Maker Pro
Maker Pro

EMI suppression

Tim Brown

Dec 31, 2017
23
Joined
Dec 31, 2017
Messages
23
Hi,

I am using some Murata ferrite beads for EMI suppression above 30MHz..These Murata ferrite beads in any particular model are available in various impedances @100MHz. For example 10 - 1000 ohms. I don't understand why anything other than the maximum impedance type would be chosen. Why are there so many impedance types?


There are so many models available that I am really not sure what to get. The attached picture shows six graphs for six devices. The DC resistance and current capability is about the same for all of them, and will suffice for my application.There are three different physical sizes, 15, 18 & 21, and any of those size are OK for me. I need one bead on the VCC line of the 64MHz MCU, and perhaps also on the 4MHz SPI line, for the clock, data in and also data out, I guess.

It seems to me that if I want to limit radiated emissions at 30MHz+ then the best choice is BLM21RK102N1, because its resistive area starts lower than any of the others, at about 7MHz (where X crosses R). But if it is that simple then I don't know why there are so many choices and why anyone would ever choose anything else.


It seems that a device with a resistive curve starting above 30MHz is no good and so none of the devices on the right are satisfactory. But again with everything commercial needing EMI certification for radiating emissions above 30MHz, I don't know why there are so many options above 30MHz.

appreciate any information. thanks.
 

Attachments

  • EMI filter graphs.png
    EMI filter graphs.png
    116.1 KB · Views: 142

tedstruk

Jan 7, 2012
476
Joined
Jan 7, 2012
Messages
476
I have heard of people using ferrite beads for impulse suppression.
The only way I know how to stop EMI is with a loaded antenna.
 

Arouse1973

Adam
Dec 18, 2013
5,178
Joined
Dec 18, 2013
Messages
5,178
Hi Tim

Not all ferrite beads are used in the power supply line. You find them being used on the output of digital circuits that may interface to the outside world via a cable as an example. If you just chose the highest impedance device you could find, you may have issues with insertion loss and impedance matching issues. Also your circuit may operate at 100 MHz and your EMI issue maybe above 1 GHz so you may need to choose a lower impedance at 100 MHz knowing that the impedance is going to increase at the 1 GHz issue frequency and attenuate the unwanted emission.

Thanks
Adam
 

Tim Brown

Dec 31, 2017
23
Joined
Dec 31, 2017
Messages
23
Hi Tim

Not all ferrite beads are used in the power supply line. You find them being used on the output of digital circuits that may interface to the outside world via a cable as an example. If you just chose the highest impedance device you could find, you may have issues with insertion loss and impedance matching issues. Also your circuit may operate at 100 MHz and your EMI issue maybe above 1 GHz so you may need to choose a lower impedance at 100 MHz knowing that the impedance is going to increase at the 1 GHz issue frequency and attenuate the unwanted emission.

Thanks
Adam
I see.
chip ferrite beads can peak up to 1500 ohms, whilst the axial type ones peak at about 60 ohms. That seems like a lot of difference. How can I workout how much impedance is required at a particular frequency to reduce it as needed.
 

Tim Brown

Dec 31, 2017
23
Joined
Dec 31, 2017
Messages
23
Hi Tim

Not all ferrite beads are used in the power supply line. You find them being used on the output of digital circuits that may interface to the outside world via a cable as an example. If you just chose the highest impedance device you could find, you may have issues with insertion loss and impedance matching issues. Also your circuit may operate at 100 MHz and your EMI issue maybe above 1 GHz so you may need to choose a lower impedance at 100 MHz knowing that the impedance is going to increase at the 1 GHz issue frequency and attenuate the unwanted emission.

Thanks
Adam
also, When it comes to putting a cap on either side to make a pi filter. How can the cap value be calculated when the ferrite beads don't have inductance values?
 

Arouse1973

Adam
Dec 18, 2013
5,178
Joined
Dec 18, 2013
Messages
5,178
It all depends on your application circuit and where the noise is coming from and its impedance. This will help you decide if you are best with a PI or a T filter topology. The inductance of a ferrite bead is not constant with frequency which is why they only mention impedance versus frequency. You have to be careful choosing the correct capacitor to make sure you don't make the situation worse by introducing resonance by using a high Q capacitor.

Are you actually having an EMI issue or are you just playing safe? It is sometime better to allow for some filter components on the circuit board when you design it and then have a look with a near field probe to see if you have any issues. Then you know what you are trying to fix which is much better than hoping it will work.

You could have a look for a calculator online and simulate your circuit, this gives you a better idea of how the ferrite is going to work in your circuit. Trying to calculate it may prove difficult because of the variation of ferrite characteristics versus frequency.

Thanks
Adam
 

Arouse1973

Adam
Dec 18, 2013
5,178
Joined
Dec 18, 2013
Messages
5,178
Here is something else that you might find helpful.

Ferrite.PNG
 

Tim Brown

Dec 31, 2017
23
Joined
Dec 31, 2017
Messages
23
ok., thanks for that. parts are ordered. I'll see how it turns out.
 

Arouse1973

Adam
Dec 18, 2013
5,178
Joined
Dec 18, 2013
Messages
5,178
Cool. Please let us know how you get on.
Thanks
Adam
 
Top