Good afternoon fellow students of life,

I have a particular question that relates to the use of a Duracell USB switching power supply as the main supply for a high sensitivity sensor board. Most of the noise is below 3.58MHz going down to 300 kHz.

enter image description here

and has this waveform.

enter image description here

The schematic uses this configuration to try and eliminate the noise by blocking it in the first place. The chip inductors are separate.

enter image description here

A variation on this that I found on the web was to use a decade of capacitors (100uF, 10uF) in front of a ferrite bead with an electrolytic, taltanium, and ceramic (10uF, 47uF, 0.1uF) after.

http://www.ti.com/lit/an/scaa048/scaa048.pdf and http://www.ftdichip.com/Support/Documents/AppNotes/AN_146_USB_Hardware_Design_Guidelines_for_FTDI_ICs.pdf

I know that there is no perfect solution to this problem

Is this a better way of accomplishing noise suppression?

As you can see there is an extra inductor (L3) to decouple the AGND and the DGND.

Thanks ahead of time.

  • 3
    \$\begingroup\$ You may be restricting answers by addressing this question to "fellow students of life". I don't feel qualified on this basis to provide an answer because at my age I'm more concerned about death. \$\endgroup\$ – Andy aka May 29 '15 at 14:14
  • \$\begingroup\$ That's kind of morbid. I was addressing the community as a whole since we should be continuously learning. Even before "death". I would still like to have your comment on the topic of filtering a noisy USB source. Although, like already commented, it is not the ideal source. \$\endgroup\$ – MiscBob May 29 '15 at 16:14

Using a switch-mode regulator (in the Duracell USB PSU) for "a high sensitivity sensor board" is your problem. Sensitive analog electronics quickly succumbs to the noise inherent with such regulators. You can go to town with filtering (and simply throwing a cap of every type is not the ideal way to go about it), but some still gets through - whether such efforts are enough for your needs can't be gleaned from what you've told us.

It's traditional to add a linear regulator stage between the raw switch-mode PSU and analog circuitry. If your analog circuitry will happily operate at less than 5V, then a LDO (low drop-out voltage) linear reg can be added, but this assumes a lot about the analog circuitry (head-room, voltage reference in an ADC, etc).

| improve this answer | |
  • \$\begingroup\$ There are LDO's later down the line to help with further filtering. Actually there are two in the sensors pathway. The measured sensor is a very low current sensor. The LDO linear regulators are a REF3040 and a REF3020 with a RC filter in between. I was just wondering is there was a better way to accomplish the filtering at the source in a more efficient manner and if it is necessary to have the inductor on the ground path immediately from the USB. Also would a low impedance capacitor in front of the inductors help or cause more harmonics? \$\endgroup\$ – MiscBob May 29 '15 at 16:11
  • \$\begingroup\$ ok. so do you actually have a problem with this 300kHz (& its harmonics) getting into your analog readings (i'm guessing yes, or you wouldn't be here ;) ? if so then I'd be looking at tuning your LC filtering to target these frequencies, rather than just a scatter-gun approach thus far. \$\endgroup\$ – Techydude May 30 '15 at 1:37
  • \$\begingroup\$ I added an RC filter in between the regulator and the 5Vusb after the inductor. Seems to work fine. Thank you for your help. \$\endgroup\$ – MiscBob Jun 7 '15 at 2:45

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.