Hello TI Team,
For a study I get some sample of the LM3429.
I made with Webench LED designer the components calculation for 3x Nichia NS6W083 with 0,3A current from 6V to 9V Input voltage.
and soldered the chip on a test board.
The circuit works well without PWM Signal (I measured 303mA) but I have some problems with this signal:
Some information about the components: Rivp2 = 10k, Rivp1 = 3.3k, Rdim = 21k, nDIM Mosfet 30V/0.5A, L1 = 47µH, Cout = 10µF....
- The LED current is not proportional at all to the PWM signal (run at 147Hz)
- By 5% duty cycle => I measured 44,7mA (15%)
- By 20% duty cycle => 79,1mA (26%)
- By 50% duty cycle => 168,1mA (55%)
- By 95% duty cycle => 303,3mA (99%)
Are some Problems known? Should I place the nDIM transistor before the Rdim resistor like by LM3421 design??
- Wenn nDIM = 0V, The LEDs don't turn off properly (residual current of approximately 200μA in the LEDs chain), Glim effect !! Especially if the operating voltage is going bigger than 8,5V. Is there a solution for this phenomena??
- I choosed a 30V MOSFET 0.5 A for the PWM control switch. But it's always die, I must put at this place a 60V or 100V type, why with so small Voltage? (is something missing, like a protection Z-diode on Gate?)
- In the datasheet, it's explain that a pull-down at the gate of the nDIM MOSFET should be used, but It means that when you turn on the Powersupply, all the LEDs are fully lighting (PWM signal from microcontroller for about 300ms high Z ).... not really logical, right?? Should I used a pull-up resistor?
Could the problems regarding the PWM accuracy and the residual current be solved with the LM3421 ?
Many thanks in advance for your support,
Alain
