going mad with beads..

Go To Last Post
18 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

 

I have not used ferrite beads before and I am trying to figur out where I need them on a new PCB I want to start to sell (in the EU). 

I think I understand the concept of the beads, but I am a bit confused to exactly where they need to go on my prototype and how many I need.

I have put one on all VCC input lines on all the ICs together with an capacitor.

the I have about 20 I/O to the outside world for diffrent external components like power LED, DC motor, on off switch etc. 

But its starting to be alot of beads.. Maybe I don´t need so many?..or maybe I do..

 

So..help a fella out.. any tumb rules for where exactly I should put them? what is prioritized? where can I skip? datalines? powerlines? I/O? 

 

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 1

EMC is a complicated subject - it is difficult to predict how some circuits behave because there are a lot of variables. Just sprinkling them around is an expensive solution and may cause more problems than what they solve. Usually one builds a prototype board then tests it. The result of these tests determine if you need to make changes - ie add beads, capacitors, alter pcb layout, change component specs, add common mode chokes etc.

The specification of the beads can be somewhat tricky - their effectiveness changes with the amount of current flowing through them and temperature factors in as well. 

Most of my experience has been with switched mode converters - the pcb layout and component selection is critical. Another oversight that caught me out once was signals to a character lcd - I'd used a level translator ic that was particularly fast and the outcome was that the flat cable to the lcd module was radiating. Some series resistors cured this. I had emissions up around 100MHz with that.

 

With CE there's not only emissions, but susceptance. With every test I usually learn something new.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Well..I don´t know.. they cost 0.1 dollar each and if you compare with redesigning and retesting I feel that is quite cheap.. worst case senario you put a 0ohm ressitor or delete for production.

but as you say I don´t want to make it worse either. And not put unnecessary ones.. 

 

It just feels like there should be design rules that one could look at. havent found it yet thou. 

 

Edit: I should also mention I don´t have the posibility to test EMC without paying for it. So that adds to the cost if I need to test extra. 

Last Edited: Thu. Nov 15, 2018 - 11:39 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 1

You’re caught in a catch 22 situation - you want to sprinkle the beads like magic dust to solve a problem you’ve not identified yet. You need to test to identify the problem!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Would a used spectrum analyzer  with an antenna be feasible?

Or even those much cheaper USB analyzers for $20-80. I have zero experience, but im gonna buy a USB one just for the heck of it. 

~William

Last Edited: Thu. Nov 15, 2018 - 01:04 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

In the old days we used to send boards off to a test house they had equipment like static discharge and TEMS cells and so on and could test and profile the design with highlights of where and what it was radiating and what it was susceptible to too. It's not the kind of test work you can easily do just on your own bench so it's probably worth paying a few $,000 to get it done professionally.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

So I´m the only one that thinks is a good idea to be a bit preemptive and add this stuff from the begining?

I mean it´s easier to remove a footprint or change from a bead to zero ohm ressitor than it is to add a footrint afterwards.

Also true i guess that if you don´t know where to add it you will need to add it in a heck alot of places i guess...

 

it just seems like there should be the usual suspects we you could be a bit preemtive. like at the microcontroller or whatnot..

I don´t know..Maybe I am looking at this the wrong way like you say.

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

mannen wrote:
So I´m the only one that thinks is a good idea to be a bit preemptive and add this stuff from the begining?
Probably. How can you possibly guess where the issues that need fixing are? It's a bit like filling the rust holes on your car before the rust has actually appeared.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Consider what ferrite beads do. At a minimum, they are a higher impedance at high frequencies than at low frequencies. SOME are designed to be resistive at high frequencies so that they absorb the high frequency signal energy. So, they can be used to isolate portions of circuits with significant high frequency energy.

 

To do this isolation, each ferrite needs to be part of a voltage divider. The divider is made from a ferrite bead as a series element and an impedance as a shunt element. The divider is most effective if the shunt element is a capacitor so that you have a high series impedance and low shunt impedance at high frequencies. This gives you the greatest attenuation at high frequencies while allowing the desired low frequencies, through.

