• State Resolved
  • Locked Locked
  • Replies 3 replies
  • Subscribers 62 subscribers
  • Views 610 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.

TLC5946 Ground Noise Issue

david greenberg
Prodigy 30 points
Other Parts Discussed in Thread: TLC5946

Hello, I'm having a problem with ground noise in a circuit I've made using a TLC5946.  I have connected a the TLC5946 to a microcontroller that is sending it data.  The LEDs connected to the TLC have a power dissipation resistor between them and +V so that the TLC5946 will have to dissipate less power.  When I set +V to be the exact turn-on voltage for the leds (or a fews 10s of mV above that) the circuit functions properly.  As I increase +V by 1-2V, the LEDs become very flickery and I see ~2.5-3MHz spikes on the chip's ground, which have total amplitude of 1V to 1.5V, with the severity increasing as +V increases.  I am running the grayscale clock at 1MHz.  If I send animation (time-varying data) to the TLC5946 with a +V above the threshold, it seems to flicker for <0.5 seconds, and then some number of the outputs dim or turn off, and then the animation continues.  I suspect that this is because the ground spikes are causing the MODE line to falsely toggle, thus setting the prescaling values to turn off several of the outputs, and then the noise is reduced due to reduced current consumption.  Do you have any experience with this issue, or any ideas on how to debug it?

  • 0
    TI__Mastermind 36110 points

    This is probably a layout issue.  You may have large switching currents in the ground plane flowing past the more sensitive analog ground currents associated with the control signals.  When your +V voltage is very low, the TLC5946 probably turns the outputs on and off slower than when the +V is higher.  The lower di/dt in the ground plane generates less noise.  The solution will be to add capacitance to some strategic locations to keep high frequency currents contained, reroute your grounds, or filter the MODE pin.  This will probably be a trial and error solution

  • 0 in reply to Matthew Day
    Prodigy 30 points

    I tried using 22uF bypass capacitors across the TLC5946's VCC and GND, to no avail.  I also added an additional path to ground (the board is very cramped and has very few grounds, so only this additional path exists).  Could bypassing +V be a potential solution?  Also, are there descriptions on how the TLC5946 works internally when there is a nonzero potential between OUTn and GND?

     

    Thanks,

    David

  • 0 in reply to david greenberg
    TI__Mastermind 36110 points

    There is no description of how the IC works with different voltage on OUTn.  The IC requires a minimum voltage (specified in the datasheet) on OUTn to meet specifications.   Adding additional capacitance may help.  Specifically, you need to follow the path of an electron as it flows from the input voltage through the LED, into the LED driver, to ground, and back to the input.  You must place capacitors so that they provide a very short path for high di/dt currents to flow.  As an example, if you have no bypass capacitance at all, the high di/dt currents must come from the output capacitance of the +V power supply.  If these currents must flow through a long path, they will create voltatge drops due to resistance and parasitic inductance.  Placing the capacitors near the high di/dt load minimized the physical loop that the currents flow through.  Each IC must have a good quality ceramic capacitor connected directly from Vcc to ground.  You should also have good quality ceramics, plus some bulk capacitance at the anode of your LEDs to provide the high di/dt current to your LEDs.