Current regulation confusion

[Michael]

I've been toying with these Luxeon white leds and the data calls for a
current regulated supply.

I'll assume the reasoning for any LED to require a current regulated supply
is for a consistent, unwavering brightness level. (?)

So, I went ahead and built an LM317 in current regulated mode with a couple
of these Luxeon's in series as per the application info and all seems well
(except one seems a bit dimmer....that could be another issue....they talk
about bin numbers alot which could be "batch differences"?).

Fact is, every application shows these LED's wired in series (yuck?) with
"current regulated power supply".

Anyway, my question is this....

If the LED is the only thing in the circuit and is fed through a series
limiting resistor, the current is going to be constant anyway, right? I mean
there's nothing else in the circuit to affect it.

If the power supply was robust enough, wouldn't it be the same thing to use
an lm317 as a voltage regulator and run
my (2) led's with resistors in parallel to the supply buss?

I don't get it....the current is constant either way. ?

I'd rather do the latter because I wouldn't need a higher voltage for the
series supply.

If I did parallel, I could just whip out a good old 6.3vac xformer.

Thanks.

 

[Chris]

I've been toying with these Luxeon white leds and the data calls for a
current regulated supply.

I'll assume the reasoning for any LED to require a current regulated supply
is for a consistent, unwavering brightness level. (?)

So, I went ahead and built an LM317 in current regulated mode with a couple
of these Luxeon's in series as per the application info and all seems well
(except one seems a bit dimmer....that could be another issue....they talk
about bin numbers alot which could be "batch differences"?).

Fact is, every application shows these LED's wired in series (yuck?) with
"current regulated power supply".

Anyway, my question is this....

If the LED is the only thing in the circuit and is fed through a series
limiting resistor, the current is going to be constant anyway, right? I mean
there's nothing else in the circuit to affect it.

If the power supply was robust enough, wouldn't it be the same thing to use
an lm317 as a voltage regulator and run
my (2) led's with resistors in parallel to the supply buss?

I don't get it....the current is constant either way. ?

I'd rather do the latter because I wouldn't need a higher voltage for the
series supply.

If I did parallel, I could just whip out a good old 6.3vac xformer.

Thanks.

Hi, Michael. Luxeon is staying away from all the customer's stuff by
specifying a current source. But how the customer develops a current
source is his problem.

Your feeling that a good voltage source with a series LED to limit
current would work just as well is correct. There are some variations
in LED voltage between LEDs, and also the forward voltage will tend to
decrease as the LED warms up, but if your voltage source is at least
twice the LED forward voltage (6VDC or so for the Vf=3V of the white
LED), you shouldn't have any problem in just using a current-limiting
resistor and being done with it.

Good luck
Chris

 

[TuT]

Hi, Michael. Luxeon is staying away from all the customer's stuff by
specifying a current source. But how the customer develops a current
source is his problem.

Your feeling that a good voltage source with a series LED to limit
current would work just as well is correct. There are some variations
in LED voltage between LEDs, and also the forward voltage will tend to
decrease as the LED warms up, but if your voltage source is at least
twice the LED forward voltage (6VDC or so for the Vf=3V of the white
LED), you shouldn't have any problem in just using a current-limiting
resistor and being done with it.

Good luck
Chris

You would however be well advised to provide each LED with its own series
resistor for this approach, since variations in forward voltage in the LEDS
could otherwise result in one LED hogging most of the current.

 

[ehsjr]

You would however be well advised to provide each LED with its own series
resistor for this approach, since variations in forward voltage in the LEDS
could otherwise result in one LED hogging most of the current.

Read the last line of what Chris said.

 

[Michael]

You would however be well advised to provide each LED with its own series

resistor for this approach, since variations in forward voltage in the
LEDS could otherwise result in one LED hogging most of the current.

Exactly.

That's why the "yuck" when I mentioned series.

