A Cycling & bikes forum. CycleBanter.com

Go Back   Home » CycleBanter.com forum » rec.bicycles » Techniques
Site Map Home Register Authors List Search Today's Posts Mark Forums Read Web Partners

Electronic tech: Dynamo headlamp circuit



 
 
Thread Tools Display Modes
  #21  
Old December 30th 13, 11:21 PM posted to rec.bicycles.tech
Ralph Barone[_3_]
external usenet poster
 
Posts: 321
Default Electronic tech: Dynamo headlamp circuit

Joe Riel wrote:
Ralph Barone writes:

Frank Krygowski wrote:
I've been wondering what electronic circuits are inside dynamo powered
LED headlamps. Rectification, almost certainly, since bike dynamos
output AC, but LEDs conduct in only one direction.

(There is the useful scheme of connecting two or more LEDs
anti-parallel, so no other rectification is needed and each protects its
mate from excess back voltage.)

Most importantly, I figured no current regulation would be needed, since
bike dynamos are almost all self-regulated by design to 0.5 Amp output, a
value which works well with most high output LEDs. This idea was
confirmed by a friend who is an LED product engineer.

Well, an Avenir 3 Watt dynamo headlamp gave me a reason to find out about
the circuit. It was inexpensive (about $30), and had optics that
produced a pretty nice road beam, rather than a simple round blob of
light. Not nearly as much light output as my Cyo, but adequate for the
bike to which I fitted it.

But in less than a year, it failed. And it was no longer available on
the market. Avenir treated me well, sending me a fairly nice battery
light (but with "round blob of light" optics - all they had) plus a
multi-tool and seat bag as consolation. Better, they didn't want the
broken light back, and gave me permission to dissect it instead.

Turned out the main LED had failed, something that's very rare with LED
lamps. With much trouble, I was able to fit another, higher output LED I
had on hand (sorry, no specs - it was a gift from that LED engineer).
Larger die, tougher to fit, optics a bit less precise, but total light
output is greater. It'll work fine. Unless, that is, something in their
circuit cooks LEDs!

I spent considerable time investigating the circuit. For those
interested, it's posted on Flicker at
http://www.flickr.com/photos/16972296@N08/11635110843/

Dynamo input is at left. I didn't fit the taillight, but its connections
are shown. P6KE6.80 seems to be to absorb any input spikes, and/or
provide some simple input voltage regulation; I'm not sure. U2 must be a
full-wave rectifier; it's got AC to the left, DC (positive at the top) to
its right. The two push-buttons shown are actually one DPST switch on
the circuit board. The 1F capacitor powers the standlight feature. One
tiny LED is an on/off indicator. The main LED is at far right. I
_think_ D1802 S 34 is a fast-switching power transistor used to regulate
voltage to the main LED.

But I'm not an EE. Comments more than welcome.

- Frank Krygowski


The transistor appears to be a capacitance multiplier, which buys you a
smaller capacitor at the expense of more voltage drop. Seems a bit
goofy...


That could be the purpose. Any idea how much inductance is in a typical
generator (hub and/or bottle)? Using a simple model, the capacitance
multiplier reduces the peak-peak ripple current through the LED from
about 14mA to 1mA. That was out of an average current of 250mA. I
doubt that affect would be noticeable.



I didn't do any analysis, but thought that at high speeds, the flicker
wouldn't be noticeable, while at low speeds, the loss of efficiency
involved on masking the flicker would be unacceptable.
Ads
  #22  
Old December 30th 13, 11:56 PM posted to rec.bicycles.tech
Ralph Barone[_3_]
external usenet poster
 
Posts: 321
Default Electronic tech: Dynamo headlamp circuit

sms wrote:
On 12/30/2013 11:14 AM, Jeff Liebermann wrote:

The real question is at what temperature is the LED operating. Too
high, and the projected lifetime becomes rather short:
http://www.digikey.com/us/en/techzone/lighting/resources/articles/LED-Heat-Dissipation.html


Note that the article uses common PCB material (FR4) as a worst case
example of a lousy heat sink material.


You're so wrong Jeff.

"You can't make a typical incandescent bulb work well without having a
super-hot filament. You can't have a super-hot filament without having
it burn out eventually. But you _can_ make an ordinary LED work well,
and last damned close to forever, without getting it hot."

Actually what happened was that it became possible to manage the
thermals in order to make an LED work well even when it gets very hot.
Not that every LED light manufacturer knows how to do thermal
management. Apparently Avenir didn't figure it out. And the 100,000 hour
life of LED indicator lamps doesn't extrapolate to high-power LEDs.

There's no free lunch. High power lamps have always had thermal issues
that have had to be managed.


