• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Subscribers 62 subscribers
  • Views 931 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.

General Question: Why is the range analog dimming precision so limited for most LED driver chips?

W HS
Prodigy 170 points
Other Parts Discussed in Thread: LM3409, TPS92690, TPS92691

My application requires analog dimming.  In general, I see that some chips only offer 10:1 or 20:1 for analog dimming. For 10:1 dimming, I think that is 1 part in 10 or a resolution of 10%. For 20:1 dimming, I think that is 1 part in 20, or a resolution of 5% with a constant control voltage.

This is two orders of magnitude less than PWM dimming, which provides much more resolution.

What is the  limiting factor for analog dimming? Noise in the driver chip, current ripple, etc?

Why not one of magnitude?



  • 0
    TI__Mastermind 25500 points
    In general LED drivers use relatively low maximum current sense (reference) voltages due to the power dissipation in the current sense resistor (a 200mV reference will burn 200mW in the resistor at 1A where a 1.24V reference would burn 1.24W in the resistor at 1A). So when you reduce this reference voltage the error amp offsets can become significant. If you have +/-5% at 200mV that is +/-10mV. The offset will not change so if you reduce the reference to 20mV with analog dimming you are all of the sudden at +/-50%. Most of them you can analog dim further than what is specified, it's generally just limited to a range where the accuracy is somewhat acceptable for most applications.
  • 0 in reply to Clinton Jensen
    Prodigy 170 points

    Thank you for a concise and easy-to-understand reply. 

    It seems to me that if an external feedback loop  (software with adc/cs resistor/amplifier or analog circuit w/o adc) then resolution can be improved and is only limited by drift and noise. Extra considerations for stability of this loop will be necessary.

  • 0 in reply to W HS
    TI__Mastermind 25500 points
    All error amps have some offset, but of course it is possible to make it very small, but at the cost of die area and increased IC cost. But there are other tricks such as that used in the LM3409 where you swap the inputs of the error amp each cycle to cancel out the offsets. But if you have a fast current rise/fall time using PWM dimming with the higher sense voltage you will always be able to achieve better accuracy and a wider accurate range of dimming.
  • 0 in reply to Clinton Jensen
    Prodigy 170 points
    Yes, understood on the PWM, but it cannot be used on my application.

    Since my input power is a well-filtered DC supply (fully monitored) , and since the control loop input intensity command change rise time is 50mS , and since there is no PWM, I do not require a high-bandwidth control loop. Compensation is only required for DC variation and electrical parameter changes due to LED temperature increase--should be an easier layout.

    On the other hand I need to supply about 4 channels of 90W each at 40V each (ie--no chip with integrated switches)

    If you have a recommendation for a control chip
    -24 VDC input, commericial
    -Boost Topology to 38- 42 V, 2.1 Amp load current (external transistor)
    -Best Analog Dimming Ratio (cant seem to find that in power-bench)

    I would certainly pursue it with Arrow before asking for an FAE. PM me for company contact info if necessary.
  • 0 in reply to W HS
    TI__Mastermind 25500 points

    90W is pretty high. But the TPS92690 or TPS92691 are controllers that should be able to do it and will provide the widest and most accurate analog dimming range of any controllers I am aware of.