• State TI Thinks Resolved
  • Locked Locked
  • Replies 9 replies
  • Answers 1 answer
  • Subscribers 63 subscribers
  • Views 1508 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.

LM3424-Q1: Start-up issues at low LED current or overall low LED power

Hendrik Ohm
Prodigy 50 points
Part Number: LM3424-Q1

We have created a design where we can switch the led driver to either drive 2 leds at 12V and 2.45 Amps total (approx. 30W led power) or drive 2 leds at 6 V and 1 Amps total (approx. 6W led power). We have worked quite a bit to find the right configuration to have both strings operate stable over the full input voltage range of 8 to 33 VDC. However, we have one remaining issue. When operating the low power string we have to take care that the led current setting (by resistor at CSH) does not fall below 1.05 Amps. Below 1.05 Amps led current the driver does not start up over the full input voltage range. It will start up in a voltage window from approx. 12.5V to approx. 25V but outside that window it will not start up. Also, we encounter a delay in startup if the driver has been operated before. Means, we have to wait a couple seconds after operation until we can restart the device. Once started, it operates stable over the whole range. We are not sure why and how this comes together.

The used topology is a SEPIC driver. We are using 50 mOhm for the led sense with 22.8k at CSH;

R(hsp)=R(hsn) =1k
R(fs)=10R
C(fs)=100nF
C(cmp)=C(ss)=470nF
R(lim)=33mOhm

Operating frequency is 480 kHz. Any ideas why we encounter these startup issues?

Thanky you!

  • 0
    TI__Mastermind 28655 points

    Hello,

    I've started looking at this and have a few questions to start out:

    How are you adjusting between the two current thresholds, it appears you are doing something with CSH?  Is CSH resistor 22.8 Kohm for 1A or for 2.45A?  Are these two different designs and you are just changing the CSH resistor?  If so what are the two values?

    If you have a circuit tied to CSH what is it?

    Regards,

  • 0 in reply to Irwin Nederbragt
    Prodigy 50 points

    Hello,

    yes, that's correct. We do have a little switching circuit tied to CSH. For operating mode A (1 Amp led current) we just have a resistor from CSH to GND. The value is 22.8 kOhm. When switching to operating mode B (2.45 Amps led current) we are activating a BSS138 N-channel mosfet which connects a resistor value of 17.4 kOhm in parallel from CSH to GND so that we have approx. 9.87 kOhms in total from CSH to GND. All three components are located very close to the IC and share the same GND with the IC.

    In general this method appears to operate very stable. Once operating the voltage drop over R(sns) varies only about 0.1 mV over the whole input voltage range. The only remaining issue is the startup at the lower power condition (mode A). We have difficulties with turning the device on at led currents below 1.05A. The problem disappears when the led current setting is 1.05 Amps or higher...

    Best

    Hendrik

  • 0 in reply to Hendrik Ohm
    TI__Mastermind 28655 points

    Hello Hendrik,

    Seems this should work.  What you are trying to do is basically analog dimming down to one new operating condition which has been done many times before.  I calculated a few things:

    I LED high = 2.513A

    I LED low = 1.088A

    How are you getting this down to 1.0A, are you changing the CSH resistors or doing something else?

    What capacitor value to you have for VCC and Vin?  Are you using the thermal foldback?  If so what are you using for Cref and CNTC?

    If you have an oscilloscope available it would be good to look at a few signals:  Comp, SS, CSH, EN and nDIM, Vin and Vout and Vovp during the issues you see.

    What value is your inductor, is it a coupled inductor or two separate inductors?

    Regards,

  • 0 in reply to Irwin Nederbragt
    Prodigy 50 points

    Hello Irwin,

    first of all thanks so much for taking care of this issue. Regarding the led current you are absolutely right. With the setting of 22.8k the led current is slightly above 1 Amp. This is already the value that includes the safety margin to assure startup over the whole input range. The original value was 24k for ~1 Amp but that was creating the startup issue as discussed.

    The capacitor at Vcc is 2,2µF. The input capacity is 2x4.7µF plus 1µF directly located at the Vin-Pin. We are using two seperate inductors with 33µH each (Wuerth Electronics 7443551331).

    Please find below the screen shots from the oscilloscope as requested.

    Comp:

    SS:

    CSH:

    EN=nDIM=Vin:

    Vin:

    Vout:

    Vovp:

    I hiope this will help to find it.

    Best

    Hendrik

  • 0 in reply to Hendrik Ohm
    Prodigy 50 points

    I am sorry, I mixed up one of the screen shots above. Please discard the single shot of "Vin"  - that shows "Vout" again....

  • 0 in reply to Hendrik Ohm
    TI__Mastermind 28655 points

    Hello Hendrik,

    The person that is more familiar with this part is out until next Monday.  I looked at the oscilloscope pictures and noticed Vout is fairly low when it's trying to start.  Is the small load 6V at 1.05 amp or is it a lower stack voltage.  Vout looks like it's rising to approximately 2.4 volts at the peak.  I need to find out about operating in this configuration.  The other waveforms appear to be okay except comp.  The comp pin voltage doesn't seem to be operating correct.

    Regards,

  • 0 in reply to Irwin Nederbragt
    TI__Mastermind 28655 points
    Hello Hendrik,

    I have something to try. Can you raise the 50 mohm LED sense resistor value and change the CSH resistors to get back to the correct currents? The CSH current is fairly low for running at 1A and trying to start. Maybe 70 mohm and scale the CSH resistors down by that much (17.8 Kohm for 1A and 11.8 Kohm in parallel for 2.5A) to see if it works better.

    What does the output voltage versus output current look like? The error amplifier gets disabled when HSP/HSN are pulled to GND, they need to be above 3V for the error amplifier to operate correct.

    Regards,
  • 0 in reply to Irwin Nederbragt
    Prodigy 50 points

    Hello Irwin,

    this is it. This works a lot better. Have set the resistors to the values that you have proposed. Any idea why this is? Is I(csh) getting too small with values higher than approx. 22k at CSH?

    Best

    Hendrik

  • 0 in reply to Hendrik Ohm
    TI__Mastermind 28655 points

    Hello Hendrik,

    I believe you are trying to start into a relatively low output voltage with low CSH current so it's having a hard time starting.  Glad it's working now.

    Regards,