Electromagnetic shielding for DC-DC converter

[han]

hi all,

i'm on DIY differential probe project and have to use dc-dc converter due the size constrain.
The diferential probe work very well, but have a large noise (100 mVpp).
I probe the circuit and found the noise came from DC-DC converter. and it EM radiation,
i have put many capacitor decoupling near the DC-DC converter, and it didn't help
is make cage from FR4 pcb to isolate the DC-DC converter will help?


anybody have a experience for eliminating transient noise from DC-DC converter?

thanks

[T3sl4co1l]

Sounds like an EMC problem to me.

I'd have to see your circuit and layout to offer much help.  And $50/hr...

[Psi]

Try an LC filter on the output of the DC-DC.

I find this works really well for DC-DC powersupplys on my RC plane to eliminating switchmode noise that gets into the camera picture.

RC filter would work too but the voltage loss across the resistor will vary with current and effect regulation, so its not ideal.

[han]

Sounds like an EMC problem to me.

I'd have to see your circuit and layout to offer much help.  And $50/hr...

unfortunately the device it self less then $50 ...

[han]

Try an LC filter on the output of the DC-DC.

I find this works really well for DC-DC powersupplys on my RC plane to eliminating switchmode noise that gets into the camera picture.

RC filter would work too but the voltage loss across the resistor will vary with current and effect regulation, so its not ideal.

i have clean the signal with RC filter and have a clean signal on +/- Vsupply  if i probe using coax soldered to circuit, the problem is the rest of circuit behave like antenna.
the frequency is 500KHz but i doubt FR4 shielding will work, so anyone try it before?

[DaveW]

You can LC filter the input and output (on the output making a pi filter) which works quite well. If you LC filter the input, it can help to add an RC snubber to compensate the the source inductance that has been added.
Layout for SMPSs is pretty critical-something like this is helpful, but you are basically keeping the size of any loops down which are carrying pulsed high currents
http://www.edn.com/design/components-and-packaging/4390811/PCB-layout-considerations-for-non-isolated-switching-power-supplies

Also have a look at the slew rates that your controller is using. You can sometimes slow the switching of the FET which will reduce emissions quite dramatically/

If things require more effort than that, metal cans over the analog input side can help. Also Wurth make ferrite tiles which fit over the critical part of the SMPS which reduces noise