can use that to build a light, and make the whole body the heatsink. Hmmm...

I think I have some of that DX glue. Its likely much better than JB for heat transfer.

1 inch aluminum makes a decent housing. I've used it on a bunch of lights. Here's a pretty simple one.

unterhausen 12-30-11 07:25 PM


Originally Posted by roashru (Post 13656148)
heat sinks are a major factor, but anything pass 5 volts damages solid state electronic components (parts). amps can be controlled with a heat sink. volts is what burns them up.

What you have posted here is incorrect. There are many solid state devices that can take much higher voltages than 5 volts. If an led can take the max current generated by a current source, it doesn't really need regulation. Many currently available power leds can withstand over 1A, they are not going to burn out using a dynohub that puts out less than half of that. The forward voltage drop of the led is going to determine the voltage coming out of the dynohub to a great degree.

salek 01-02-12 01:31 PM

Just be ready to do the math and don't limit yourself to "old" style leds. I have installed LED lighting around my house and on my Volkswagons. Most of my lights are 6.7VDC at 700mA. I have four new ones that are 14VDC at 1Amp. I have a single killer that is 6.7VDC at 8 amps. Yes, eight. If I can learn how to adjust the power, I'm interested in this new 40W led that runs at 44VDC and the appropriate amperage.
I am considering building a twin mount of either the 6.7V or 14V for my Big Dummy.

roashru 01-06-12 09:20 PM

interesting info on gallium nitride l.e.d.s:

http://www.oksolar.com/led/led_color_chart.htm

a three junction digital component is still limited to 5 volts everything else is an i.c. chip
.

christ0ph 01-09-12 01:51 AM

watch out for capacitors in that situation if you ever turn your led off. Voltage will be stored at its peak value. Which could get quite high.

Mondoman 01-17-12 03:02 PM


Originally Posted by znomit (Post 13657397)
I think I have some of that DX glue.

That DX "glue" looks like standard silicone heat sink thermal compound, which is not a glue, but rather a grease. It will provide some stiction, especially as the recommended extremely thin layer, but will allow the "attached" component to move around and be removed. You probably want to use something like this instead: http://www.arcticsilver.com/arctic_s...l_adhesive.htm


All times are GMT -6. The time now is 09:46 AM.


Copyright © 2024 MH Sub I, LLC dba Internet Brands. All rights reserved. Use of this site indicates your consent to the Terms of Use.