bike headlamp LED, 4.5V (AA), resistor in series needed?

[Mike S]

I was given a broken LED light (on a headband) that takes 3 AA
batteries, can't find the lumen spec as there is no company name or part
number on the ct. board. If I just run the 3 batteries directly to the
LED will the batteries drain too fast or will I damage the LED? I don't
know if I have to limit the current for either case.

[mike]

It depends...
What components are on the circuit board?
Does it look like a voltage or current regulator?

A typical cheapo light might use crap "Heavy Duty"
batteries and expect the series resistance of the battery
to limit the current. Putting Alkaline cells, which
have lower voltage, can overheat the cheapest ones.

Others have sophisticated control circuitry.

There's a lot you can do to figger it out, but the nature
of your question suggests that you don't have the tools
to do that...or you'd have done it already.

I put an 18650 lithium battery into a 9-LED
Harborfreight 3X-AA cell light.
4.2V is less than 4.5, right?
Well, it was bright...but the LED's burned out
in short order. I put 1.5 ohms in series in the next
one because that's what I had. Has been working fine.

You can actually calculate the best resistor, but
since you don't know the specs of the led, there's nothing
to put into the equation. And the range of voltage over
the life of the battery
can exceed the drop across the resistor at full charge.

Safest thing is to put something like 10 ohms in series with
a full battery and drop the value until it gets as bright
as you dare.

Sometimes, you can use google images to find lights that look
like yours and go searching down those paths for representative
specs.

[Mike S]

Thanks, I looked through a lot of google images and ebay head lights and
couldn't find this model. There was a component of some sort on the ct.
board that had a large dollop of very tough epoxy completely covering
it, I would have destroyed it getting the cement off, after about 10
minutes of trying to carefully shave/chip it away I gave up.
There were 3 resistors and a transistor on the m/b in addition to the
mystery component, and since something was malfunctioning and reflowing
the solder joints didn't help I couldn't measure the voltage at the LED.
There were 4 smaller LEDs on the board as well, complicating the ct.
enough that without knowing the mystery component I didn't know how to
sort it out. I like your approach, starting high, try decreasing values
of series R until I have decent brightness, then see how the batteries
and LED hold up. Thanks for your thoughts.

[Clifford Heath]

You wouldn't have learned anything even if you'd successfully
decapped the chip. It's what they call COB (Chip On Board)
technology; you would have seen a bare silicon wafer with
bonds to the PCB.

These chips usually provide dimming and flashing function.
If there's no external inductor - and I'm betting there's
not - then the device is relying on bang-bang control for
dimming. Peak current is limited by the battery's internal
resistance, average current by the on/off duty cycle.

The best way to estimate the power capability is to look
at how much heat can be removed from the LED. If it's not
got much metal attached, keep the current below 100mA and
you'll be safe and still get very good brightness.

Bear in mind, that ten times the current only doubles the
perceived brightness. I have an 8W headlamp for night hiking,
but I usually use it on the low setting; under 1W. So if you
run it below its limit, it'll still be useful.

Clifford Heath

[tabb...@gmail.com]

If you had uncovered the chip your light would be dead for sure.
4.5v is far too high to run an LED direct - at least with most LEDs.
If it's a 20mA LED, which I can't assume, a new battery of 4.65v would need the R to drop about 1.15v at 20mA = 57 ohms. A near flat battery at 3.5v would want a resistor of 0 ohms. And there's the problem with running an LED off a battery & resistor. What's really wanted is a constant current circuit. There are simple ways to do nearly that, but it's not as simple as a resistor.


NT

[Jeff Liebermann]

On Sun, 15 Apr 2018 16:39:14 -0700, Mike S <ms...@yahoo.com> wrote:

>I was given a broken LED light (on a headband) that takes 3 AA
>batteries, can't find the lumen spec as there is no company name or part
>number on the ct. board.

I'll assume that there was only one LED in the headlight. Lumens are
fairly easy to guess. The commodity LED du jure is the Cree T6
series. It produces about 100 lumens/watt of power consumed (minus
whatever the optics attenuates). There are LED's with higher efficacy
available, but you won't find them in commodity lights. The LED will
be mounted on a COB (chip on board) which also acts as heat sink.
Somewhere in the package, either on the COB or on the on/off/dim
switch, will be some electronics to control the light. These also
include a current source to regulate the current drawn by the LED. You
should not need any additional current limiting resistors. On the
really cheap junk LED lights, there might be one or two tiny chip
resistors on the COB to provide some current limiting. Three AA
alkaline cells in series is about typical for powering a low end,
single LED light.

>If I just run the 3 batteries directly to the
>LED will the batteries drain too fast or will I damage the LED? I don't
>know if I have to limit the current for either case.

