Link to home
Start Free TrialLog in
Avatar of AJAXNoob
AJAXNoob

asked on

I have a SMPS that has a 1.4MHz transient I'm trying to shunt to ground. How do I select the correct capacitor to filter 1.4MHz?

I have built a circuit.  I took a snapshot of the waveform.  It looks like there is a transient with a frequency of ~0.75 micro seconds (~1.4MHz).  Any idea how I can remove the transient from my circuit?
User generated imageUser generated image
Avatar of d-glitch
d-glitch
Flag of United States of America image

You should make sure you have a smaller, high quality ceramic capacitor in parallel with the large electrolytic.  Try  0.1, 0.22, or ever 1.0 uF.

Large electrolytic capacitors tend to have poor high-frequency characteristics.
Correcting  ....

Is that a real scope output or a simulation?
What are the amplitude of the 1.4 MHz and the nominal output voltage.

The natural frequency of the LC components you have chosen is 2 kHz.

So you would expect to see approx 0.2% ripple on the output.
If this is more than you can tolerate, you may need to add additional filter stages.
Avatar of AJAXNoob
AJAXNoob

ASKER

That is real scope output I captured.  Amplitude (peak to peak) is 120mV.  Nominal output is 5VDC.  Is there a way to calculate the correct size ceramic capacitor to use?  I used this calculator to determine the correct capacitor (http://www.midnightscience.com/formulas-calculators.html).
How big is the input voltage to the SMPS?
And how much ripple would be acceptable?

For the ceramic capacitor...   Try any value you have and see if it makes a difference.
0.1uF should me very common.  Make sure the leads are short.  

And think about the grounding.  
Is this constructed on a PC board or one of those white prototype boards?  
There are inevitably large current pulses between the input power supply and the switch.  Make sure these are contributing to the output noise.  You may want to include some imput filtering as well.

If
The input is 48VDC.  Ideal peak to peak ripple would be 20mV.  Current peak to peak ripple is 30mV.  I made the circuit on a PCB board.
ASKER CERTIFIED SOLUTION
Avatar of d-glitch
d-glitch
Flag of United States of America image

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
I have empty pads to add additional capacitors in parallel, but didn't leave space for an extra inductor.  I might have to think about a redesign to test the added inductors.
Adding a second cap in parallel will give you a factor of 2 improvement.
Adding a second LC stage like the one you already have would give a factor of 800.
I'll try to add a second LC filter over the next few days.  Thanks!