Constant Current Led Drive

[v1.5]

Hi.

Im here for searching some kind of solution for driving paralled leds using constant current led driver chips like tl4242.

Here what im looking for. Im planning to control leds brigness using microcontroller based system. As you know , for variable brigness of leds , best way is to use a current source for driving leds. For that purposes lots of ic available on market, but i was searching some kind of different solution.
As known , constant current based led driver chips can source up to 500mA (or even more ) current for serial connected leds. This is great way to adjust brighness of the leds using microcontroller. But there is an issue on such as these systems. If when one of the serial connected led is open circuited , the current path will be open circuited to and none of the leds will shine anymore. This will make some problems on real world.
For that issue im here to ask you that what can i do to eliminate that problem ? Is there any method can be done such as these circuits out of there ?

Thanks.

 

[wkrug]

Put LED's parallel without a Resistor at any LED was never a good Idea - It's a veritable LED Killer.
Put them together serial.

Some LED Driver Chips has an PWM Input.
This Input You can connect to a PWM Output of the Microcontroller.

The pulsewidth steer the brightness of the LED's.

The Advantige of the Solution is, there is no colour change at the LED's, The brightness is made by the On/Off Time of the LED's.
The Problem is the LED will flicker with the PWM Frequency, so that must be carefully choosen.

Other LED driver Chips has an Brigtness Input that works with constant Voltages.
When the Controller has an integrated DAC You can use this.
Otherwise You can use the PWM again and put it through a Low Pass filter to get an constant adjusting Voltage.
But this can make colour mistakes at lower Current of the LED's.

The "Killer" Application is the TCA9685.
It has 16 Outputs and can be connected to a Microcontroller via I²C.
So You can adjust the Brightness for any Output seperatly.
But the Protocoll is not easy to undestand.

 

[Diver300]

You could put some a zener diode in parallel with each LED.

If you are trying to increase reliability, be aware that if a zener diode fails short circuit, that will increase the heating in the driver chip.

Also, LEDs have good reliability if not driven too hard. They go dim after being on for a long time. Zeners won't help that.

If one LED fails, unless you can increase the power to the other LEDs, the overall brightness will be reduced. If you can accept that reduced overall brightness of one less LED, you might improve reliability more by simply increasing the value of the sense resistor slightly.

 

[rjenkinsgb]

What's the application - it could be anything from indicator lights to room or vehicle lighting and different problems may have very different solutions.

 

[v1.5]

What's the application - it could be anything from indicator lights to room or vehicle lighting and different problems may have very different solutions.

Yes. I thought the solution will be common so no need to tell about it but it seems that solution directly dependent of problem and not that easy like i thought .
The project is measuring width of wooden plate for my backyard works. The measurement accuarcy is not that too much important at least 5cm is good for me. Im planning to build a barrier of my wood cutter table with using ir transmitter and receiver on each side of the barrier. This will contactless measure approximate width of the material. This project will do my job fast enough.
Increasing current will increase the intensity of the led , so the receiver will handle with variable distances between receiver and transmitter barriers by incresing or decreasing of the intensity. The goal is as much as said. The ir led must be 100 pieces and so the source/sink current will at least 20mA and to max 80mA.
This is what i have thought for solution on my backyard wood cutter table. So how you think of i should handle of driving leds ?

 

[crutschow]

You could put some a zener diode in parallel with each LED.

I don't see how that can help (?)l

 

[Diver300]

I don't see how that can help (?)l

A zener diode in parallel with an LED will take the current if the LED goes open circuit. If the zener is chosen to have a voltage just a bit larger than the voltage of the LED, then the overall voltage of the string won't go up by much, so the current will be maintained and the rest of the LEDs will stay alight.

 

[v1.5]

A zener diode in parallel with an LED will take the current if the LED goes open circuit. If the zener is chosen to have a voltage just a bit larger than the voltage of the LED, then the overall voltage of the string won't go up by much, so the current will be maintained and the rest of the LEDs will stay alight.

Perfect. Thank you for such a good information

 

[crutschow]

If the zener is chosen to have a voltage just a bit larger than the voltage of the LED,

That's not possible with typical tolerances for Zener and LED voltages.

 

[Diver300]

If the zener is chosen to have a voltage just a bit larger than the voltage of the LED, then the overall voltage of the string won't go up by much, so the current will be maintained and the rest of the LEDs will stay alight.

That's not possible with typical tolerances for Zener and LED voltages.

It depends how you understand "a bit".

Whether zeners would work depends on how much spare voltage there is, and how the tolerance of the voltages stack up.

If there's a 24 V supply, and 4 white LEDs in series, there's still about 12 V left for the resistor voltage and the minimum voltage across the control element, where less than 1 V is needed.

A 6.8 V zener in parallel with each LED would mean that the voltage across the string would go up by around 3.6 V if an LED went open circuit, and the rest of the LEDs would continue to work.

Is it correct to describe a 3.6 V change as "a bit"? I don't know for sure, but having zeners in parallel with the LEDs can work. Like any circuit design, it needs tolerances to be considered, and it's entirely possible to engineer any circuit well or badly.

 

[gophert]

I would use a 12V supply and forget about the constant current driver BS
Then wire up groups of 6 IR LEDs in series You'll have a forward voltage of 8.1 to 9.6 Rarely will you have 6 in the whole series very far off from the typical value (1.35Vf), so, according to OSRAM, you can generally design to the typical VF when you use 6 or more LEDs in series for typical T-1 3/4 LEDs or 3 or more for highly specified (modern) LEDs from a reputable supplier.

so, assuming 7 in a series, 9.45V drop from Vf, 12VDC supply, one resistor for the 2.55V drop, you'll need a 51 ohm resistor for 50mA.