Jeff's so wrong about what? Or did you snip the part you objected to?
  #23  
Old December 31st 13, 12:00 AM posted to rec.bicycles.tech
James[_8_]
external usenet poster
 
Posts: 6,153
Default Electronic tech: Dynamo headlamp circuit

On 31/12/13 06:14, Jeff Liebermann wrote:
On Mon, 30 Dec 2013 10:04:07 -0800 (PST), Andy M-S
wrote:

Why NOT a waterproof casing with decent heat sinking, rather than an extra circuit to reduce power?


Because heat sinking an LED is difficult:


I do not agree.

--
JS
  #24  
Old December 31st 13, 12:36 AM posted to rec.bicycles.tech
Frank Krygowski[_2_]
external usenet poster
 
Posts: 7,511
Default Electronic tech: Dynamo headlamp circuit

On Monday, December 30, 2013 7:01:07 AM UTC-5, Andy M-S wrote:
Interesting. I've built much cruder LED headlights (2x3W Luxeons, in series, run off a fw bridge, and my first, a 3W Luxeon with a bridge, built into a chunk of threadless stem for built-in heat sinking) and had no burn-out problems, but then I tend to design, um, "conservatively" with regard to heat sinks. As in, the LEDs are thermal-epoxied to a wide, thickish (1/8") sheet of aluminum that is then attached/thermal epoxied to a metallic case.. I ran that light for several years of commuting, and it was "on" pretty much all of the time in the Wisconsin winter (of course that aids cooling, too--the winter, that it). The source in all cases was a 3W Shimano dynohub,



I have one of the Avenir lights, and the first thing that struck me (after how much better the light was than my home-brew unit in terms of lighting the road) was that the casing was made of plastic. No exposed heat sink at all, so once the sink is hot, there's not much place for heat to go. I only used it for a little while, but I'm planning to put it on my "go fast" bike with a BB generator as an emergency "after dark" system. If I have LED troubles, I'll likely just gut it and devise a sink with surfaces outside the shell for heat dissipation.


The Avenir's LED has thermal paste to a surface on an internal heat sink, which forms the chassis, so to speak, and which has fins. But you're correct, all that can do is dump the heat to the inside of the plastic shell. Not nearly as good as the B&M lights that have the LED thermally linked to the metal top of the light, in the cooling wind.

- Frank Krygowski
  #25  
Old December 31st 13, 01:07 AM posted to rec.bicycles.tech
SMS
external usenet poster
 
Posts: 9,477
Default Electronic tech: Dynamo headlamp circuit

On 12/30/2013 2:56 PM, Ralph Barone wrote:
sms wrote:
On 12/30/2013 11:14 AM, Jeff Liebermann wrote:

The real question is at what temperature is the LED operating. Too
high, and the projected lifetime becomes rather short:
http://www.digikey.com/us/en/techzone/lighting/resources/articles/LED-Heat-Dissipation.html


Note that the article uses common PCB material (FR4) as a worst case
example of a lousy heat sink material.


You're so wrong Jeff.

"You can't make a typical incandescent bulb work well without having a
super-hot filament. You can't have a super-hot filament without having
it burn out eventually. But you _can_ make an ordinary LED work well,
and last damned close to forever, without getting it hot."

Actually what happened was that it became possible to manage the
thermals in order to make an LED work well even when it gets very hot.
Not that every LED light manufacturer knows how to do thermal
management. Apparently Avenir didn't figure it out. And the 100,000 hour
life of LED indicator lamps doesn't extrapolate to high-power LEDs.

There's no free lunch. High power lamps have always had thermal issues
that have had to be managed.


Jeff's so wrong about what? Or did you snip the part you objected to?


Sorry, sarcasm doesn't come across well on Usenet. Jeff is completely
correct of course. I was referring to something our resident troll wrote
a while back.


  #26  
Old December 31st 13, 01:09 AM posted to rec.bicycles.tech
SMS
external usenet poster
 
Posts: 9,477
Default Electronic tech: Dynamo headlamp circuit

On 12/30/2013 3:00 PM, James wrote:
On 31/12/13 06:14, Jeff Liebermann wrote:
On Mon, 30 Dec 2013 10:04:07 -0800 (PST), Andy M-S
wrote:

Why NOT a waterproof casing with decent heat sinking, rather than an extra circuit to reduce power?


Because heat sinking an LED is difficult:


I do not agree.


Not agreeing does not make it untrue. The difficulty in transferring the
heat from the tiny semiconductor junction is the limiting factor in
increasing the output from an LED.

