Brief note: modification to create super-bright tail light

I just modified my King Sword KS-303F stock 3 LED flasher into what may
be one of the brightest LED taillights around, and would like to share
how I did it with the group to see if others would be interested in
doing the same/similar stuff. The idea of doing the modification came
about from trying to salvage an old KS light that had a bad switch; it
then dawned on me that the light can be greatly enhanced with the
latest LEDs.

Basically, the King Sword KS-303F is a very simple light. It comes with
3 standard 5 mm deep red LEDs, and is a flash only device. All the LEDs
are connected in parallel, and the switch apparently contains a
built-in flasher. No obvious dropping resistors were noted, and I
believe none are built in (more on this later).

With a soldering gun, I removed the three stock LEDs, and replaced them
with 3 5 mm red THC (High Current, High Candela, High Capability) LEDs
from lsdiodes.com (Note: US postal mail is for some reason extra slow
these days, and the LEDs took 9 days to arrive). For the two additional
positions at the ends of the board, I added 2 3 mm LEDs from
lsdiodes.com as well. One should note that those positions will take 5
mm LEDs; however, 5 mm LEDs will prevent one from replacing the lens
cap. All in all, the upgrade gave me a whopping 51000 mcd of light;
presumably the lens spreads enough light at the edges so that the light
is visible from the side as well. It should be noted that the batteries
used to power the lights need to have a combined voltage of 2.4V, not 3
V. LEDs are very voltage sensitive, and one cannot safely overvolt a
LED and expect much additional output. Rather overvolting an LED by a
little bit causes a HUGE increase in current, and burns the LED out. My
guess is that the light as shipped has the flasher set up in such a way
that the momentary overvoltage doesn't hurt the stock LEDs. Anyhow, 2.4
V total is readily obtained with a set of rechargeable NiMH AA cells
(AA NiCads would work nicely too), and the additional 3 mm LEDs are
rated at 2.4V max for the 5500 mcd variety (the 6000 mcd ones have a
reg rating of 1.7 V, which is far too low).

After closing up the lens and attaching the light to my bike (the screw
and the tab nicely fit the standard rear reflector bracket), I found
that the light is very bright as expected. Somewhat surprisingly, the
light went from a deep red to a orangeish-red. Though initially a bit
annoyed, since the LEDs were sold as "red," I soon found myself liking
the tinge of orange. Aren't red traffic lights tinged orange to enable
color-blind people to see them anyhow? It seems the light is now
comparable in brightness to some car taillights, at a cost of US$8.12
including shipping and handling for the LEDs, and another $8 or so for
the original light. Current draw is presumably greatly increased (if it
weren't, then the new LEDs probably wouldn't of all lit up as they did
with a limited current, though I didn't rigorously measure the
currents), so I estimate the run time is now 12 hrs or so vs the
standard 200, but that's plenty for me.

This makes me wonder how it seems most taillight manufacturers are
reluctant to share specs about their lights besides offering useless
adjectives such as "3 super-bright LEDs" and the like making it
impossible to objectively compare models. Would anybody have comments
on this or similar taillight modifications, and/or know how close I am
to some "ultimate" taillight? What about numerical brightness
specifications for taillights on the road? Thanks.

Later,
Nelson Chen

Jan 21, 8:48 pm show options

Newsgroups: rec.bicycles.tech
From: - Find messages by this author
Date: 21 Jan 2005 20:48:52 -0800
Local: Fri, Jan 21 2005 8:48 pm
Subject: Brief note: modification to create super-bright tail light
Reply | Reply to Author | Forward | Print | Individual Message | Show
original | Report Abuse

I just modified my King Sword KS-303F stock 3 LED flasher into what may
be one of the brightest LED taillights around, and would like to share
how I did it with the group
SNIP


Hi, Nelson, it sounds like you may have a good idea, here.

I am _NO_ kind of electronics person but I did stumble into this
website the other day:-
http://www.superbrightleds.com/led_info.htm

There is a link on there, 'Why do I need a resistor with an LED?'

Their explanation goes way over my head but I thought it might be
relevant to your project.

Lewis.

