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.

LM3406HV-Q1: Remote LED EMI concerns

Part Number: LM3406HV-Q1
Other Parts Discussed in Thread: TIDA-050030, TIDA-050029, TIDA-050040, PMP15004

Hello,

We are currently evaluating various possible solutions for lighting a high-power LED with 1 A flow current in automotive environment.

The LED will be mounted on a separate PCB, which is 20 cm to 30 cm away from the driver board. 

Probably the LED PCB has to be so small that there would be no space for an output capacitor on the LED PCB.

That and the fact of the long wires with high current (+ ripple current) is why we are concerned about EMC

Is the current through the LED one of the critical paths regarding EMC? We are concerned that the ripple current through the long wire will generate a lot of noise end emissions.

Another solution would be to use a DC DC buck converter which generates a constant voltage in the range of the forward voltage of the LED. The constant current would then be generated by linear drivers, thus avoiding an AC component on the long supply line to the LED. 

Would you prefer the second solution to the first with regard to EMC?

Thanks a lot!

  • Hello,

    Yes, the current through the LED is one of the critical paths regarding EMC, both differential mode and common mode.  How much effort in reducing this depends on the level of EMI trying to be achieved.  A capacitor across the output at the drive reduces differential mode noise, common mode noise is dealt with a different way (common mode choke, bypass capacitors from both LED leads to the enclosure, etc.).

    Creating a DC voltage with a linear on the output will reduce low frequency differential mode noise, high frequency may still get through.  Depending on the layout and design EMC still can be an issue, both common mode and differential mode.

    Best Regards,

  • Thanks for your answer.

    So you think both solutions would need input and output filters (depending on EMI level tried to be achieved)?

    Would using a shielded cable from driver PCB to LED PCB a plus?

  • Hello Clemens,

    Yes, depending on the level of EMI trying to be achieved.  PMP15004 is a direct from battery to low voltage LED string that meets CISPR25 Class 3 levels with our local unofficial testing without an enclosure.  It needed a common mode choke on the output to get there (common mode noise).  I took this design and placed it in an enclosure bypassing the inputs and outputs to the enclosure to meet CISPR25 Class 5 again our own test equipment.  We have several TIDA designs that are tested to EMC, conducted, such as TIDA-050029, TIDA-050030, TIDA-050040.

    Shielded cables do work but they really need to be tied off on both ends to the chassis.  di/dt noise cannot be shunted by the shield if one end is open since there is no continuous conduction path for the shield.  We don't use shielded cables to meet these levels since many customer designs cannot use this such as LEDs on a long PCB for turn signal or daytime running lights.  It makes more sense to try and make the input and output look like pure DC.

    Best Regards,

  • Hello Irwin,

    thanks for your answer.

    So did I understand correctly that with a design like PMP15004 or any other design with very clean input and output, a relatively large load or a load placed far away will not significantly degrade EMI performance?

  • Hello Clemens,

    If the output is filtered well/correct for EMI reduction there can be leads to the LED string.  It is typical in lighting applications for automotive to have a wire harness and exposed PCB traces from the LED driver and some of these solutions require very stringent EMI limits.  For PMP15004, to reach the conducted EMI levels required, a common mode choke was needed on the output.  There are other ways to meet this but it takes time trying and testing these solutions.

    Best Regards,