LP5036: Need an example of deriving 9 pwm bits by multiplying 8 color bits with 8 intensity bits

Hi Bhargav,

    I am out of office now and I will reply to you soon.

Hi Bhargav,

    1.The 9 bits PWM signal are coming from the combination of two 8 bits and reserve 9 MSB bits from the multiplication result, which means the PWM resolution is 9 bit only and will lost information in this process. To improve the accuracy, you can enable the dithering digital control. The dithering digital control is the internal logic mentioned in the reply of Kenneth Du.

    2.The final output PWM duty cycle has 12 bits of control accuracy, which is achieved by a 9 bits of pure PWM resolution and 3 bits of dithering digital control. For 3-bit dithering, every eighth pulse is made 1 LSB longer to increase the average value by 1 / 8th.

Hi Charles,

Thanks a lot for the quick response!

So what you are saying is that if say for example, color=10000011 and intensity=11000011, then their multiplication would produce the 16 bit number 0110001111001001.  Here the 9 MSBs of the product i.e. 011000111 (blue font) would form the pwm signal.

And if pwm dithering is enabled, then the 3 bits adjacent to these 9 MSB bits, i.e. 100 (saffron font) would be used for choosing the dither pattern.

Could you please let me know if I my understood is correct here?

Also, if the above understanding is correct, then I had one more query below about this scheme.

If say for example for an RGB module, color(for R/G/B)=11111111 and intensity=00000001, then their product would be 011111111(for R/G/B). The 9 bit pwm pattern (for R/G/B) obtained in this example is also 011111111. That is PWM_R=PWM_G=PWM_B=011111111. This means that the pwm duty cycle is ~50% for R/G/B leds.

My query here is that even though the intensity required for the RGB module is the least, that is 00000001 (1 LSB), the pwm duty cycle that would go to the RGB module leds is ~50%. So wouldn't this make the intensity of RGB module far greater than the intended least intensity of 1LSB? Or could you please let me know if I am missing something here?

Thanks again for your quick response.

Best Regards,

Vijay Hugar

Hi Charles,

I am sorry for asking an incorrect question in the second part of my query (regarding multiplication of color(for R/G/B)=11111111 and intensity=00000001).

I had actually missed the fact that the product of these two would be 16 bits (00000000 11111111) and so the 9 MSBs would be 000000001.

I am sorry for any confusion. Kindly ignore the second question. 

Best Regards,

Bhargav K

Hi Bhargav,

     Actually, the 9 MSB bits is obtained a little bit differently from the process you descripted. However, the result is right and you can just understand it in your way because it is just related to the internal calculation logic.

     For dithering, you could say that it is an interpolation way to get the approximation of 12 MSB bits of multiplication result. Regarding to the dithering method, the users don't need to know about it in detail since it is a little bit complicated, which would be done automatically by the chip.