I don't know. Take it apart and see what's inside. Most battery
compartments include a label or molded description of the type of
battery to install. Look for it. Unless you were given a bag of
parts, it's unlikely that you'll need to add anything to make the
light work. Start with three alkaline cells and measure the current.
If the LED burns about 1 watt or less as in:
voltage_across_LED * current through LED = watts
it should work.

Also, while it might be possible that your headlight runs on LiIon
cells, I doubt it. The lights that use AA size cells (14500) tend to
use alkaline or NiMH, while the one's I like use one (or two in
parallel) 18650 LiIon cells. I haven't seen any 14500 LiIon cells
used in headlights yet, although such a light might be possible.

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

[Mike S]

Thanks Clifford, I get good brightness using 35 ohms, I'll measure the
current to see where that is. There's no heat sink on the LED. This
doesn't have to be super-bright, it's in addition to a handlebar mounted
light so it's basically to increase visibility to cars, being higher and
moving out of sync with the handlebar light. Once I make sure the
current is < 100mA and install a small toggle switch and I'm good.
Thanks again.

[mike]

What you really want is a resistor that automagically drops
the series resistance as the battery discharges.
Lucky for you, they exist.

They're called incandescent light bulbs.
Try a lamp from a single-cell incandescent pen light.
I just tried one. Looks like the sweet spot is around 150ma.
But the lamp was unmarked, so I can't be sure what I had.
Gotta test 'em.

Also might try a lamp from a string of tiny incandescent Christmas
tree lights. I can't find one, but, as I recall, those were about 3V
nominal. Might be too high.

Not enough current, parallel some. Too much current,
series some. Could even switch to change brightness of the LED.

Are we having fun yet?

[Mike S]

I was thinking of building a simple voltage/current regulator, and
wondering how much current would be spent in the ct. This is another
idea I never thought of. I found a small pot in an old blown up computer
pwr supply I had saved, was thinking of using it to vary the brightness
as the batteries were drained. It's funny how complicated a simple head
lamp can become.

[whit3rd]

Yeah, for battery powered gizmos, that's important.
The easy solution is a funny chip that has a builtin
current regulator, and has (as well as current limit) thermal
protection (when mounted thermally to the LED parts).

Look on eBay for AMC7135 (about 1W delivered to an LED, use two or
three in parallel for higher power lamps).. This is a good solution
for a 4.2V battery input, which a power-wasting resistor can never
improve on. Alas, battery and LED options must be driven by the
choice of such a chip.

[Mike S]

$3 for 10!
10pcs AMC7135 350mA LED driver SOT-89
Thank you!

[~misfit~]

https://www.aliexpress.com/item/10PCS-AMC7135-350mA-2-7V-6V-LED-driver-high-LED-driver-chip-New-Original-Free-Shipping/32811284788.html

Just over a buck for 10 there - including shipping.
--
Shaun.

"Humans will have advanced a long, long way when religious belief has a cozy
little classification in the DSM*."
David Melville (in r.a.s.f1)
(*Diagnostic and Statistical Manual of Mental Disorders)

[Mike S]

On 5/5/2018 5:39 PM, ~misfit~ wrote:
> Once upon a time on usenet Mike S wrote:
>> On 4/29/2018 3:50 PM, whit3rd wrote:
>>> On Sunday, April 15, 2018 at 9:30:31 PM UTC-7, Mike S wrote:
>>>> On 4/15/2018 8:30 PM, mike wrote:
>>>
>>>> I was thinking of building a simple voltage/current regulator, and
>>>> wondering how much current would be spent in the ct.
>>>
>>> Yeah, for battery powered gizmos, that's important.
>>> The easy solution is a funny chip that has a builtin
>>> current regulator, and has (as well as current limit) thermal
>>> protection (when mounted thermally to the LED parts).
>>>
>>> Look on eBay for AMC7135 (about 1W delivered to an LED, use two or
>>> three in parallel for higher power lamps).. This is a good solution
>>> for a 4.2V battery input, which a power-wasting resistor can never
>>> improve on. Alas, battery and LED options must be driven by the
>>> choice of such a chip.
>> $3 for 10!
>> 10pcs AMC7135 350mA LED driver SOT-89
>> Thank you!
>
> https://www.aliexpress.com/item/10PCS-AMC7135-350mA-2-7V-6V-LED-driver-high-LED-driver-chip-New-Original-Free-Shipping/32811284788.html
>
> Just over a buck for 10 there - including shipping.

Just ordered them, thank you. Can you recommend a circuit, for 2 AA
batteries? I found several with a quick search, if you have one that has
a very efficient use of current I would like to see it.