# Ideas for High Power LEDs

#### Xenon02

Joined Feb 24, 2021
500
Hello !

I've been working on a project and needed some help with reading datasheet to LEDs because I don't know one thing and ask about one of my ideas.
So I was working on to create a Lamp with high power LEDs independently controlled to change the dimming.

And so the LEDs are 6W, 1A so I thought of how to resolve it.
So I thought of 3 options :
- Use high power Drivers that can give 1A output but I have to buy for each LED diod.
- Use Low power Drivers that can give me 120mA but somehow boost the current, I tried to do some experiments like here :

So I was able to come up with from 120mA to 268 mA, but the resistors where pretty small and couldn't much more + the power that was in each of these resistors wasn't that small ...

- Last option is to again use Low Power Drivers but 120mA connected to the transistor that will work as a "switch"

But the problem is in the LED because their forward voltage changes from 5.9V to 6.3V (model : SPHWH2L5N603YET5A3 ). I don't know if the forward voltage is 5.9V and the diode will have let's say 6V (I know saturation voltage etc) so it will be 1A like if the forward voltage was 6V so the diode will take 1A ? I mean I saw the datasheet but it doesn't say what happens if the forward voltage is not 6V but 5.9V then if I give him 6V will it still be 1A ?

And wondered if the "current boost" is possible, I tried something and somehow boosted but not alot.
So I wondered if there is a way.

If nothing will work then I will leave it with high power drivers each for 1 LED.

#### Michal Podmanický

Joined May 11, 2019
72
In a Led dimming don’t care about forward voltage but rather controlling the current.
For that Led you need a 7V power supply at least. Also too high voltage isn’t good because it causes waist (>9V).

Make a reference from tl431, supply a pot with this voltage, viper output sets the demanded led current. The feedback about led current is taken from series shunt.

Last edited:

#### Xenon02

Joined Feb 24, 2021
500
@Michal Podmanický

I wanted to use the drivers because they have the control of the dimming + the size because it is IC.
So I thought if I use the transistor as a switch then the voltage forward matters ? I didn't know how to better describe it.
So I thought that if the transistor that works as a switch (that was the first option) will have different forward voltage but still the full voltage (like 6V) will be on him no matter what the forward voltage will be the different current he will "give"

The second option was to boost the output current from 120mA to 1A and change it in some way.

Last was to use bigger driver for 1A output.

But I wanted to think about these first 2 options. I tried to think about it in some way but yea.

#### Danko

Joined Nov 22, 2017
1,842
You can connect up to 14 LEDs in series using 95 V power supply
with one driver MP4689DN-LF-Z
or buy 10 drivers lot for \$5.20 and use separate driver for each LED.
Datasheet

Last edited:

#### Xenon02

Joined Feb 24, 2021
500
You can connect up to 14 LEDs in series using 95 V power supply
That's the thing I wanted the LEDs to be parallel so I could control them individually.

So I understand that there is no way to boost the current (120mA from a driver) ? I was able to somehow make it 260 mA so I thought it is possible.

#### Xenon02

Joined Feb 24, 2021
500
Or a different question :

From a datasheet.

What happens if the forward voltage changes from 6V to 5.8V ? If I give the diode 6V will I still get around 600mA ? And if the forward voltage didn't change and was 6V and I give him 6V will it also be around 600mA ? And if the forward voltage changes from 6V to 6.2V if I give him 6V will it still get around 600mA ?
I see there are min/typ/max when it comes to Forward voltage so I wanted to know if something changes if I all the time connect 6V straight to the LED like this. Does the forward voltage changes something here ? I only have one characteristics that doesn't say what happens if I give 6V to the diode that is 5.8Vf or 6.3Vf (voltage forward).

#### Michal Podmanický

Joined May 11, 2019
72
The schematic I posted dimms the led by turning the pot (in range 0-1.1A).

There is also possible to dimm the Led by removing the reference and pot, and feed the noninverting (+) op-amp input with 0-250mV (or 0-5V with resistor divider).

The power efficiency with 8V power supply is about 80% (very close to led switching driver)

#### Jerry-Hat-Trick

Joined Aug 31, 2022
574
How many LEDs are you looking to control? I’m thinking you could use a small controller board like a Seeeduino XIAO to control up to five, with 5 pins to read the wiper voltage of potentiometers and five pins to control the LEDs with PWM outputs.

