TPS92691: LED Power Supply working but still has noise/EMI

Part Number: TPS92691

Hi.

I am working on a Power Supply for LEDs. It's a single Power supply that multiplexes 6 PCBs LEDs. The Power Supply it's working fine, I can succesfully dimmer the power of the LEDs, even 100% duty cycle. But the issue I am experiencing is that when one of the LEDs is on, there is a lot of noise spread all over the board.

Sch_PowerLED.pdf

BOM_LED (1).xlsx

At first I thought that the source of the noise was the switching between leds, but if I keep only one LED turned on, I still get a lot of noise.

The following images is oscilloscope images, the cyan is the diode anode/dreno/inductor node(D13,Q11,L2) and the dark blue is an GND measured and two different point the PCBs. Another fact is everytime I turn on the LED I lost communication with UART devices and with the STM32 debugger.

Noise-20210504T115320Z-001.zip

  • Hello,

    If you have specific question about he TPS92691 in application then we can help you but as far as losing communication with your UART device ect we cannot help with that.  Usually, with EMI it's layout and system level design.

    Note that we cannot read the Google share drive that you have attached due to our TI firewall.  You need to attach the file or the image in order for for us to read it.

    Thanks Tuan

  • Hello,
    Is there any advice for improvement for my design? Also I updated the links attaching the files.
    Thanks

  • Hello Paulo,

    The noise are typical switching noise that is associated with switching power supplies.  If you have the ground of your scope probe on ground and the tip of your scope probe to another ground away from the board do you see the same noise signature?  If you do then your scope is picking it up with the ground loop and this is typical of switching power supplies.  Usually, this is not the cause your resetting issue. 

    What you might want to change is the SS cap that is tied to pin 2 of the TPS92691 from 22nF to 0.1uF for longer soft-start time to see if it helps.  This cap will prolong the soft-start time so your input power will not experience high current with short soft-start time.  I would also monitor the input voltage to see if it droops too low and that's why the UART gets reset.  

    Loosing communication with the UART is a system level issue with the UART communication and I cannot help you much with that.

    Thanks 

  • Hello,
    Yes, the noise signature was measured in two different GND points of the PCB. I'll try chaning the soft-start time. Other things I'll look is the MOSFETS gate capacitances and the gate charge curve.
    Thanks for the advice.

  • Hi Paulo,

    The Soft-Start cap is the only thing I can help you with the possible issue with your UART resetting here.  Other than that you will have to trouble-shoot the UART directly.  I'll close out this thread and if you have an issue related directly to the application of the TPS92691 then you can re-open it.

    Thanks Tuan

  • I solved the problem changing the MOSFET for a model with less internal capacitance. Other thing I did was putting an Rg resistor in the gate (49.9ohm). If someone wants to know more about try searching for ringing mosfet noise.