• State Resolved
  • Locked Locked
  • Replies 7 replies
  • Answers 1 answer
  • Subscribers 63 subscribers
  • Views 1396 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.

TPS92641: TPS92641 IADJ linearity & SDIM operation causes LED flicker

Scott Grodevant
Intellectual 925 points
Part Number: TPS92641

Hi All,

I have designed a 4-layer (2 oz. copper) PCB using the TSP92641 to drive a single white LED at up to 20A.  The design is more or less a copy of the reference circuit 20A EVM schematic.  The board runs from a 13.8V, 20A supply and the LED's forward voltage ranges from 2.8 to 3.6V.  The IADJ pin is driven from a 2-pole filtered PWM output.  The filters have a nominal 160 Hz cut-off.  The SDIM input is driven by a second (unfiltered) PWM ouptut.

During testing I've discovered that the LED light output is not very linear with the IADJ voltage.  In fact, simply enabling SDIM with IADJ set to its minimum output produces about 600 mA of LED current which represents about 12% of the LED's full-scale output.  In an effort to "linearize" this jump, I am attempting to use the SDIM input.  At very high duty-cycles (SDIM high >90% of the time), the driver behaves as expected and does indeed dim the LED proportionally.  However, as the duty-cycle drops to the 80% level (and lower), the LED noticeably (and randomly) flickers.  The "SW" output (TP6 on the application schematic) shows varying numbers of positive pulses which explain the flicker.

I've attached my schematic (2 pages), a scope trace shown SW (yellow) and SDIM (red) and a graph of the LED output (measured with an external photodiode) versus PWM drive duty-cycle.

Any thoughts much appreciated.

Thanks,

Scott

  • 0
    TI__Mastermind 25500 points

    Hello Scott,

    I do not see the schematic, did you forget to attach it?

    I would also like to know what your IADJ minimum is? Also what is your switching frequency? There will be a minimum on time/maximum off time limitation when analog dimming in some cases since it can limit the minimum duty cycle. If you are at that point PWM dimming of any kind might not work quite right.

    For such high current applications it is generally best to PWM dim (UDIM or SDIM) since your current sense resistor value will be low enough that IADJ can only go so far even if you do not hit D limitations. Then if you need to dim further than PWM dimming can get you then you consider reducing IADJ some to meet the requirement.

    If you attach the schematic and let me know the answers above I will take a look.

    Thanks,

    Clint

  • 0 in reply to Clinton Jensen
    Intellectual 925 points

    Hi Clint,

    I guess my schematic posts were lost, here they are again.

    IADJ can go to zero.  The PWM frequency (for both the IADJ drive and SDIM is 60 KHz) which gives me 1000 steps of control.  I have tried running SDIM at a lower rate, 5 KHz, but that didn't seem to help.  The TPS is switching between 55 and 80 KHz depending on the IADJ level.

    Thanks for your help.

    Scott

  • 0 in reply to Scott Grodevant
    Intellectual 925 points
    I've pasted images of both schematic pages three times now. I don't see them but hopefully you will.
  • 0 in reply to Scott Grodevant
    TI__Mastermind 25500 points

    Hello Scott,

    I do not see anything attached. But with IADJ at 0V I could see there being issues. You basically have COMP at 0V as a result. Generally with shunt FET dimming COMP stays constant and the output current stays constant (but through the FET). But if COMP is at zero any slight perturbation positive or negative could cause the part to not switch at times and to switch at others. Noise sources on CS and COMP due to various reasons including layout could make it worse.

    With shunt FET dimming the ideal when using IADJ is to keep it in continuous conduction mode (inductor current never falls to zero). It can work fine lower but you for sure want to stay out of the noise floor. If using SDIM I would probably keep IADJ at an absolute minimum of 10mV so that CS at 1mV is above the possible +/-600uV offsets. Depending on noise sources that could couple in maybe a bit higher.

    Regards,

    Clint

  • 0 in reply to Clinton Jensen
    Intellectual 925 points

    Hi Clint,

    I gave your idea a try last week and it seems to work well.  Linearity of light output is much improved and will work for our application.

    Thanks for your help.

    Scott