Why do they push the series arrangement so much? You should see there data
sheets for this part....unreal matrixes of led's in series/parallel.

Heck, I just want to light 2 of them for a sign!

In fact, the 317 I have now as a current regulator and 2 in series as per
their instructions is showing one of the leds dimmer.....(maybe this bin
stuff they tlak about cures that, really confusing, the bin numbers).

Maybe I should adjust my resistors in my parallel arrangement for the
variation in forward voltage and resultant brightness......we'll see. Maybe
the Luxeons are more critical than the average LED.

Thanks you guys.

 

[[email protected]]

Read the last line of what Chris said.

"Good Luck"?

 

[redbelly]

I've been toying with these Luxeon white leds and the data calls for a
current regulated supply.

I'll assume the reasoning for any LED to require a current regulated supply
is for a consistent, unwavering brightness level. (?)

So, I went ahead and built an LM317 in current regulated mode with a couple
of these Luxeon's in series as per the application info and all seems well
(except one seems a bit dimmer....that could be another issue....they talk
about bin numbers alot which could be "batch differences"?).

Fact is, every application shows these LED's wired in series (yuck?) with
"current regulated power supply".

Anyway, my question is this....

If the LED is the only thing in the circuit and is fed through a series
limiting resistor, the current is going to be constant anyway, right? I mean
there's nothing else in the circuit to affect it.

If the power supply was robust enough, wouldn't it be the same thing to use
an lm317 as a voltage regulator and run
my (2) led's with resistors in parallel to the supply buss?

I don't get it....the current is constant either way. ?

I'd rather do the latter because I wouldn't need a higher voltage for the
series supply.

If I did parallel, I could just whip out a good old 6.3vac xformer.

Thanks.

What you say is completely true for a fixed voltage source, as you get
with a regulator.

However.

For circuits that use a battery, the current can change considerably as
the battery voltage drops. Or, if you were to change to another type
of LED, you'd have to recalculate and change the resistor value.

There are some simple constant-current circuits that are easy to build:

A two-transistor ciruit that I use, except instead of the 2n4403's
shown here I use two 2n3906's (Vbe = 0.66V):
http://wolfstone.halloweenhost.com/Lighting/litled_LightEmittingDiodes.html#LEDTester
(The circuit can be modified to use npn transistors instead of pnp's)

Using a transistor and two diodes. While I don't recommend recharging
standard batteries as shown at this site, it does look like an
easy-to-build current source that should work fine for LED's. Just put
your LED (or LED's) in place of the battery shown here, and then
connect a battery between ground and where it says "12ish V":
http://www.afrotechmods.com/reallycheap/batteries/batts.htm

Regards,

Mark

 

[redbelly]

Using a transistor and two diodes. While I don't recommend recharging
standard batteries as shown at this site, it does look like an
easy-to-build current source that should work fine for LED's. Just put
your LED (or LED's) in place of the battery shown here, and then
connect a battery between ground and where it says "12ish V":
http://www.afrotechmods.com/reallycheap/batteries/batts.htm

Multiple LED's would be connected in series where the battery is drawn.
Oh, and also change the value of R2 to run a current suitable for an
LED.

Mark

 

[ehsjr]

What you say is completely true for a fixed voltage source, as you get
with a regulator.

However.

For circuits that use a battery, the current can change considerably as
the battery voltage drops. Or, if you were to change to another type
of LED, you'd have to recalculate and change the resistor value.

There are some simple constant-current circuits that are easy to build:

He already has the simplest:

-----
+ --+--Vin|LM317|Vout---+
| ----- |
[.1uF] Adj [R]
| | |
Gnd +----------+-----> Constant I output

Only 3 parts. I = 1.2/R

Ed

 

[redbelly]

redbelly wrote:

There are some simple constant-current circuits that are easy to build:

He already has the simplest:

-----
+ --+--Vin|LM317|Vout---+
| ----- |
[.1uF] Adj [R]
| | |
Gnd +----------+-----> Constant I output

Only 3 parts. I = 1.2/R

Ed

So he does.

