[SOLVED] How to use ferrite bead's on MCU SPI outputs

[Knife]

I'm designing a system which use SPI comminication. I wanna use ferrite bead for EMI issue on SCK and MOSI output. I searched on internet and I found this circuit. I think L is a ferrite bead.

My clock frequency is 1 MHz. Could you help me to choose ferrite bead on SPI output ports ?

 

[Borber]

Yes L is SMD ferrite bead. Usually straight wire surrounded with ferrite gives inductance of 1uH.

 

[Knife]

I have two idea for solution. I need 4 channel protection. One of them is discrete solution like above, tvs, fb and cap. Other one is compact solution. Could I use NUF4220MN ?

Which situation do I have to handle when I make choise ? Could you give me a direction ?

https://www.onsemi.com/pub_link/Collateral/NUF4220MN-D.PDF

 

[schmitt trigger]

If you can reliably solder a DFN8 package, and the board space is limited, I would go for the NUF4220MN.

 

[dick_freebird]

Looking at that schematic I have to wonder if much
of the EMI problem isn't of your own making.

There's EMI, and there's EMI. Thinking only in voltage
mode you might think those huge shunt caps are a
good idea. But what you're doing, is ensuring a fat
slug of current on every transition since you put them
bare to the MCU output. Add any loop-area to that
and you've got a pretty good radiator of energy.

A series resistor between output pins and shunt cap
would help that, some. On the other hand I kind of
doubt that voltage-mode risetimes are fast enough
for a ferrite bead to knock them down much, if you
are putting them in place of those shown inductors.

I think you'd be better off with a close-in RC filter
with a corner about 3MHz, where C includes the
driven load.

 

[Knife]

Thanks all for answers. I decided to use NUF4220MN or same device because of ESD and EMI are included one package.

But I want to know about how to choose corner frequency of filter ? Always is it be 3 times of working frequency or there is difference situation for all application ?

 

[FvM]

Looking at that schematic I have to wonder if much
of the EMI problem isn't of your own making.

I don't hear an actual report of EMI related problem. It looks like the OP thinks about possible EMI problems or has been to told to care about.

I agree with the RC filter suggestion for data lines, overvoltage protection would be another thing. My first doubt is that SPI isn't intended (and better shouldn't be used) as an off-board interface. So I wonder what's the actual communication range and if we have to fear external interferences or surge voltages at all?

Ferrite beads give second order filter slope compared to first order slope of RC low passes, it might be reasonable in some case for data lines, but overshoot, ringing and data corruption can easily occur with not well designed LC filters. I have e.g. used ferrite bead filters to cut the harmonics of a system clock, quite effective if the the target devices can accept a sine waveform as clock.

 

[Knife]

We have a relay network which have 20 pcs one string as you can see below. We use this system an outdoor application. We wonder about outdoor hash condition like surge, ESD, EMC issues.

The system doesn't use sine clock wave.

 

[Knife]

I solved this problem with "Murata EMI Filter Selection Simulator". Thanks for helping.

 

[schmitt trigger]

I solved this problem with "Murata EMI Filter Selection Simulator". Thanks for helping.

Thanks for coming back and advising us about the successful resolution. This is excellent forum etiquette.
That is the only payment that we desire for our advice.