Digital Current Control for LED array

--
Yeah. Put a resistor in series with each LED such that when you
switch the power supply over to that array each LED only consumes
what it needs to be as bright as the rest of the LEDs in that (or
the entire) array.

If you\'re looking for precision, the value of each of those
resistors  will need to be determined experimentally and you\'ll need
a light meter to determine what value of resistance will cause each
LED pixel to be as bright as you want it to be.

Anything else?

I think this will fit the bill. Many other devices provided by Maxim and competitors will give individual control of amplitude, often through "PWM" or Pulse Width Modulation which will turn the LED on and off faster than you can see the change.

1 amp though? Those aren't the normal LEDs I know of. I thought the 5W Luxeon lumileds were limited to ~700mA.

Yea, I'm using 3W Luxeon III lumileds which are rated to ~1400mA.

Looking at the MAX6958/6959 I can't tell whether I can alter the current directly up to 1A for each LED. It says the current is limited to 26mA for the display LEDs but if it's going straight out to descreet LEDs it looked like only 275mA. I need to be able to alter the current output from 0-1A for each LED.

the problem is I need the output current to the LED to be a set percentage of the supply current automatically. If I put a resistor in front I'll need to change the value of the resistor every time I change the supply current won't I?

A friend has suggested converting the current to a voltage and putting it through an amplifier with a gain of say 0.2 then convert it back into a current. That will allow 20% of the input through. How do I convert a current into purely voltage? Do I need to know the input current to do this or can a circuit be designed that will convert the current to voltage regardless of input size?

Thanks Evan

Circuit to convert a current into a voltage, ---/\\/\\---

Circuit to amplify by 0.2, in ---/\\/\\--o--/\\/\\--- ground - | - out

Sorry, Evan, I couldn't resist.

--
 Thanks,
    - Win

--- No, because you won't be changing the supply current, you'll be changing the supply's output _voltage_.

What the resistors will be doing is causing the current going into the LEDs to be scaled so that the weaker LEDs will get more current than the strong ones for the same light output.

If, as you say, the relationship between light output and current is linear, then once you've got them all at the same brightness then their brightnesses should track as you vary the supply voltage.

One thing I failed to mention was that your initial premise was wrong. That is, if you connect seven LEDs in parallel with the output of a power supply which is putting out seven amperes, the current _will not_ divide equally because the forward voltages of the LEDs will _not_ be equal. What will happen is that one LED will hog most of the current, and then as it heats up its Vf will drop, causing it to hog even more of the current, causing it to heat up more... Eventually it'll die, and if it faiss open the LED with the next highest Vf will give up its life, and so on...

---

--- Forget it. Just use the resistor method I outlined earlier.

BTW, when you reply to a post please leave a little of the post you're replying to in your reply so we can tell to whom and about what you're replying.

Thanks,

-- John Fields Professional Circuit Designer

I forgot to elaborate on an item for the one IC: to drive a 1 amp output from any LED array driver chips, you'd want to go through a mosfet or one (or 2) bipolar transistor stages to get the current you want from one control voltage. Current still needs to be limited "appropriately."

You may want to use individual, tiny switching supplies to give you high, fixed currents to each LED. It's more complex, but the precision control over current and improved effeciency might be worthwhile. Current-mode switchers are designed for LEDs but would typically require an external transistor and beefier inductors to switch the high loads. The LED switchers are also available from maxim-ic.com.

If you try to use one supply for all 7 LEDs in parallel, you need current limiting for each LED such as your thought for a different circuit for each LED. Often people just use resistors for the low current LEDs but that doesn't supply much tuning. Make sure in whatever scheme you use that the maximum current *cannot* be exceeded for any individual LED. It's a shame to burn one out.

It may be the best way to go is a multi-channel DAC to control MOSFETs inline with each of the 7 LEDs. To get the low voltage drop across the FET, you'll need an amplifier to drive each MOSFET with feedback from a per-LED current-sense resistor to keep the gate voltage at just the right level. Keep in mind how much power each will burn when you make your selection.

If you'd like to get into more specific detail please email me at johnhandwork at mail dotc om and I can help you come up with something decent. I should have a few 3W devices myself from a coworker in a few days and I've got a teeny-tiny Luxeon Flash at home waiting to be wired up to something. It's in my own interest to get something "nice" together.

