CN5711 trouble getting the correct current limiting setting

[Wnef]

Hello all, I'm usualy the kind of "read forum but don't post" type of guy but I'm so clueless on this one !!!

So here is my situation.
I'm designing a circuit to control a 3W RGBW led via I2C
I'm using four cn5711 chip ( datasheet link ) to control each channel.
PWM signal is coming from a PCA8632 to each one of them.
Power is from a two paralleled 18650 battery ( 4000mah total  protected on low side  )

And here is the problem.
Datasheet of the CN5711 specifies that the LED pin current is set by a resistor R-iset wich value is calculated by this formula
ILED = 1800V／R-iset
they give this example :
R-iset = 1800V/1A = 1.8kΩ

I want 750ma so 1800/0,75 = 2.4 kΩ

But mesuring with a multimeter I only get around 380ma
If I put a 1.8kΩ resistor insted like in the example I only get around 800ma

I'm pretty sure my problem is here (at the Riset) but just as sanity check I've verified the followings :
- I've use two different meters for mesure
- Use psu rather than battery
- I've feed 3.4V at 750ma with bench psu to the LED to check if they would take it ( the did and happyly blinded me ! )
- I've cheked the compatibility of the pca9632
     - outputs a 1.5khz pwm ( mesured with scope ) => in spec with the 2khz max of the CN711
     - are supposed to deliver 25ma wich is above the 1 uA required for CE pin ( chip enable PWM input of CN5711 )
I even taken the PCA9632 out of the equation and shorted CE to VCC on the chip and still I cant get it to output current it is suppose to give me with the ILED = 1800V／R-iset formula.
- Parts are supplied by lcsc some of them soldered by jlcpb or myself for the same result.

Is this normal ? Doe someone has the same experience with cn5711
I've seen CN5711 mounted on pcb with a potentiometer in place of R-iset
so I guess lot of people don't even bother to read the datasheet and just turn the pot until power requirements are met
should I just change the resistor value until i get the desired output and forget the ILED = 1800V／R-iset formula ?
I'm fairly new to electronics so I'm kind of not confortable to step aside from the datasheet


Just for context.
I've chose the cn5711 because it is cheap an combines the role of a current limiting resistor and mosfet to drive the led.
My part count and size of thoses parts is limited because the pcb need to fit in circle of 30mm diametre in order to fit in alumimun housing I have. the spx3819 is just a regulator for the PCA9632, and I'm considering add pullup resistors to the I2C SDA and SCL but I'm have not read this yet

I've linked schematic and PCB  layout (4 layers- Ground on inner 1&2 layers ) even if it does not really matter in this case since I've tried outside of the pcb.


So yeah ... do you have the same issue ?
Am I missing something ?
Should I just forget the formula ?


Hope to get so feedback !

[Wnef]

Ok I think I understand where is the issue
It seems the voltage drop of the CN5711 is way more than I expected
far from what I mesures with Riset set at 2.7kΩ

Volt   R     G       B       W
4,6   738   747   738   740
4,5   738   747   738   740
4,4   738   745   738   736
4,3   738   742   731   732
4,2   738   738   707   725
4,1   738   729   666   725
4,0   738   687   620   680
3,9   738   642   574   627
3,8   738   600   527   573
3,7   738   561   486   525
3,6   738   518   440   472
3,5   737   475   394   420
3,4   733   432   348   365
3,3   718   395   309   318

Since red led has lower foward voltage than the other colors it seems to coroborate a voltage drop around 1.2v trougth the CN5711

So maybe it is a normal behavour, and I just need to live with it !

If anyone has a suggestion for a replacement of the CN5711 it would be greatly apreciated !
 

[LinuxHata]

ONSEMI has the similarly featured chips (in fact, CN5711 is clone of one of them) and I have used them for high power RGBW led control, but I was using 12VDC supply, so really do not care that much about voltage drop.