What about a cap on Vout?

Mark

 

[default]

Anyway, my question is this....

If the LED is the only thing in the circuit and is fed through a series
limiting resistor, the current is going to be constant anyway, right? I mean
there's nothing else in the circuit to affect it.

If the power supply was robust enough, wouldn't it be the same thing to use
an lm317 as a voltage regulator and run
my (2) led's with resistors in parallel to the supply buss?

I don't get it....the current is constant either way. ?

I'd rather do the latter because I wouldn't need a higher voltage for the
series supply.

If I did parallel, I could just whip out a good old 6.3vac xformer.

They generally use series LEDs because it is simpler to have one
current source feed a string of them.

CC has the advantage of keeping the current constant with changes in
the voltage drop due to temperature variations. The data sheet on the
Luxeon, that I'm looking at, doesn't give a chart with temperature and
voltage, but it would be unusual if it wasn't a factor.

And there's nothing holy about constant current - a voltage source
with dropping resistors will be fine. Run them conservatively and
there's no need for a regulated supply - just take into account the
voltage fluctuation of the mains and the ambient temperature you will
be working at and de rate them for temperature and anticipated voltage
fluctuations if need be.

If you are building one of a kind - measure the current and set it
with a resistor. If you are manufacturing many - it is easier to use
CC and not have to worry about the individual led variations in
voltage drop. 3.2 to 4.0 for white.

The White Luxeon Stars are specified with bins- and both the intensity
and color temperature vary with the bin letter. A is highest with
maximum light and coldest color temperature and E is weak and warm.

De rate the LEDs and keep the power supply simple or run them at their
maximum output and regulate the current - and use a good heatsink.

I'm running four Cree red leds in series mounted to a 24 square inch
piece of aluminum with a resistor. At 15 volts in they are close to
their maximum power limit, and my voltage runs between 12 and 15 volts
ambient temperature to about 40 C.

A place calling themselves the ledshoppe in Hong Kong is selling 1.5
watt Luxeon star flashlights for $5 post paid - about the best price
that I've seen. I've bought LEDs from them without problems but not a
Star flashlight.

 

[James Beck]

redbelly wrote:

There are some simple constant-current circuits that are easy to build:

He already has the simplest:

-----
+ --+--Vin|LM317|Vout---+
| ----- |
[.1uF] Adj [R]
| | |
Gnd +----------+-----> Constant I output

Only 3 parts. I = 1.2/R

Ed

So he does.

What about a cap on Vout?

Mark

I would think a cap on Vout would cause instability when running it as a
CC source and possibly make it oscillate or slow the tracking to the
point large current excursions could take place, possibly frying the
load.

Jim

 

[ehsjr]

redbelly wrote:

There are some simple constant-current circuits that are easy to build:

He already has the simplest:

-----
+ --+--Vin|LM317|Vout---+
| ----- |
[.1uF] Adj [R]
| | |
Gnd +----------+-----> Constant I output

Only 3 parts. I = 1.2/R

Ed

So he does.

What about a cap on Vout?

Mark

Not needed.
Ed

 

[Rich Grise]

You would however be well advised to provide each LED with its own
series

Exactly.

That's why the "yuck" when I mentioned series.

Why do they push the series arrangement so much?

Because in a series circuit, every element passes the same current. They
recommend that LEDs be driven with a constant-current source, because
the voltage/current curve is very exponential, and biased by temperature.

If you want to use parallel strings of LEDs, each with its own resistor,
then you can use any power supply you want, and you could even trim the
resistor values to equalize the brightness. In series, if a given LED has
a different brightness from another at the same current, that's what shows
up.

Have Fun!
Rich

 

[Michael]

The White Luxeon Stars are specified with bins- and both the intensity

and color temperature vary with the bin letter. A is highest with
maximum light and coldest color temperature and E is weak and warm.

