Power management

Power management forum

State Resolved
Locked Locked
Replies 17 replies
Answers 3 answers
Subscribers 64 subscribers
Views 395 views
Users 0 members are here

Support feedback

Options

Options

Related

LP8863-Q1: Measurement result of LED PWM ON TIME

Koji Hayashi92

Mastermind 7030 points

Part Number: LP8863-Q1

Hi TI-team

When my customer measured the LED PWM ON TIME, the result looked like the black line in the figure below.

LED:152Hz

PWM input:0.016%

LED Duty:0.016/65536/100=10 => 10/8192*100=0.122%

LED PWM ON TIME: 0.122/100/152=8.03us

The time is short when measured at 1/2 of the rise and fall.

Does this mean that the time is originally delayed for the rising edge of the red line?

Does this have any effect on the ”SLOPE_SELECT[2:0]” setting in ”8.3.6.3 Sloper” in the datasheet?

Best Regards,

Koji Hayashi

over 4 years ago

0 Frank Qiu over 4 years ago

TI__Genius 15776 points

Our expert is on business trip, please wait for the reply.

0 Koji Hayashi92 over 4 years ago in reply to Frank Qiu

Mastermind 7030 points

Hi Frank-san

What is the situation?

0 Charles-Chen over 4 years ago

TI__Genius 11836 points

Hi Koji-san,

The rising edge and falling edge both have delay and the PWM on time is supposed to be accurate if they also consider the delay of falling edge. The delay of rising and falling edge are caused by the parasitic capacitors of LED cathode. Slope select will take effect when the brightness is changed. For the same brightness, the slope doesn't have any effect.

0 Koji Hayashi92 over 4 years ago in reply to Charles-Chen

Mastermind 7030 points

Hi Charles-san

Thank you for response.

What percentage of error do you expect as the design value of the IC?

(e.g., when 0.122% is detected, there is a variation of ±10%, etc.)

Best Regards,

Koji Hayashi

0 Charles-Chen over 4 years ago in reply to Koji Hayashi92

TI__Genius 11836 points

Hi Koji-san,

Please clarify the variation of parameter. The pwm input variation depends on the PWM input signal accuracy.

0 Koji Hayashi92 over 4 years ago in reply to Charles-Chen

Mastermind 7030 points

Hi Charles-san

The variation will be the respective variation of the rise and fall.

0 Frank Qiu over 4 years ago in reply to Koji Hayashi92

TI__Genius 15776 points

hi Koji-san, Charles will be out of office today, and he will replay you later.

0 Charles-Chen over 4 years ago in reply to Koji Hayashi92

TI__Genius 11836 points

Hi Koji-san,

Your question is confusing. The variation of rise and fall time is related to the input PWM signal and the external capacitors of output pins.

0 Koji Hayashi92 over 4 years ago in reply to Charles-Chen

Mastermind 7030 points

Hi Charles-san

Does that mean there is no delay inside the IC ?

0 Charles-Chen over 4 years ago in reply to Koji Hayashi92

TI__Genius 11836 points

Hi Koji-san,

The detection frequency is 20MHz so that I think the delay can be ignored.

0 Charles-Chen over 4 years ago in reply to Charles-Chen

TI__Genius 11836 points

Hi Koji-san,

Do you have any more questions?

0 Koji Hayashi92 over 4 years ago in reply to Charles-Chen

Mastermind 7030 points

Hi Charles-san

Sorry for the late reply.

I am aware that the variation of 20MHz is 7%, but my understanding is that if I detect 0.122%, the variation will not be 0.113% to 0.130%.

Is that correct?

0 Frank Qiu over 4 years ago in reply to Koji Hayashi92

TI__Genius 15776 points

Hi Charles, please keep following this case

0 Charles-Chen over 4 years ago in reply to Koji Hayashi92

TI__Genius 11836 points

Hi Koji-san,

Your understanding is correct. Please close this thread if you don't have any more questions.

0 Koji Hayashi92 over 4 years ago in reply to Charles-Chen

Mastermind 7030 points

Thank you for response.

I think that LED PWM ON TIME: 0.122/100/152=8.03us

20MHz is 50ns and the variation is ±3.5ns at 7%.

Does taht mean that 8.03us is ±3.5ns variation ?

or is it affected by the 152Hz variation ?

+1 Charles-Chen over 4 years ago in reply to Koji Hayashi92

TI__Genius 11836 points

Hi Koji-san,

Even if the on time has the variation caused by the variation of 20MHz, the duty cycle is accurate.

0 Koji Hayashi92 over 4 years ago in reply to Charles-Chen

Mastermind 7030 points

Hi Charies-san

I'm sorry for the late reply.

I answered my customer.