You'll need 14 sets for 98 lights. Not your requested 100 but its your design to modify.

At 50mA, you'll need 2.5 amps so buy a 4 amp x 12VDC power supply. You're constant current will be constant as long as you have 12VDC from your supply - which is always

 

[v1.5]

I would use a 12V supply and forget about the constant current driver BS
Then wire up groups of 6 IR LEDs in series You'll have a forward voltage of 8.1 to 9.6 Rarely will you have 6 in the whole series very far off from the typical value (1.35Vf), so, according to OSRAM, you can generally design to the typical VF when you use 6 or more LEDs in series for typical T-1 3/4 LEDs or 3 or more for highly specified (modern) LEDs from a reputable supplier.

so, assuming 7 in a series, 9.45V drop from Vf, 12VDC supply, one resistor for the 2.55V drop, you'll need a 51 ohm resistor for 50mA.

You'll need 14 sets for 98 lights. Not your requested 100 but its your design to modify.

At 50mA, you'll need 2.5 amps so buy a 4 amp x 12VDC power supply. You're constant current will be constant as long as you have 12VDC from your supply - which is always

Thanks.
This will work but it doesnt solve any of the intensity problems. As i described above, each side of the barriers distances will vary dependent on the measured material diamention. So voltage dependent led drives doesnt support this option automatic. Everytime increasing or decreasing distance , i must change the value of the resistor. This is why i thought CC source at first.
So what do you think about this problem ? I havent tried yet anything since im waiting my orders delivering but what if i choose the current at near to the max raiting like 80mA (100ma absolute peak current obtained on the datasheet ) ? How does it behavour when barriers is too close or too far ? Also sunlight is another problem to solve

 

[wkrug]

How does it behavour when barriers is too close or too far ? Also sunlight is another problem to solve

So the flicker can be a part of the solution.

When give a rectangular 50% pulses to the LED and at the receiver side put behind the receiver a small bandpass filter with the same frequency, constant light should make no more problems.
Do not Choose a frequency that's a multiply of 50 or 60Hz ( Lamp Flicker )

After this cicuit You can took an active rectifier and a Comparator.

I've tryed this a few Years ago.
It works, but at direct sunlight at the receiver diode it will be blind for the Frequency signal.
That must be solved mechanical!

When the Receiver Diode has an optical filter that's according to the sending Diodes, the Sun effect would be better.

Sun will give about 1000W Ray per qm that's many power.

 

[Diver300]

IR light beams, as well as IR remote controls, use light that is switched on and off very rapidly and the detectors look for the rapidly changing IR light level, not the background level.

The background level is affected by sunlight. As far as I can tell, IR LEDs produce light that is only slightly longer wavelength than visible light, and objects have to be quiet hot (>200 °C) to produce IR light that is at the same wavelength as that. However, I don't know what you are using to detect the IR light, and the sensors may be sensitive to the IR with wavelengths over 1 - 2 µm that will come from any object at room temperature.

Can you explain how you are intending to do the measurement? Can you draw a picture of the arrangement?

 

[v1.5]

Wow. Things getting complicated

So the flicker can be a part of the solution.

I have just searched some infos on the net and as Diver300 said it s the best way to handle with sun. The problem getting increased when i use hf signals for driving ir beams. Especially for comparator circuitry. Since im plannin lots of ir pairs , comparators will needed. That will increase the cost and design time. Driving the leds is much easier than reading the receiver diode. Now i should worry about receiver circuit too .

Can you draw a picture of the arrangement?

Here what i have found on aliexpress that looks typical same as i thought but with 5cm accuarcy. And also on the picture it has much higher distances , all i need max 20cm lenght between them. This will be placed on the wood cutter table for my backyard projects

 

[v1.5]

It works, but at direct sunlight at the receiver diode it will be blind for the Frequency signal.
That must be solved mechanical!

So for every option i have to put sensors in some kind of material for not to touch directly with sunlights right ?

 

[wkrug]

So for every option i have to put sensors in some kind of material for not to touch directly with sunlights right ?

I would say so.
It depend of the used Sensors, witch Filters are included.
Some are better some less.
When the Sun Brightnes is high enough and shine directly into the sensor, any will go into saturation and You can't filer out any Signal.

From Sharp it gives Optical Range sensors. With different range values.
They produce a IR Light Point onto the Material and scan this with an LDR Bar inside.
The output Voltage according to the disdance but its not linear ( Look into Datasheet ).
Like they would work into Sunlight I could not say - Never played around with them.

 

[KeepItSimpleStupid]

let's assumeyour measuring the intensity at the other end. Can;t you just do that?

e.g. A small I-V converter fed to a Sample and Hold to get the value of the nominal intensity. Store this value somewhere on calibrate.
Feed this to a differential amplifer and comparitor or sets of a comparitors.

our PC would output the voltage to the sample and hold fast enough so it would not droop.

Depending on the number of LEDs, you could generate an interrupt if any went out of range and then if you had too figure out whichones

==
The other problem you may have is ambient light. You may need a lot of TSSOP devices from Vishay. They do autogain and you would probably have to modulate the IR source just like remote controls do.

I know I don;t fully understand the problem.

 

[ChrisP58]

Here is an LED bypass part that is intended to solve the problem of open LEDs in a series string.

**broken link removed**

 

[wkrug]

Have You thought about IR Receiver Chips like for a TV Set?
The sending Diode has to be modulated with 38kHz and the Part spit's out a digital Low / High signal.

Vishay writes something about 45m Range.