When I ordered my Luxeon 1 stars from future electronics, I ordered 5 of
them.

When I placed the order, there is nothing on the order form that says
anything about being able to designate a bin number.

On the parts I have, there are some numbers printed on the back, one looks
like a common part number (all 5 have this) and the other number is....
RX0JW 0122186 and
RX0JW 0122205

When you order, how do the bin numbers come into play and how can you
choose?

Like I say, I have 2 of the stars in series and one is obviously a mellower
white and dimmer YET they both have the 0122205 number.

(always something).

THANKS.

 

[redbelly]

There are some simple constant-current circuits that are easy to build:

He already has the simplest:

-----
+ --+--Vin|LM317|Vout---+
| ----- |
[.1uF] Adj [R]
| | |
Gnd +----------+-----> Constant I output

Only 3 parts. I = 1.2/R

Ed

What about a cap on Vout?

Mark

Not needed.
Ed

I would think a cap on Vout would cause instability when running it as a
CC source and possibly make it oscillate or slow the tracking to the
point large current excursions could take place, possibly frying the
load.

Jim

Ed, Jim,
Thanks for answering. I still have some learning to do when it comes
to linear regulators and when to use capacitors. ON Semiconductor's
datasheet for the LM317 does say:

"Although the LM317L is stable with no output
capacitance, like any feedback circuit, certain values of
external capacitance can cause excessive ringing. An output
capacitance (CO) in the form of a 1.0 uF tantalum or 25 uF
aluminum electrolytic capacitor on the output swamps this
effect and insures stability."

I'll have to do a fair amount of self-study before I understand this
better.

Mark

 

[James Thompson]

There are some simple constant-current circuits that are easy to
build:

He already has the simplest:

-----
+ --+--Vin|LM317|Vout---+
| ----- |
[.1uF] Adj [R]
| | |
Gnd +----------+-----> Constant I output

Only 3 parts. I = 1.2/R

Ed

What about a cap on Vout?

Mark

Not needed.
Ed

I would think a cap on Vout would cause instability when running it as a
CC source and possibly make it oscillate or slow the tracking to the
point large current excursions could take place, possibly frying the
load.

Jim

Ed, Jim,
Thanks for answering. I still have some learning to do when it comes
to linear regulators and when to use capacitors. ON Semiconductor's
datasheet for the LM317 does say:

"Although the LM317L is stable with no output
capacitance, like any feedback circuit, certain values of
external capacitance can cause excessive ringing. An output
capacitance (CO) in the form of a 1.0 uF tantalum or 25 uF
aluminum electrolytic capacitor on the output swamps this
effect and insures stability."

I'll have to do a fair amount of self-study before I understand this
better.

Mark

That only applies to using it as a voltage regulator, not using it in
current source configuration. For using it as voltage regulator, generally
a 10 uf cap is good enough.

 

[Tim Auton]

When I ordered my Luxeon 1 stars from future electronics, I ordered 5 of
them.

When I placed the order, there is nothing on the order form that says
anything about being able to designate a bin number.

On the parts I have, there are some numbers printed on the back, one looks
like a common part number (all 5 have this) and the other number is....
RX0JW 0122186 and
RX0JW 0122205

You might want to read up on Luxeon codes here:

http://www.lumileds.com/pdfs/AB21.pdf

The first part is the bin code. All LEDs from a particular bin are
similar, but there is still some visible variation within bins as you've
discovered - up to 20% for your LEDs, which are from intensity bin 'R'.
If you want better intensity matching than Lumileds do at the factory
you've got to do it yourself - either by testing a whole bunch and
selecting similar ones or by adjusting the current to the LEDs
individually.

When you order, how do the bin numbers come into play and how can you
choose?

You could order several thousand from Lumileds and pick exactly the bin
you want Assuming you don't want that many you'll have to shop around
for a vendor who carries the bin you want in small quantities.


Tim