0
\$\begingroup\$

I'm planning to buy the following line filter: http://uk.rs-online.com/web/p/iec-filters/4540956/

And I was planning to check the differences between 10A or 15A models. But I have difficulty to understand the "Typical filter attenuation" graphs in the third page of the datasheet:

graphs

Aren't these supposed to attenuate the high freq. noise? I'm aiming to filter 70kHz switching noise coming into a system through AC lines.

At 70kHz it looks like the filter graphs are flat. In some graphs the attenuation is -dB at 70kHz.

I'm confused if it would make a change if I use it.

\$\endgroup\$
4
\$\begingroup\$

I think you may be somewhat underestimating the problems caused by a local switching regulator. There is a very good site that I took this diagram from - it shows the noise produced by a switching regulator operating at 280 kHz (4 times higher than yours): -

enter image description here

As you can see, switching noise is not contained just to those frequencies around the switchig frequency. So, the line filters you are considering are going to be pretty good at attenuating all the noise in the 100 kHz to over 10 MHz range.

However, you may still have problems above 10 MHz because the graphs you show don't specify that range. For this region, ferrite beads are generally accepted as good solutions but, they still have to be designed with care: -

enter image description here

enter image description here

And, if you cascaded two ferrite bead circuits you would get a significantly improved response: -

enter image description here

enter image description here

So, I think that 70 kHz may be the smaller end of the problem even though it appears to pose the biggest threat in terms of amplitude. I say this because if your product has some form of voltage regulator (and I know your input is AC) then, this by itself is likely to have sufficient rejection at 70 kHz to significantly reduce its effect.

All the great images above taken from this superb site.

\$\endgroup\$
  • \$\begingroup\$ I found the damn source. It was a power supply(probably SMPS type) of an LED lighting system. It is probably called common mode noise if Im not wrong. \$\endgroup\$ – HelpMee Dec 14 '16 at 16:04
1
\$\begingroup\$

Neither of those filters attenuate much at 70kHz. It's your job to add filter components to your design to handle emissions at those frequencies. The filtered inlet manages higher frequencies.

If you've ever pulled a TV or PC PSU apart, then you'll have seen a small PCB often used behind a filtered inlet, with a few capacitors and one or two wound toroids on it, handling frequencies lower than the inlet does.

\$\endgroup\$
  • \$\begingroup\$ do I have to design a choke filter for that specific switching freq.? is that a kind of notch filter? \$\endgroup\$ – HelpMee Dec 14 '16 at 10:06
  • \$\begingroup\$ I also found this one: uk.rs-online.com/web/p/power-line-filters/8456827 But they dont have freq. response graphs \$\endgroup\$ – HelpMee Dec 14 '16 at 10:20
  • \$\begingroup\$ That sort of component costs a lot because it's designed for the system designer who has just presented a system for sale, and it's been rejected because of emissions. They will lose sales revenue until the damned thing is approved, so buying and bolting in something like this is cost effective. It's rather expensive to be a design-in of first choice. Read Andy's post, and follow his links to the axotron site. \$\endgroup\$ – Neil_UK Dec 14 '16 at 12:46

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.