- John_H

As far as I know connecting LEDS in parallel is a bad idea as if any one of them has a slightly low resistance it will light brighter and all the others will be dim . So unless they have the same model number, come from the same shop , have same age you run a risk. And even then these things degrade at different rates so you still run a risk . Do a bit of Math and put them in series and add a resistors in parallel to the each of the series connected LEDs .

"Go easy with the whisky"

theJackal

--- If this is what you're advocating:

+V | +----+ | | [LED] [R] | | +----+ | | [LED] [R] | | +----+ | | [LED] [R] | | +----+ | | [LED] [R] | | +----+ | | [LED] [R] | | +----+ | | [LED] [R] | | +----+ | | [LED] [R] | | +----+ | GND

I'd like for you to explain how to choose the values of the resistors so that as the current through the string is varied the brightness of all of the LEDs will vary identically.

-- John Fields Professional Circuit Designer

--
You seem to have missed the OP\'s post which stated that he had an
array of seven LEDs which he wanted to dim, with all of the LEDs
being equally bright as dimming proceeded.

Your suggestion of a series string with a single series resistance
in it won\'t work, simply because if the LEDs arent equally bright to
begin with, (which they won\'t be) then the differences in brightness
between the lamps will persist as their intensities are adjusted.

--
You\'re still full of shit.

--- Hmm...

As well as being abjectly stupid about LEDs, binary counters, and just about everything else you've had the misfortune of trying to write about, I see you have to quote others since you obviously have no wit of your own.

Jackal? More like jackass.

-- John Fields Professional Circuit Designer

--- Perhaps, but in the world of _written_ communications it's easy to tell.

For example, your use of 'Vn' for 'Vf' (the universally accepted symbol for the forward voltage of a diode) and 'In' for 'If' (the universally accepted symbol for the forward current of a diode) coupled with the ridiculous arithmetic you used to try to illustrate what you didn't understand in the first place labels you as having sat in the dunce chair for a long, long time.

Not to mention the fabulously funny flip-flop fiasco, where you made the old saying that every man's reach should exceed his grasp come to life.

Jackal? More like Jackoff...

-- John Fields Professional Circuit Designer

I actually had in mind both the circuit above and a series connection of LEDs with a single resistor connected at the other end depending on which case was taken in consideration. Well the Op should know 2 values regarding the LEDs from their datasheet

1) Every specific colored/type LED has a specified forward voltage drop (Vn) for nth LED 2) He should know the specified current limit of the LEDs he is using. (In)

Assuming the LEDs are of the same type then simply R = Vdcsupply

-V1*7/I1 He can modulate the brightness using a PWM source being careful that the current * the frequency of his PWM source is less or equal to I1.

I'm not sure whether the OP is using identical LEDs or differently colored ones in which case the above values at 1) and 2) are different.

Lets take the more complicated case which is the circuit you've drawn above in which each LED is different so has a different limiting current through it , needs a differently valued resistor and has a different forward voltage. I would add another Resistor (Rd) in series with the parallel array of LEDS and resistors.

if In is the mesh current flowing through the nth mesh Then I1*R1 = Voltage drop on LED 1 =V1 R1 = V1/I1 R2=V2/I2 . . . R7= V7 /I7

I8 is the mesh current through the mesh containing Rd and the power supply so

-V + (I8-I1)*R1 + (I8-I2)*R2 + ... +(I8-I7)*R7 + Rd*I8= 0

All the resistors can easily be calculated from the above equations He can vary the individual brightness of each LED by varying Rn separetely or otherwise varying Rd or the power supply voltage (Pulsed).

theJackal

Using the convention of currents I've used actually

R1 = V1/I1 R2=V2/I2 . . . R7= V7 /I7

Should be

R1= V1/ I1-I8 R2=V2/I2-I8 . . . Rn= Vn/In-I8

theJackal

LMAO Egotism is the anesthetic that dulls the pain of stupidity.

Frank Leahy

"Go easy on the whisky"

theJackal

Talk to an idiot and you won't be able know who is the idiot

"Go easy on the whisky"

theJackal

Is this to correct for "shading" variances across the arrays?

How exactly do you intend to measure the actual light output of each LED?

Is this a one-off correction, or an ongoing process?