0
\$\begingroup\$

My project has two micros.

One is an ESP8266 - WiFi radio and the other a general purpose micro.

The power to the ESP needs to be clean.

I initially looked at a pi filter using a ferrite bead but then found several articles that state the bead should not be used in the power line to a micro. The resistance of the caps at the noise freq does not supplement current enough to compensate what the bead is blocking (from a TI paper and another post in this forum on ferrite beads)

Can someone suggest how to make sure the 3.3v line to the ESP is clean?

The other micro on the board .. sitting about 1.5 inches away, is running off an internal 48 mhz clock and has several SPI lines that are clocking at 1 mhz.

Thanks

\$\endgroup\$
2
  • \$\begingroup\$ "The resistance of the caps at the noise freq does not supplement current enough to compensate what the bead is blocking"- That statement seems too general. Which bead? Which caps and what capacitance? Seems like if it's designed properly you could make it work. Otherwise maybe a more traditional filter with an inductor rather than a ferrite bead (if you know the frequency or frequencies you're trying to reject.) Maybe just good quality decoupling caps will be enough, do you have any reason to believe otherwise? \$\endgroup\$
    – John D
    May 3 '16 at 18:18
  • \$\begingroup\$ The statement is general. I paraphrased it from another post on this forum as to why ferrite should no be placed in micro power lines. I don't know if I have an issue. I am just designing the board and am aware of the possible issue so I wanted to understand the design options (besides a standard bypass cap) that could be used. I'm trying not to have to make a rev2 on the board to address an issue I'm aware might exist. \$\endgroup\$
    – JHinkle
    May 3 '16 at 18:30
1
\$\begingroup\$

The fact that plenty of people have used an esp by just jamming in onto an arduino board with no special filtering or anything tells me that the esp has a better tolerance to supply noise than you might think. The average micro draws next to nothing (and will probably be using it's internal regulator to feed its core anyway, giving you extra isolation anyway). Even in mobile phones (where the radios are more sensitive and the processors draw way more power, a few ceramic capacitors is usually all that gets used, maybe a ferrite bead on the radio but that'd be it, as long as you have sufficient capacitance at the load end, you can put a ferrite bead anywhere, it's when there's not enough bulk capacitance that ferrite beads cause problems as they act like high value resistors to high frequencies.

\$\endgroup\$
0
\$\begingroup\$

You may use 0 ohm resistor in series with the ESP and power source. Also make sure the LPF cap of 0.1 uF 50 (or 100) VDC CER Disc is placed very close to the ESP on board. Also it may be helpful to allow some settling time for the data communication between the MCU and chip. At times the ESP may need some time to settle depending on what you are trying to accomplish. For example ... frequency tuning and other labour intensive operation don't settle fast - so give this baby some time to relax before you push it hard :)

\$\endgroup\$
1
  • \$\begingroup\$ Thanks for your replies. I'm not an arduino guy. The main micro is an arm. I'm laying out the board now - all SMD. I will have adequate capacitance at the ESP's power in and add a zero-ohm resistor in case I need to switch in a ferrite bead after the fact. \$\endgroup\$
    – JHinkle
    May 7 '16 at 2:24

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.