5
\$\begingroup\$

I want to use a charge pump to convert 5V to regulated 3.3V output. The charge pump I want to use has the switching frequency of 1200 kHz max. Do I need to be concerned about this switching frequency to avoid EMC issues? How is the EMI from charge pump in comparison to inductor-based regulators?

\$\endgroup\$
  • 1
    \$\begingroup\$ Why can't you just use a linear regulator and, thus, get rid of any EMI issues? \$\endgroup\$ – EM Fields Sep 7 '14 at 11:07
1
\$\begingroup\$

EMC = Electromagnetic Compatibility. EMI = Electromagnetic Interference.

You can try to make it compatible by using smart PCB layout techniques. But it depends on your goals and final application. New buck converters are smaller, their EMI radiation is smaller as well (EMI is proportional to line length of the components). Most of the EMI you have to worry about is from the charge/discharge in the feedback loop. So, it is very important that you get your layout optimized for reducing it all. For example,

  • Large ground planes to take all that corrent away.
  • Parallel traces/caps to ground or V- in order to cancel noise and EMI.
  • Choosing caps to match your application.
  • You can also shield your converter to prevent noise to nearby elements.
  • If you have large Inductor, you might want to consider placing another one on the opposite side of the PCB to cancel the noise/EMI it can cause. You can use also Ferrite Beads to get rid of some of the EMI.

Bottom line, I guess it all matters how much noise/EMI you are willing to accept, and how much you are willing to invest in getting rid of it. As far as Inductor based switches are concered, I used both and aside of the size, I did not notice much difference. I guess in my case it all boiled down to the design.

|improve this answer|||||
\$\endgroup\$
  • \$\begingroup\$ Since this is a charge pump, it doesn't need inductor. But didn't find details about EMC in datasheet. \$\endgroup\$ – zud Jun 17 '14 at 15:28
  • \$\begingroup\$ Well, think about this way - lets say you design a two layer PCB. If you lay the traces incorrectly - you are at risk of creating antennas (long traces) or capacitors (parallel traces). Usually datasheets do not provide EMI information. But if you will look on page 18 - the PCB layout is optimized for output voltage ripple performance. The less ripple you have the less noise and disturbance you will see across the rest of the circuit. On the same note, some people create seperate circuits on the same PCB, and just connect ground planes via 1M resistor to isolate noise and etc. \$\endgroup\$ – KingsInnerSoul Jun 17 '14 at 20:03
0
\$\begingroup\$

I would guess, that - in terms of radiated emissions - a charge pump circuit will perform better than a switched inductor. But since you draw pulsed current, the conducted emissions might still be a problem. I recommend to add a filters not only to the output but also to the input.

Depending on the current you are going to consume - with the low voltage difference in mind - I would also consider using an LDO.

|improve this answer|||||
\$\endgroup\$

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.