  #27  
Old December 31st 13, 01:52 AM posted to rec.bicycles.tech
Frank Krygowski[_2_]
external usenet poster
 
Posts: 7,511
Default Electronic tech: Dynamo headlamp circuit

On Monday, December 30, 2013 11:19:00 AM UTC-5, sms wrote:

If an LED (or other semiconductor) fails in a situation where it can't
have too much current forced through it then it's a pretty safe bet that
it's a heat related failure caused by an inadequate thermal solution.

It's become non-trivial to cool LED bicycle lights. At Interbike this
year I saw several lights that have incorporated temperature sensors and
that automatically reduce power to the LED when it becomes too hot--you
cannot run it at maximum power continuously, and the amount of time you
can run at maximum power depends on the airflow over the light.


FWIW: When replacing the original failed LED with the more powerful one I'd been given, I first used jumpers to clip it into the circuit. I then ran it from a bike generator driven by my drill press. While I didn't measure the temperature of the LED's backing plate, I did hold it between my fingers for quite a while. I never got too hot to hold. I estimate it was at 110 degrees F or so.

According to my LED engineer friend (who gave me the high-powered ones I have) the main problem from insufficient cooling would be a somewhat shorter life. But as he said, with a rated life of around 10,000 hours, a bike headlight could accept a much shorter life with no problem.

I think it's more likely that the failure of the LED in the Avenir was just an example of random early death. "Burning in" of electronics is supposed to catch most of those. Maybe the manufacturer of these lights didn't do a burn-in. It was inexpensive.

For bicycle lighting you have the additional constraints that the light
needs to be waterproof so you can't have the most optimal airflow and
you also don't want massively heavy heat sinks. At least you don't have
to hold a bicycle light in your hand so it doesn't matter if the outer
case gets very hot.


I'll also note that most LED bike lights - the battery powered ones with foggy round beams - are plastic, with no great amount of external heat sinking. And IME admittedly limited experience, they don't get extremely hot.

Amusingly, when LED lights were first being used for bicycles someone on
Usenet gushed about how wonderful they were because there was no "white
hot filament." What they didn't understand was that there's no free
lunch and that the "white hot filament" had been replaced by a
semiconductor junction that also got extremely hot, but that there
wasn't a visible indication of the heat, and that the heat was difficult
to dissipate. You get three guesses as to the source of this wisdom, and
the first two don't count.


Is there a white hot filament in an LED? Is the amount of waste heat per lumen for an LED not much less than for an incandescent bulb?

The main difference, thermally, is that the heat has to leave by conduction then convection, instead of mainly by radiation. But of course, there's always been conduction from incandescent bulbs as well. Many years ago, I increased the bulb wattage in one of my battery powered lights and partially melted their bases in the plastic reflectors. That was entirely conduction.
  #28  
Old December 31st 13, 02:45 AM posted to rec.bicycles.tech
James[_8_]
external usenet poster
 
Posts: 6,153
Default Electronic tech: Dynamo headlamp circuit

On 31/12/13 11:09, sms wrote:
On 12/30/2013 3:00 PM, James wrote:
On 31/12/13 06:14, Jeff Liebermann wrote:
On Mon, 30 Dec 2013 10:04:07 -0800 (PST), Andy M-S
wrote:

Why NOT a waterproof casing with decent heat sinking, rather than an
extra circuit to reduce power?

Because heat sinking an LED is difficult:


I do not agree.


Not agreeing does not make it untrue. The difficulty in transferring the
heat from the tiny semiconductor junction is the limiting factor in
increasing the output from an LED.


It's not difficult at all, and saying it is doesn't make it so.

--
JS
  #29  
Old December 31st 13, 07:05 AM posted to rec.bicycles.tech
Jeff Liebermann
external usenet poster
 
Posts: 4,018
Default Electronic tech: Dynamo headlamp circuit

On Mon, 30 Dec 2013 14:16:13 -0800, sms
wrote:

On 12/30/2013 11:14 AM, Jeff Liebermann wrote:

The real question is at what temperature is the LED operating. Too
high, and the projected lifetime becomes rather short:
http://www.digikey.com/us/en/techzone/lighting/resources/articles/LED-Heat-Dissipation.html


Note that the article uses common PCB material (FR4) as a worst case
example of a lousy heat sink material.


You're so wrong Jeff.


Oh goodie. I just hate it when everyone agrees with me.

FR4 fiberglass is a rather bad heat sink. Copper is much better. Vias
filled with solder connected to copper is so-so. What did I get wrong
(this time)?

Actually what happened was that it became possible to manage the
thermals in order to make an LED work well even when it gets very hot.
Not that every LED light manufacturer knows how to do thermal
management. Apparently Avenir didn't figure it out. And the 100,000 hour
life of LED indicator lamps doesn't extrapolate to high-power LEDs.


