• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Answers 1 answer
  • Subscribers 62 subscribers
  • Views 1608 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.

TPS92641: Glitches and oscillations in TPS92641 PWM output

Niyas Khader
Prodigy 40 points
Part Number: TPS92641

I am currently working on a 100W LED driver application with TPS92641.

The application scenario is as follows:

- Input voltage: 50V Nom (46-52V)

- Output voltage: 31-34V (single LED)

- Output current: 1.5A - 3A

- Switching frequency: currently 150kHz

- UVLO Threshold : 46V (Turn ON at 48V)

- Dimming: Analog

Schematic for reference.

 

I set UVLO threshold to 46V and turn ON voltage to 48V for our prototype boards as above circuit, we observed small glitches in PWM output (waveform-1) in this configuration. I am able to  resolve this issue by increasing RUDIM3/R19 resistor to 20K or above (Waveform-2), however UVLO hysteresis increased to 10V and also PWM frequency jumping in to 2MHz (Waveform-3) when the input voltage is near to UVLO threshold.

 

Waveform 1- PWM output at normal working voltage of 48V with RUDIM3 3.24K Ohms

 

 

Waveform-2 PWM output at normal working voltage of 48V with RUDIM3 20K Ohms

 

 

Waveform-3 PWM output at near under voltage region

 There is an oscillation in the PWM when output voltage is near to OVP threshold(Waveform-4).

Waveform-4 PWM output at near to OVP threshold region

 

Regards,

Niyas

 

  • 0
    TI__Mastermind 25500 points

    Hello Niyas,

    There isn't a ton of info to go on here since the schematic is partial and I can't see the SW node, UDIM, or input voltage.

    But one thing you don't want to do is operate right near the thresholds of UDIM or OVP.

    UVLO using UDIM is meant to keep from operating below the output voltage (not necessary for all applications). So if you want to use it I would set it at a point a bit above the maximum output voltage. If you have an LED with 36V max Vf I would set the UVLO level at around 38V to 40V, not near the nominal operating range. It is a dual function pin for fast PWM dimming so if there are any glitches or noise on it then it can turn the high side switch off and back on very quickly. Some capacitance from UDIM to ground can really help, but I would still recommend on operating so closely to the threshold if at all possible.

    The OVP should not be near normal operation either. It can also very quickly turn the part on and off. The OVP function is more for protecting any output capacitors you may have. I would recommend setting that level well above the max LED voltage, at least 40V (or higher if the over temp max of the LED string is >36V) and then adjust RON if needed.

    Let me know if that helps or not.

    Thanks,

    Clint

  • 0 in reply to Clinton Jensen
    Prodigy 40 points

    Hi Clint,

    Here is the full schematic.

    I will try your suggestions.

    Is there any impact on RUDIM3 resistor value (>10K?) on normal operation? Can RUDIM3 be zero, so that  UVLO hysteresis will retain 2.1V?

    I want to use UDIM pin for under voltage lock out when the previous PFC stage output goes below 45V.

    Regards,

    Niyas

  • 0 in reply to Niyas Khader
    TI__Mastermind 25500 points

    Hello Niyas,

    RUDIM3 can be just about any value, it only affects the hysteresis, and that includes zero or >10k.

    But really the first thing I would try is a lot more input capacitance. Buck regulators have discontinuous input current, so they produce a lot of noise on the input and that noise will be seen at UDIM as well through the divider. At such high power levels 2.2uF + 1uF is probably not nearly enough, especially at 150kHz switching frequency. I would try at least 10x that, but more likely I would add 2 or 3 22uF capacitors at the input. That alone may solve your problem. It also wouldn't hurt to add a 1nF capacitor from UDIM to ground for a little extra filtering.

    Regards,

    Clint

  • 0 in reply to Clinton Jensen
    Prodigy 40 points

    Hi Clint,

    We have 2 x 220uF Aluminum capacitor at the input of LED buck driver (at PFC output, previous stage). We tried 2 x 10uF ceramic capacitor at the Buck input and 1nF to 10nF from UDIM to ground, but oscillation near UVLO threshold still exists.

    We observed the same issue in TPS92641 EVM board also. Could you please verify this and confirm?

    Regards,

    Niyas

  • 0 in reply to Niyas Khader
    TI__Mastermind 25500 points

    Hello Niyas,

    The UVLO function really is more to ensure operation doesn't start until VIN > VOUT. It isn't intended to protect the input supply, but it can be used for that. But it still should not be operated right near the threshold or oscillation can occur. If you need to have a set level very near a UVLO threshold you may want to use an external comparator for the UVLO function and feed its output to the UDIM pin.

    Regards,

    Clint

  • 0 in reply to Clinton Jensen
    Prodigy 40 points

    Hi Clint,

    We implemented external UVLO comparator and fed to UDIM pin, under voltage cut off  is functioning without any issues now.

    Thanks for all the suggestions and support.

    Regards,

    Niyas