PWM Led drivers which source current?

[davivid]

Are there any PWM LED drivers that source current in a similar way to the PWM outputs on the Arduino?

I need to drive 30 LEDs (10 RGB) on my LED driver boards (which are made, example below), and consist of a high current BJT switching circuit. Therefore I need to drive the base of the transistor with 0 - 5v.

All the IC's I have found are current sink. Any advice would be greatly appreciated!

[johnwasser]

If the current-sinking PWM driver you like has open-collector (or open-drain) outputs you can use pull-up resistors (to 5V) to turn those output into 5V logic outputs.  When the output is off the pull-up resistor will take the signal HIGH.  When the output is on it will pull the signal LOW.

[retrolefty]

Do you have a datasheet or link to that CCR2 device? Most constant current driver devices have active outputs and would not work the way you are attempting to PWM it.

Lefty

[fungus]

The thing labelled "CCR" on that diagram looks like a constant current regulator so you don't need any current regulation, all you need is PWM.

If so... put a PNP transistor where it says "7.5V" in your diagram and use a PWM chip to switch it, eg. TLC5940.

[davivid]

If the current-sinking PWM driver you like has open-collector (or open-drain) outputs you can use pull-up resistors (to 5V) to turn those output into 5V logic outputs.  When the output is off the pull-up resistor will take the signal HIGH.  When the output is on it will pull the signal LOW.

That sounds like a great solution, how do I work out the optimum pull-up value?

Do you have a datasheet or link to that CCR2 device? Most constant current driver devices have active outputs and would not work the way you are attempting to PWM it.

Lefty

The circuit is taken from the figure 11 in the datasheet: http://www.onsemi.com/pub_link/Collateral/NSI50350AS-D.PDF

[CrossRoads]

http://www.onsemi.com/pub_link/Collateral/NSI50350AS-D.PDF
Circuit follows the device's data sheet, except for C11.

[davivid]

http://www.onsemi.com/pub_link/Collateral/NSI50350AS-D.PDF
Circuit follows the device's data sheet, except for C11.

I was advised elsewhere to add C11 to reduce any EMI effects (also mentioned in the data sheet), as the rest of the circuit is a touch sensor.

[fungus]

http://www.onsemi.com/pub_link/Collateral/NSI50350AS-D.PDF
Circuit follows the device's data sheet, except for C11.

I was advised elsewhere to add C11 to reduce any EMI effects (also mentioned in the data sheet), as the rest of the circuit is a touch sensor.

I was confused when you said this was a "built" circuit. I thought you couldn't modify it.

C11 is preventing you from using the BJT for PWM control. If you remove C11 you can connect a PWM signal to the base of Q2 (any current sink chip will do the job - just place a smallish resistor between Vcc and the chip input and connect the base of Q2 to the resistor).

EMI effects? They're almost inevitable in a PWM circuit. You'll need to add lots of decoupling and maybe try different PWM frequencies to minimize them.

[davivid]

http://www.onsemi.com/pub_link/Collateral/NSI50350AS-D.PDF
Circuit follows the device's data sheet, except for C11.

I was advised elsewhere to add C11 to reduce any EMI effects (also mentioned in the data sheet), as the rest of the circuit is a touch sensor.

I was confused when you said this was a "built" circuit. I thought you couldn't modify it.

C11 is preventing you from using the BJT for PWM control. If you remove C11 you can connect a PWM signal to the base of Q2 (any current sink chip will do the job - just place a smallish resistor between Vcc and the chip input and connect the base of Q2 to the resistor).

EMI effects? They're almost inevitable in a PWM circuit. You'll need to add lots of decoupling and maybe try different PWM frequencies to minimize them.

Built as in my PCB's are fabricated and populated... but I could replace C or R with other values.

The circuit works fine with C11? (should it not?) I have it connected to a PWM output on an arduino. Now I will try connecting it up to a current sink chip with a pull up resistor - hopefully this will work too!

[fungus]

The circuit works fine with C11? (should it not?)

That depends on the value of the capacitor.

If it's a small value it might just be rounding the corners of the square wave coming from the PWM controller, not enough to completely prevent it from working but enough to distort the colors and maybe heat up the transistor.

[Grumpy_Mike]

The circuit works fine with C11? (should it not?)

What makes you think it works fine? Have you looked at the waveforms on a scope. What is happening is that the FET is spending longer working in the linear zone and so dissipating more heat that it should be. You might not think it is getting hot but it is getting hotter than it needs to be and that is bad for reliability. If the capacitor is too large then it will not give you much if any variability of control.