I think (not sure) that was 100,000 hr MTBF, not lifetime. There's a
difference. 100,000 hr MTBF (57 years) means that it you take a large
number of such lights, all running at roughly the same temperature,
you'll see one failure for every 100,000 hrs of COMBINED operation.
Start with 100 LED's. Run them all for 1 year or 876,000 hrs.
Therefore, one would expect:
876,000 / 100,000 = 8.76 failures/year
or one LED failure roughly every 1.4 months.

Lifetime is simply when an individual LED is expected to deteriorate
in performance down to some measurable level. Energy Star reference
for 35,000 hrs (minimum) is 70% of initial luminous output.

There's no free lunch. High power lamps have always had thermal issues
that have had to be managed.


Something about the light at the end of the tunnel being a fire.

Cree XLamp LEDs
Solder Joint Reliability Study
http://www.cree.com/~/media/Files/Cree/LED%20Components%20and%20Modules/XLamp/XLamp%20Application%20Notes/Solder%20Joint%20Reliability.pdf


--
Jeff Liebermann
150 Felker St #D
http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558
  #30  
Old December 31st 13, 07:42 AM posted to rec.bicycles.tech
Jeff Liebermann
external usenet poster
 
Posts: 4,018
Default Electronic tech: Dynamo headlamp circuit

On Tue, 31 Dec 2013 10:00:39 +1100, James
wrote:

On 31/12/13 06:14, Jeff Liebermann wrote:
On Mon, 30 Dec 2013 10:04:07 -0800 (PST), Andy M-S
wrote:

Why NOT a waterproof casing with decent heat sinking, rather than an extra circuit to reduce power?


Because heat sinking an LED is difficult:


I do not agree.


Methinks you might be trivializing the problems. Make one light is
probably easy enough. Doing it in production is not so easy.

http://www.cree.com/LED-Components-and-Modules/Products/XLamp
If you select a large die LED, possibly in the range of 10-20 mm per
side, those are fairly easy to handle and attach. Just slop some
solder paste on the heat sink, heat, reflow, inspect and you're done.
Plenty of room for the solder balls and flux to splatter.

However, if you're working with a 2.5-5mm per side LED, things are far
more critical. Small alignment mistakes produce large increases in
thermal resistance. The small dies also tend to "float" on top of the
solder or glue. Rework requires a microscope.

I tried to built a few prototypes using a hot air SMT desoldering gun
and gave up. The air would push the LED's out of position. Because
thermal conductivity is all important, I couldn't use cyanoacrylate
(super-glue) adhesive to keep the dies in place. I eventually
switched to an IR reflow gun, which is much better if I'm careful not
to delaminate the PCB traces with the concentrated heat. If I used to
much solder paste, I had a short under the LED. If I used too little,
the LED would overheat.

The situation is no better in production. Because the LED is balanced
on two blobs of solder paste, there is no guarantee that either IR
reflow would allow the LED to settle parallel to the PCB. The large
thermal mass of the heat sink required a rather long reflow cycle time
for the heat sink to pre-heat, reflow, and cool-down. There are
several tricks required (that I can't talk about) that are required to
properly position the LED.

Testing isn't much easier as the temperature gradient between the LED
and head sink had to be measured with an IR camera. That's where I
discovered that the vias were full of gas and flux bubbles.

I would just love to have the cheap FR4 PCB replaced with a much more
thermally conductive and insulating alumina hybrid. Made thick
enough, it would probably survive the shocks of cycling. I've
proposed hybrids a few times, but that was rejected due to the high
cost, handling problems, and lack of modern hybrid production
machinery. I expect the next generation of LED's to have a built AlN
(aluminum nitride) substrate:
http://ledsmagazine.com/features/10/9/7
which should eventually morph into LED arrays on AlN. AlN has
something like 8 times the thermal conductivity of alumina.


--
Jeff Liebermann
150 Felker St #D
http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558
 




Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

vB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
Idea for future of electronic gear shift. Beware, contains actual"tech" content. James[_8_] Techniques 670 August 7th 12 07:21 PM
dynamo lights with switches and bottle dynamo/generators Keiron[_4_] Techniques 12 September 18th 09 10:30 AM
CATEYE HL-EL320 AS A HEADLAMP Andre Jute[_2_] Techniques 7 March 19th 09 11:00 PM
will this headlamp be ok for muniing? wobbling bear Unicycling 22 October 22nd 05 06:51 AM
Headlamp ist Scheinwerfer buzzgun General 2 August 10th 05 05:10 PM


All times are GMT +1. The time now is 06:42 AM.


Powered by vBulletin® Version 3.6.4
Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.
Copyright ©2004-2024 CycleBanter.com.
The comments are property of their posters.