OT(maybe): LED Lighting?

Ok guys I was thinking about taking quite a few of those 10,000mcd LED's and
putting them in a steel plate to make a light fixture. My first question
is, will this work? Basically what I was thinking is this. Take a sheet of
metal and drill holes in it for the LED to stick through, then I could
either take a thin piece of wire and run it across all the same polarity
contacts on the LED and then run power to those, or maybe just weld the
LED's to each other? Anyways after that I would coat them in silicon and
put another piece of metal on top to protect them. Kind of sandwiching them.
Any thoughts? Also power is a concern of mine, as in how much? And will it
be safe?

TIA
_Chris

Chris Stolworthy wrote:

Ok guys I was thinking about taking quite a few of those 10,000mcd
LED's and putting them in a steel plate to make a light fixture.
My first question is, will this work? Basically what I was
thinking is this. Take a sheet of metal and drill holes in it for
the LED to stick through, then I could either take a thin piece of
wire and run it across all the same polarity contacts on the LED
and then run power to those, or maybe just weld the LED's to each
other? Anyways after that I would coat them in silicon and put
another piece of metal on top to protect them. Kind of sandwiching
them. Any thoughts? Also power is a concern of mine, as in how
much? And will it be safe?

Probably won't work. If you wire them in series any failure kills
the whole system. If you put them in parallel the mismatches will
steal all the current to a subset. In either case you have to
control the actual current.

--
Chuck F ) )
Available for consulting/temporary embedded and systems.
http://cbfalconer.home.att.net USE worldnet address!

"CBFalconer" wrote in message
...
Chris Stolworthy wrote:

Ok guys I was thinking about taking quite a few of those 10,000mcd
LED's and putting them in a steel plate to make a light fixture.
My first question is, will this work? Basically what I was
thinking is this. Take a sheet of metal and drill holes in it for
the LED to stick through, then I could either take a thin piece of
wire and run it across all the same polarity contacts on the LED
and then run power to those, or maybe just weld the LED's to each
other? Anyways after that I would coat them in silicon and put
another piece of metal on top to protect them. Kind of sandwiching
them. Any thoughts? Also power is a concern of mine, as in how
much? And will it be safe?

Probably won't work. If you wire them in series any failure kills
the whole system. If you put them in parallel the mismatches will
steal all the current to a subset. In either case you have to
control the actual current.

--
Chuck F ) )
Available for consulting/temporary embedded and systems.
http://cbfalconer.home.att.net USE worldnet address!

Ouch so you mean that I would have to run current to each of them
individually if I wanted this to work? Doesn't really sound too feasible to
me then. Mostly because I figured that I would need ALOT of LED's to put
off much light. Does anyone know how many 10000mcd LED's it would take to
equate to say a 60w bulb?

"Chris Stolworthy" wrote in message
...
Ok guys I was thinking about taking quite a few of those 10,000mcd LED's
and
putting them in a steel plate to make a light fixture. My first question
is, will this work? Basically what I was thinking is this. Take a sheet
of
metal and drill holes in it for the LED to stick through, then I could
either take a thin piece of wire and run it across all the same polarity
contacts on the LED and then run power to those, or maybe just weld the
LED's to each other? Anyways after that I would coat them in silicon and
put another piece of metal on top to protect them. Kind of sandwiching
them.
Any thoughts? Also power is a concern of mine, as in how much? And will
it
be safe?


although they only take 2 volts or so...and thus safe

you'd probably need 1000 or them or more

if you used a 5 volt supply, you could put a resistor in series with each
one

(200 ohms or so) but it seems like a bit too much work to be practical

Chris Stolworthy wrote:
"CBFalconer" wrote in message
...

Chris Stolworthy wrote:

Ok guys I was thinking about taking quite a few of those 10,000mcd
LED's and putting them in a steel plate to make a light fixture.
My first question is, will this work? Basically what I was
thinking is this. Take a sheet of metal and drill holes in it for
the LED to stick through, then I could either take a thin piece of
wire and run it across all the same polarity contacts on the LED
and then run power to those, or maybe just weld the LED's to each
other? Anyways after that I would coat them in silicon and put
another piece of metal on top to protect them. Kind of sandwiching
them. Any thoughts? Also power is a concern of mine, as in how
much? And will it be safe?

Probably won't work. If you wire them in series any failure kills
the whole system. If you put them in parallel the mismatches will
steal all the current to a subset. In either case you have to
control the actual current.

--
Chuck F ) )
Available for consulting/temporary embedded and systems.
http://cbfalconer.home.att.net USE worldnet address!


Ouch so you mean that I would have to run current to each of them
individually if I wanted this to work? Doesn't really sound too feasible to
me then. Mostly because I figured that I would need ALOT of LED's to put
off much light. Does anyone know how many 10000mcd LED's it would take to
equate to say a 60w bulb?


There are a lot of LED traffic lights using an array of LEDs.
Series/parallel connections work, since the current is controlled by the
series resistor(s), and not by any individual LED. I'm sure a traffic
signal is a lot brighter than a 60W bulb! Try a Google search.

