• State Resolved
  • Locked Locked
  • Replies 8 replies
  • Subscribers 63 subscribers
  • Views 413 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TPS92512HV: weird behavior of LED current output

Humphrey Appleby
Prodigy 10 points
Part Number: TPS92512HV
Other Parts Discussed in Thread: TPS92512, TPS922055

Hi there,

We design a led drive board with TPS92512HV with design parameters as below:

min max Typical
Vin 22V 26V 24V
Vout 3.7V
Iled 2A

The design is based on the guidelines of datasheet(section 9.4, section 11). However we observed weird behavior of led current output as shown in the videos attached here.

this is what appears on osciloscope of led-

this is led+ on osciloscope

this led is used to illuminate microscope, here is what we capture with live camera

Note that we have three output capacitors on this board.

1) 47uF Electrolytic capacitor

2) 4.7uF ceramic capacitor

3) 0.1uF ceramic capacitor

The minimum required Cout(based on datasheet guidelines) is 12uF.

Deleted:

"""One thing noteworthy is that we tried to remove 4.7uf and 0.1uF ceramic capacitors, then the led drive board is not output any current at all(led+ is 24V and led- is almost 0V).

Does this mean that there is a problem of using Electrolytic capacitor as Cout for tps92512?"""

Sorry, the deleted phenomenon observed purely because of not connecting led drive board to the LED after removing ceramic capsJoy

  • 0
    TI__Mastermind 23150 points

    Hi Humphrey,

    I am currently out of office. I will reply to you within 24 hours. Thanks for your understanding.

    Best Regards,

    Steven

  • 0
    TI__Mastermind 23150 points

    Hi Humphrey,

    Based on the video you posted, I do not clearly understand the problem (the weird behavior) you meet. Could you give me more detailed information? 

    Best Regards,

    Steven

  • 0 in reply to Steven Li
    Prodigy 10 points

    Hi Steven,

    Pardon me for not clarifying the problem. Say, we generate PWM at 250Hz with 50% duty. Based on the design, the current flowing through led should be around 1A and we should've observed a constant illumination on the live microscope images, but instead we saw the RGB spectrum curves sliding b/w 0 to 255 at a frequency like 1Hz(as shown in the 3rd video).

    what could be the possible reason for all this?

  • 0 in reply to Humphrey Appleby
    TI__Mastermind 23150 points

    Hi Humphrey,

    Humphrey Appleby said:
    but instead we saw the RGB spectrum curves sliding b/w 0 to 255 at a frequency like 1Hz(as shown in the 3rd video).

    Do you mean the LED is flickering at 1Hz? If so, do you have any oscilloscope capture that illustrates the 1Hz flickering?

    Best Regards,

    Steven

  • 0 in reply to Steven Li
    Prodigy 10 points

    Hi Steven,

    I think we've found there the problem is. The 1st video(oscilloscope monitoring LED-, and thus the led current) actually does not show 1Hz flickering. Yet we observed it from high speed camera connected to our microscope at an exposure time of around 20 microseconds. Compared with pwm at 250Hz, 20us time window is certainly not enough to include enough current cycles to make illuminations constant from camera point of view. We tried larger exposure time and things improved.

    Now if we plan to stick with 20us exposture time as required by our application, we need to drop the PWM option and adopt analog regulation of LED current through pin IADJ, to make a constant led current for such a narrow exposure window. 

  • 0 in reply to Humphrey Appleby
    TI__Mastermind 23150 points

    Hi Humphrey,

    Noted. Based on my understanding, using high-frequency PWM dimming (like 20-kHz supported by TPS922055) may also help. However, I would say that analog dimming is more suitable here as even with the 20-kHZ PWM dimming, such narrow 20-us window won't include many cycles.

    Do you have any other questions that I can help with? If not, I am going to close this thread.

    Best Regards,

    Steven

  • 0 in reply to Steven Li
    Prodigy 10 points

    Hi Steven,

    Yepp, we decided to go for analog dimming option by providing regulating voltage to IADJ pin with RC filter like this:

    3.3V PWM -----> R2 -------> IADJ

                                             |

                                         R1 | | C1

                                             |

                                         GND

    where: R2=8.2MOhm, R1=10MOhm, C1=0.01uF. PWM's frequency could go as high as 1MHz(we use FPGA)

    calculate pwm2analog v_out: 1.808219 V
    calculate pwm2analog f_cutoff: 3.520700 hz
    attenuation in db at freq 1000 Hz: -49.067473 dB
    attenuation in db at freq 10000 Hz: -69.067420 dB
    attenuation in db at freq 50000 Hz: -83.046819 dB
    attenuation in db at freq 100000 Hz: -89.067419 dB

    Since led is used for illuminating microscope at a constant level mostly, we don't care much about fast adjusting led current.

    Hope this scheme would work with TPS92512.

  • 0 in reply to Humphrey Appleby
    TI__Mastermind 23150 points

    Hi Humphrey,

    Humphrey Appleby said:

    Yepp, we decided to go for analog dimming option by providing regulating voltage to IADJ pin with RC filter like this:

    3.3V PWM -----> R2 -------> IADJ

                                             |

                                         R1 | | C1

                                             |

                                         GND

    where: R2=8.2MOhm, R1=10MOhm, C1=0.01uF. PWM's frequency could go as high as 1MHz(we use FPGA)

    This is a commonly seen approach to drive IADJ pin with a PWM signal. Just make sure the voltage divider will generate the correct IADJ voltage within the 180mV to 1.8V range. Also make sure the RC filter has the appropriate cutoff frequency to smooth out the PWM signal into your target analog signal.

    Best Regards,

    Steven