• State
  • Locked Locked
  • Replies 15 replies
  • Subscribers 64 subscribers
  • Views 2990 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

TPS40211 LED driver and camera flicker

Craig Easdale19
Expert 1425 points
Other Parts Discussed in Thread: LM3429, TPS92690, LM3421, LM3424

I've recently been working with the TSP40211 to produce an LED panel driver. To the eye the output is perfect, however this is to be used as a film light and there is noticable flickering even though the LED PWM output is set at 500KHz.

Is there anything I can do to resolve this problem and get rid of the flicker?

So far I have tried adding the 10uF electrolytic capacitor to try to smooth the output, and I have also tried a 100uF but with no success.

Here's the current schematic.

  • 0
    TI__Mastermind 25500 points

    Are you saying the switching frequency is 500kHz or that you are trying to PWM dim at 500kHz? If you are trying to PWM dim (the signal labelled "PWM") at that high a frequency you are probably oscillating as it will not be able to do that. Is this happening with a PWM signal or just while fully on? What are the switching and PWM dimming frequencies you are using?

  • 0 in reply to Clinton Jensen
    Expert 1425 points
    I am using around 130Hz for the PWM signal into the IC. The PWM output of the IC is set at a fixed 500 KHz by the components. Seems to be happening at all times, but I need to double check that.
  • 0 in reply to Craig Easdale19
    TI__Mastermind 25500 points
    I have seen this happen if the camera refresh rate is close to the PWM signal or a multiple of. Do you know what the camera frame rate its? Did you try adjusting the PWM frequency to see if there is a difference?
  • 0 in reply to Clinton Jensen
    Expert 1425 points
    The camera frame rate is likely to be anywhere from 25 - 250 FPS. I was seeing flicker using my phone camera which is at 30 FPS, as well as with another camera at 240 FPS. What's the maximum I could PWM at?
  • 0 in reply to Craig Easdale19
    TI__Mastermind 25500 points
    The maximum frequency really depends on the dimming ratio required. The LED current slew rate is finite so the faster you go the lower the overall dimming ratio will be. You can reduce the output capacitance to speed up the slew if needed. But you may be fine up to 1kHz or more depending on your requirements.
  • 0 in reply to Clinton Jensen
    Expert 1425 points
    I've found reducing the output capacitance reduced the smoothing on the LED output itself so more of the PWM ripple comes through. Adding a chunky 100uF capacitor smooths it a little better, but that's not ideal I suppose.
    Is there any way to limit the ripple of the output / reduce the PWM frequency noise in it?
  • 0 in reply to Craig Easdale19
    TI__Mastermind 25500 points

    Unfortunately that is the nature of PWM dimming and adding a large output capacitor can smooth the flicker but it also changes the dimming profile so you don't know what you will really have through the LEDs based on the PWM signal duty cycle. People PWM dim so that they have this control of brightness with duty cycle. The other reason people PWM dim is that the "on" current of the LED is known so the color spectrum is known (it can shift with current) and constant. Adding a large output capacitor would defeat that purpose.

    That being said, do you have a need for PWM dimming? Or could you analog dim the current? Analog dimming does not have low frequency content in the LED light so you would not have flicker. But you always will with PWM dimming, so the only way to get around flicker is to get the PWM frequency well out of the way of anything you would like to use it with. If a boost converter is required as well as PWM dimming we do have other parts that are designed to PWM dim much faster.

  • 0 in reply to Clinton Jensen
    Expert 1425 points

    Thank you for the feedback, Clinton.

    I had initially ruled out analogue dimming of the current because I thought the risk of LED colour shift was going to be more of an issue. It seems however that would be far more preferable than any flicker or strobing on camera.

    Could you suggest any parts that might fit my requirements? The panel voltage is 32V, max current 0.7A, and the input ranges from 6.5-24 VDC.

  • 0 in reply to Craig Easdale19
    TI__Mastermind 25500 points

    The LM3429 can do this and has an evaluation board close to what you need. But if you need very accurate analog dimming the TPS92690 has the best analog dimming profile. It also has an EVM that is already close to the specs you need. Both might be slightly overkill as they are designed to do 1A or more, but they are still close. You probably wouldn't need to do much more than set the current levels and maybe the input UVLO settings.

  • 0 in reply to Clinton Jensen
    Expert 1425 points

    So how do these parts differ in a way that they'll avoid the flicker I'm currently experiencing? The seem to still make use of PWM dimming, is that not where the problem arises? If I was to use the PWM dimming, what is the maximum speed I could PWM these chips at?

    I'm guessing if I were to make use of the analog dimming, a digital potentiometer or DAC would be required in order to control it digitally...?

    As I'm already exploring a new IC anyway, is there anything I can do that might avoid the LED going on at 22V? I'd like to power it with up to 24V but just due to the LEDs it will illumate on it's own at this voltage even when not driven.

  • 0 in reply to Craig Easdale19
    TI__Mastermind 25500 points
    They are capable of PWM dimming, but they also have analog dimming capability. That just adjusts the current through the LEDs for brightness control rather than turning them on and off at a certain duty cycle so there will be no flicker. If you PWM with those parts you can likely go up to around 1kHz PWM frequency.
    As for the leakage problem, that is common for a boost converter since you always have an inductor and diode path from Vin to Vout. But you could check out the LM3421 or the LM3424 if you need a boost. They can control a series disconnect FET which will eliminate the leakage when shut down, it also allows you to PWM dim much faster, up to 10s of kHz. They can also analog dim if you would rather. The only other way to get rid of the leakage at high input voltage is to use a buck-boost topology (or SEPIC or Cuk, they are also buck-boost). The TPS92690 and the LM342x parts are all capable of doing some form of buck-boost.
  • 0 in reply to Clinton Jensen
    Expert 1425 points

    Skimming over the datasheet for those two parts, it's not initially clear how a series disconnect FET might be operated by the IC...?

    I think I'll look at the buck-boost topology right now. As for the analog dimming, what determines the value of the adjustment resistor? I notice it's usually set at 1M Ohm, however I'll have to control it digitally and digital potentiometers generally don't go that high without being costly. Would 100K do the same job, or would the lower value have and adverse effect? 

    Edit: I've also noticed that the LM3429 mentions an enable/disable feature but don't explain how this might be implemented? Would it be as simple as using a FET to ground the nDIM pin to activate the UVLO?

  • 0 in reply to Craig Easdale19
    TI__Mastermind 25500 points

    All of those datasheets have a section dedicated to analog dimming and PWM dimming. I'm not sure where 1Mohm comes from? You shouldn't need anything even close to that, closer to 10kohm range. And yes, you can use the nDIM pin for both PWM dimming and for an enable/disable.

  • 0 in reply to Clinton Jensen
    Expert 1425 points
    I read the section on analog dimming where it suggests the use of a series potentiometer on the CSH pin. In the application schematics further down the datasheet, the BOM lists 'Radj' (for analog dimming in this manner) as 1M Ohm.
  • 0 in reply to Craig Easdale19
    TI__Mastermind 25500 points
    Ah yes, I see. That is to get as much dimming range as possible for that way of doing it. You could use 100k but you would just have limited dimming range. There is another method and an example shown in the LM3421/23 datasheet that uses a 100k pot and will give you much wider range (and could be used with any of the LM342x parts). Another option is the TPS92690 since it has a dedicated analog adjust pin. You can use a resistor/pot divider with a resistor from VREF and a pot to ground. You could use a much lower value for that.