Virg Wall
--
A foolish consistency is the
hobgoblin of little minds,........
Ralph Waldo Emerson
(Microsoft programmer's manual.)

"Chris Stolworthy" wrote:
Ok guys I was thinking about taking quite a few of those 10,000mcd LED's and
putting them in a steel plate to make a light fixture. My first question
is, will this work? Basically what I was thinking is this. Take a sheet of
metal and drill holes in it for the LED to stick through, then I could
either take a thin piece of wire and run it across all the same polarity
contacts on the LED and then run power to those, or maybe just weld the
LED's to each other? Anyways after that I would coat them in silicon and
put another piece of metal on top to protect them. Kind of sandwiching them.
Any thoughts? Also power is a concern of mine, as in how much? And will it
be safe?

What is the maximum current rating for the LED's you are planning on using, how many LED's will you be
using, and at what voltage do you wish to run them at?

Jon

On Thu, 27 May 2004 10:12:22 -0600, "Chris Stolworthy"
wrote:


Probably won't work. If you wire them in series any failure kills
the whole system. If you put them in parallel the mismatches will
steal all the current to a subset. In either case you have to
control the actual current.

--
Chuck F ) )
Available for consulting/temporary embedded and systems.
http://cbfalconer.home.att.net USE worldnet address!

Ouch so you mean that I would have to run current to each of them
individually if I wanted this to work?

No, that is only necessary if you want a failure of a single LED to only
cause that one LED to stop "shining"... ever seen those cheap christmas
lights were one goes out and the whole thing does work? Then there are
other light strands where one goes out and 1/3 of the set isn't lit, and
others where one light goes out and it's only that one light out... same
principle, the number of lights in each series dictates how many would go
out when one fails, and each series needs it's own current limiter of one
form or another (a resistor in the simplest form). It's all a matter of
how you design it, how difficult or costly it is to replace one, and the
designed and/or expected failure rate. In other words, there's nothing
wrong with putting a few in series but it's probably not a good idea to
put ALL in series unless the main goal is absolute lowest cost, simplest
construction... which many people would scoff at but see the link I
provided below, with LEDs @ $0.14 each, and it's not too expensive to just
make a few replacement modules and swap them IF necessary.


Doesn't really sound too feasible to
me then. Mostly because I figured that I would need ALOT of LED's to put
off much light. Does anyone know how many 10000mcd LED's it would take to
equate to say a 60w bulb?



First of all, LEDs are only more efficient than alternatives for
spot-lighting, like a flashlight. Fluorescents can be up to (roughly)
twice as efficient when wider coverage is needed. Many reports of an
LED's great efficiency is ignoring the focused beam, resulting light
intensity in measurements, and not investigating (or being able to
measure) total light output of a fluorescent easily.

The question of how to construct the circuit can depend on a lot of
things, like the amount of space, what materials you have, how hard you're
driving the LEDs (long long they're expected to last vs. the light
output), and the replacement method if/when they fail.

If you really wanted to get fancy with this you could also use an LED
driver IC, there are quite a few but as an example,
http://www.maxim-ic.com/quick_view2.cfm/qv_pk/3756
However, when buying a large quantity of LEDs in bulk there might be no
cost effectiveness to using such a driver. Then again, it could matter
how small you want this light, since it may be more costly per MCD output
to use 10K LEDs than lower MCD value but more LEDs. Likewise, driving
them with lower current you'd get lower output at higher cost, but their
lifespan may be multiple times higher. In other words, using a 10K MCD
LED doesn't _necessarily_ mean that for some designs or uses you'd WANT to
get 10K MCD out of one.

Superbright LED prices have dropped a lot in the past few years, making it
possible that the slight savings in time or support part costs could
outweigh the value of having each LED independantly current-limited.

You could have 5 parallel sets of 10 in a series, for example, then when
one LED goes out you only have that one series out, a 20% light
reduction. If you construct this light so it's somewhat modular, you can
replace strings of LEDs if necessary, by just plugging in a replacement
module, similar to replacing any other light bulb. Then you still have a
working light and can investigate which LED in that series failed if you
have the inclination to do so, at your leisure.

It's not really necessary to use those steel parts you've mentioned. If
driving the LEDs hard you may want their leads on soldered to a
significant metal (whatever) to 'sink the heat, but that could as easily
be a copper-clad PCB as pieces of steel, were're not talking about
heatsinks of any external construction, just using the circuit itself as
the heatsink is usually sufficient. As for using a 2nd piece of metal to
protect the LEDs, it's probably not necessary unless this is going to be
mounted on some outdoor, high-velocity vehicle. LEDs are pretty sturdy,
relatively speaking.

I'll link a couple pictures to give you some ideas, alternatives,
whatever...

http://www.allelectronics.com/cgi-bi...-16&type=store
http://www.allelectronics.com/cgi-bi...S-4&type=store
http://www.directleds.com/product_in...products_id=28

"kony" wrote in message
...
On Thu, 27 May 2004 10:12:22 -0600, "Chris Stolworthy"
wrote:


Probably won't work. If you wire them in series any failure kills
the whole system. If you put them in parallel the mismatches will
steal all the current to a subset. In either case you have to
control the actual current.