The way to control the brightness of LEDs efficiently with minimal power loss is to turn them full on and off rapidly, faster than the eye can detect. The greater the mark/space ratio of the Pulse Width Modulation signal, the brighter the LED.

You can drive your 1A current the way you sketched, with the processor pin via a resistor into the base of a transistor but since the XIAO can only drive a small output current you will need to use a Darlington Pair transistor with a gain of around 1,000. The value of the resistor in series with the LED needs to be small and to be related to the voltage of your power supply. 100 ohms at 1 amp would drop a voltage of 100 volts and dissipate a power of 100 watt! Let’s say you have a 12V power supply and you want to have the maximum voltage across the LED of 6.0V, dropping the other 6.0V across the resistor at 1.0A suggests a resistor value of 6.0 ohm but even that would dissipate 6 watts. As already suggested, a suppy voltage of 7.0V just above your maximum LED voltage of 6.3V would be ideal. In that case, dropping 1.0V across the resistor at 1.0A suggests a resistor value of 1.0 ohm which would dissipate a more manageable 1.0 watt.
In practice, you can drive the LED from a higher voltage by reducing the mark/space ratio of your PWM signal - for example from a 12V supply as long as the mark/space ratio doesn’t exceed 50/50. LEDs connected to a high voltage will be destroyed by passing too high a current experiencing thermal runaway but pulsing them so that the average current is low enough will prevent this, although their lifetime will be slightly reduced.

Just to clarify, the LED has a characteristic voltage/current curve such that when you increase the voltage from 5.9V to 6.3V you will increase the current through it. It will probably be dimly lit at 5.0V or lower. They are like ordinary diodes which may have a nominal voltage of 0.6V but with a more rounded ‘knee’

#### Audioguru again

Joined Oct 21, 2019
6,758
LED forward voltage is different for each Led unless you buy matched ones or match them yourself. The forward voltage also changes with temperature or current change.
Modern LED dimmer circuits use Pulse-Width-Modulation.

#### Xenon02

Joined Feb 24, 2021
500
I have for now 2 Ideas :

This one uses transistors as "Switches"

And this one :

A problem came up.
Dedicated 1 driver per 1 LED with high current like 1A in the output requires a special PWM input, so my microcontroller has to give for each Driver dedicated PWM, like 30 dedicated PWM ... know here is not alot but I couldn't put them all in one place.
So yea basically problematic ...

Tried to find maybe an PWM generator with multichannel but the problem is every PWM with multichannel I found was for LEDs and was just a current source.

#### Jerry-Hat-Trick

Joined Aug 31, 2022
574
Wow, 30 x 6 watt LEDs! That’s roughly equivalent to 1.8kW of conventional lighting, more than most theatre spotlights. If they are in a cluster you certainly won’t be able to look at it to tell if individual LEDs have different brightness, so maybe the ability to turn individual LEDs fully on or off would work?

And will you really have 30 individual potentiometers to adjust? As already suggested, would having maybe six parallel banks of five in series be acceptable? I’m wondering what your application is.

#### Danko

Joined Nov 22, 2017
1,842
A problem came up.
Dedicated 1 driver per 1 LED with high current like 1A in the output requires a special PWM input, so my microcontroller has to give for each Driver dedicated PWM, like 30 dedicated PWM ... know here is not alot but I couldn't put them all in one place.
So yea basically problematic ...

Tried to find maybe an PWM generator with multichannel but the problem is every PWM with multichannel I found was for LEDs and was just a current source.
You can easily solve this problem. See how it can be realized:

#### Jerry-Hat-Trick

Joined Aug 31, 2022
574
Back to the idea of driving transistors for each LED (where you could consider using logic level Mosfets instead of darlington pairs) I’m wondering if you could bit bang four 8 bit serial to parallel converters to achieve PWM with sufficient resolution and fast enough to avoid flickering.

For example, each LED is allocated a brightness level of 0 to 99 and the ICs are updated in a loop updating 100 times with each output pin value changing from 1 to 0 when the loop count equals the corresponding LED brightness level.

#### BobTPH

Joined Jun 5, 2013
9,132
To get 100 levels at 100 updates per second for 30 LEDs is a data rate of 300K bps

For comparison, WS2812B strips can be run at 800K bps. So it is no sweat to control them with serial to parallel shift registers.