******

Hi,
I saw the webpage too, and have read up on other pages about using a
resistor as well. Basically, the reason people say a resistor is needed
is that diodes (LEDs included) are highly non-linear in the amount of
resistance offered depending on the voltage applied. So, if a LED is
rated at 2.4 V for a current of say 20 mA, giving it 2.5 V would cause
the current to shoot up drastically say to 50 mA or so, and may well
kill it. That is why I noted that for my modified light, I make a point
to use a set of AA NiMH or NiCad batteries to power the thing. Those
batteries have a combined maximum voltage of 2.4 V (which does not
exceed the LED ratings), therefore keeping me safe. Resistors waste
energy that could be turned into light; IMO if one carefully notes the
voltage source and is sure the voltage source wouldn't exceed the given
values, one can safely hook up an LED that way. I guess I'll see in a
little bit whether my mod lasts or not. If the LEDs die in a little bit
we'll know there was a problem.

Later,
Nelson

Resistors waste
energy that could be turned into light; IMO if one carefully notes the
voltage source and is sure the voltage source wouldn't exceed the given
values, one can safely hook up an LED that way. I guess I'll see in a
little bit whether my mod lasts or not. If the LEDs die in a little bit
we'll know there was a problem.

Other than the diode drop, a diode is basically a short circuit. Without a
resistor to limit current, what do you suppose your battery life will be?

wrote:
So, if a LED is
rated at 2.4 V for a current of say 20 mA, giving it 2.5 V would cause
the current to shoot up drastically say to 50 mA or so, and may well
kill it. That is why I noted that for my modified light, I make a point
to use a set of AA NiMH or NiCad batteries to power the thing. Those
batteries have a combined maximum voltage of 2.4 V (which does not
exceed the LED ratings), therefore keeping me safe.

The nominal voltage of NiMH and NiCd cells is 1.2 V, but when freshly
charged the actual voltage can be somewhat higher. I've seen voltages
of about 1.35/cell in my GPS using NiMH initially and staying above
1.25 for quite a long time. There's a typical discharge curve at:
http://data.energizer.com/PDFs/nh15.pdf which agrees with what I've
observed.

Well, for a given voltage, a real diode has an actual resistance0; as
such, it's not a short circuit. Yes, for an *ideal* diode it would be a
short, and would be toast in short order. In the specs, it's noted that
for a given voltage, there is a finite current that flows for a real
device. My LEDs are rated 80 mA each for the high-powered 5 mm ones,
and 20 mA each for the 3 mm ones.

Later,
Nelson Chen

You have a good point. My voltmeter wasn't that precise in terms of
measuring the voltage to that degree. The 3 mm LEDs are rated for a
maximum of 2.4 V, the 5 mm ones can take just a little bit more. FWIW,
my light is a simple flasher, and the original LEDs were hooked up so
that they could take even 2 regular AA batteries for a nominal total of
3V. Was the light a fundamentally flawed design, or was it designed so
that with the flashing, the momentary current spikes are such that
there is time for the LEDs to cool b/t flashes and thus keep things
safe? Or did it use deep red LEDs *rated* for 3 V (do those even
exist)? It would be interesting to see the complete circuit diagrams of
the original light. However for some reason manufacturers tend to not
release that information, just as they don't release information on
their light brightnesses. We'll see in a few days or weeks whether my
setup really works, or whether my LEDs would be toast. But anyhow,
thanks for the note; if this is the case, then my scheme can only work
for this flashing light, not a steady one.

Later,
Nelson Chen

While a resistor does waste some energy, it is not nearly as much as running
two or three times the current. As far as being a "dead short," since a
diode is converting electric power into light, that is impossible.

On 21 Jan 2005 20:48:52 -0800, wrote:

I just modified my King Sword KS-303F stock 3 LED flasher into what may
be one of the brightest LED taillights around, and would like to share
how I did it with the group to see if others would be interested in
doing the same/similar stuff. The idea of doing the modification came
about from trying to salvage an old KS light that had a bad switch; it
then dawned on me that the light can be greatly enhanced with the
latest LEDs.