--
Chuck F ) )
Available for consulting/temporary embedded and systems.
http://cbfalconer.home.att.net USE worldnet address!

Ouch so you mean that I would have to run current to each of them
individually if I wanted this to work?

No, that is only necessary if you want a failure of a single LED to only
cause that one LED to stop "shining"... ever seen those cheap christmas
lights were one goes out and the whole thing does work? Then there are
other light strands where one goes out and 1/3 of the set isn't lit, and
others where one light goes out and it's only that one light out... same
principle, the number of lights in each series dictates how many would go
out when one fails, and each series needs it's own current limiter of one
form or another (a resistor in the simplest form). It's all a matter of
how you design it, how difficult or costly it is to replace one, and the
designed and/or expected failure rate. In other words, there's nothing
wrong with putting a few in series but it's probably not a good idea to
put ALL in series unless the main goal is absolute lowest cost, simplest
construction...

One thing you would have to take in to account is that LED's require
a bias voltage to make them work, I think it is about 1/2 a volt but I will
look it up as that sounds quite large.
Now I am thinking that if they are in series you would have to add the bias
voltages togeather so that say 100 leds would require 50volts for them
to work!!!.
I must admit I am a little unsure about this and possibly once you applied
over 1/2 a volt all the LED's would work?
Having said that I do go for the former theory where you add the biases
togeather.
Indeed that is correct.
http://www.kpsec.freeuk.com/components/led.htm
(very informative page)
Also the bias vooltage given there is 2V which is larger than I assumed
earlier.
Also from
http://www.maxim-ic.com/appnotes.cfm...te_number/1883
"LEDs have a MTBF (mean time between failures) usually in the range of
100,000 to over 1,000,000 hours. This is a long time for continuous
operation, considering a year is 8760 or 8784 hours. In practice, the useful
measure of LED lifetime is its half-life, that is an LED is deemed to have
reached the end of its life when the light output falls off to half the
original."

Also.
"Incandescent traffic lamps draw somewhere between 75W and 150W, depending
on size (20cm or 30cm) and color (due to differences in the transmissivity
of the red, green, and orange filters used). LED traffic lamps draw around
7W - 15W, and can be replaced every 5 years instead of every year for
incandescents."

So the LED lights require 5 times less maintainance.


which many people would scoff at but see the link I
provided below, with LEDs @ $0.14 each, and it's not too expensive to just
make a few replacement modules and swap them IF necessary.


Doesn't really sound too feasible to
me then. Mostly because I figured that I would need ALOT of LED's to put
off much light. Does anyone know how many 10000mcd LED's it would take
to
equate to say a 60w bulb?



First of all, LEDs are only more efficient than alternatives for
spot-lighting, like a flashlight. Fluorescents can be up to (roughly)
twice as efficient when wider coverage is needed. Many reports of an
LED's great efficiency is ignoring the focused beam, resulting light
intensity in measurements, and not investigating (or being able to
measure) total light output of a fluorescent easily.

The question of how to construct the circuit can depend on a lot of
things, like the amount of space, what materials you have, how hard you're
driving the LEDs (long long they're expected to last vs. the light
output), and the replacement method if/when they fail.

If you really wanted to get fancy with this you could also use an LED
driver IC, there are quite a few but as an example,
http://www.maxim-ic.com/quick_view2.cfm/qv_pk/3756
However, when buying a large quantity of LEDs in bulk there might be no
cost effectiveness to using such a driver. Then again, it could matter
how small you want this light, since it may be more costly per MCD output
to use 10K LEDs than lower MCD value but more LEDs. Likewise, driving
them with lower current you'd get lower output at higher cost, but their
lifespan may be multiple times higher. In other words, using a 10K MCD
LED doesn't _necessarily_ mean that for some designs or uses you'd WANT to
get 10K MCD out of one.

Superbright LED prices have dropped a lot in the past few years, making it
possible that the slight savings in time or support part costs could
outweigh the value of having each LED independantly current-limited.

You could have 5 parallel sets of 10 in a series, for example, then when
one LED goes out you only have that one series out, a 20% light
reduction. If you construct this light so it's somewhat modular, you can
replace strings of LEDs if necessary, by just plugging in a replacement
module, similar to replacing any other light bulb. Then you still have a
working light and can investigate which LED in that series failed if you
have the inclination to do so, at your leisure.

It's not really necessary to use those steel parts you've mentioned. If
driving the LEDs hard you may want their leads on soldered to a
significant metal (whatever) to 'sink the heat, but that could as easily
be a copper-clad PCB as pieces of steel, were're not talking about
heatsinks of any external construction, just using the circuit itself as
the heatsink is usually sufficient. As for using a 2nd piece of metal to
protect the LEDs, it's probably not necessary unless this is going to be
mounted on some outdoor, high-velocity vehicle. LEDs are pretty sturdy,
relatively speaking.

I'll link a couple pictures to give you some ideas, alternatives,
whatever...


http://www.allelectronics.com/cgi-bi...-16&type=store

http://www.allelectronics.com/cgi-bi...S-4&type=store
http://www.directleds.com/product_in...products_id=28