Basically, the King Sword KS-303F is a very simple light. It comes with
3 standard 5 mm deep red LEDs, and is a flash only device. All the LEDs
are connected in parallel, and the switch apparently contains a
built-in flasher. No obvious dropping resistors were noted, and I
believe none are built in (more on this later).

With a soldering gun, I removed the three stock LEDs, and replaced them
with 3 5 mm red THC (High Current, High Candela, High Capability) LEDs
from lsdiodes.com (Note: US postal mail is for some reason extra slow
these days, and the LEDs took 9 days to arrive). For the two additional
positions at the ends of the board, I added 2 3 mm LEDs from
lsdiodes.com as well. One should note that those positions will take 5
mm LEDs; however, 5 mm LEDs will prevent one from replacing the lens
cap. All in all, the upgrade gave me a whopping 51000 mcd of light;
presumably the lens spreads enough light at the edges so that the light
is visible from the side as well. It should be noted that the batteries
used to power the lights need to have a combined voltage of 2.4V, not 3
V. LEDs are very voltage sensitive, and one cannot safely overvolt a
LED and expect much additional output. Rather overvolting an LED by a
little bit causes a HUGE increase in current, and burns the LED out. My
guess is that the light as shipped has the flasher set up in such a way
that the momentary overvoltage doesn't hurt the stock LEDs. Anyhow, 2.4
V total is readily obtained with a set of rechargeable NiMH AA cells
(AA NiCads would work nicely too), and the additional 3 mm LEDs are
rated at 2.4V max for the 5500 mcd variety (the 6000 mcd ones have a
reg rating of 1.7 V, which is far too low).

After closing up the lens and attaching the light to my bike (the screw
and the tab nicely fit the standard rear reflector bracket), I found
that the light is very bright as expected. Somewhat surprisingly, the
light went from a deep red to a orangeish-red. Though initially a bit
annoyed, since the LEDs were sold as "red," I soon found myself liking
the tinge of orange. Aren't red traffic lights tinged orange to enable
color-blind people to see them anyhow? It seems the light is now
comparable in brightness to some car taillights, at a cost of US$8.12
including shipping and handling for the LEDs, and another $8 or so for
the original light. Current draw is presumably greatly increased (if it
weren't, then the new LEDs probably wouldn't of all lit up as they did
with a limited current, though I didn't rigorously measure the
currents), so I estimate the run time is now 12 hrs or so vs the
standard 200, but that's plenty for me.

This makes me wonder how it seems most taillight manufacturers are
reluctant to share specs about their lights besides offering useless
adjectives such as "3 super-bright LEDs" and the like making it
impossible to objectively compare models. Would anybody have comments
on this or similar taillight modifications, and/or know how close I am
to some "ultimate" taillight? What about numerical brightness
specifications for taillights on the road? Thanks.

Later,
Nelson Chen

Nothing so technical to offer, but a comment on the color. Yes, Orange works
much better for the most common color vision deficiencies and color blindness.
Red may as well be brown or hunter green. Not exactly stand-out colors.

Since we're talking about some 20% of the maile population and any one of them
would regret inadvertently running you over, it's worthwhile to have a tail
light that accomodates their perceptive limitation.

Ron

RonSonic wrote:

(snip light-building details)

Nothing so technical to offer, but a comment on the color. Yes, Orange works
much better for the most common color vision deficiencies and color blindness.
Red may as well be brown or hunter green. Not exactly stand-out colors.

Since we're talking about some 20% of the maile population and any one of them
would regret inadvertently running you over, it's worthwhile to have a tail
light that accomodates their perceptive limitation.

Ron

A couple of quibbles: color blindness is quite different from color
deficiencies as the former would imply non-functioning cones in the
retina, leaving only rods. True color blindness is exceedingly rare and
would also involve debilitating acuity loss.

It may well be that 20% of the male population has some degree of
deficiency in color perception, but such defects have a wide range of
severity, and it is a wild exaggeration to imply that those men cannot
see red, or that red looks the same to them as brown or hunter green.
Otherwise there would be carnage at every red signal light.

--
Ted Bennett
Portland, OR