 

Given that, you CAN generate some basic rules of thumb:

 

1) Ferrite beads tend to be most useful when there are circuit blocks that generate high frequency signals.

 

An example is a microcontroller power circuit. The MCU takes relatively high peak currents at every clock edge. These current peaks are narrow pulses, about the width of the logic rise time for that technology. So, for these circuits, you really want a bypass cap to supply the peak energy to the MCU plus a ferrite bead attenuating the remainder of the current spikes propagating outward from the MCU (plus, don't forget it, a cap on the outer side of the ferrite bead, to make the needed attenuator).

 

2) The most high frequency energy is present when the signal repetition rate is high.

 

Thus, while the MCU output that drives a relay may have very fast rise and fall times, if that signal changes state infrequently, then the average high frequency energy will be low. On the other hand, if you have an SPI port that is heavily used and clocked at the maximum possible rate, then its potential for generating undesirable high frequency signals is quite high.

 

3) Radiated signals come from antennas. Sort of figures. In the circuit world, antennas are wires or traces with high frequency current flowing through. If you can keep that high frequency current out of the wire or trace, it won't radiate. Wires and traces radiate the most strongly when they are an odd multiple of a quarter wavelength for the current through. Example: 2.4GHz has a WAVELENGTH of about 10cm. 1.2GHz has a wevelength of about 20cm.

 

4) EMI tests do not care about all the high frequency signals bouncing around inside your thing, as long as they don't result in radiated signals (item 3, above) or escape onto exterior wiring. Keeping EMI out of exterior wiring can be a special challenge. For lines carrying significant power, common mode chokes are often more effective than individual ferrite beads. Sometimes, connectors with built-in ferrite are very effective (and they can be a real savior when trying to retrofit an existing design). PWM motor drives can be a real problem, because you often want the PWM repetition frequency to be as high as possible so you choose a fast driver with fast switching times. Then, when you use a cable to connect the motor to the driver, you end up with lots of radiated noise. Using shielded twisted pair with a common mode choke at the driver end can all help more than individual ferrite beads. This points to another "solution": don't use faster circuits (ICs) than what you really (really) need!

 

5) One effective way of reducing EMI is to not generate the signal in the first place. Using the internal oscillator of an MCU goes a long ways. If you MUST use an external timing source (e.g. crystal), then use one of the newer chips with low power oscillator. These oscillators do not limit so hard, so the rising and falling edges are much softer, and the high frequency content is much lower. You still need to bypass and isolate the power circuit for these, but they can be much quieter for radiated signals.

 

One could go on, here, (and on, and on) about enclosures, circuit board layout, component choices, and more. But, the original question was really about ferrite beads. Here, it is important to point out that, often, ferrites are a waste if you do not have the supporting good board layout and do not use appropriate components. Good board layout includes an effective ground plane on at least a two-sided board.

 

An important corollary is that if a circuit block does not generate high frequency signals, then no isolation ferrite is needed. An example MIGHT be an LM7805 voltage regulator. Yes, you need bypass caps at the input and the output but that is for stability, not EMI. Generally, you won't need a ferrite or common-mode choke at the input because of those bypass caps (if the board layout is done well) because this is a slow analog circuit. But, there ARE exceptions even to this example.

 

So, there are design guidelines. Unfortunately, this stuff is rarely taught in engineering schools and is usually only learned in the "school of hard knocks"

 

Jim

 

 

 

 

Until Black Lives Matter, we do not have "All Lives Matter"!

 

 

Last Edited: Thu. Nov 15, 2018 - 05:31 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

ka7ehk wrote:
Unfortunately, this stuff is rarely taught in engineering schools and is usually only learned in the "school of hard knocks"
and seminars and webinars :

High Frequency Measurements Site Index

(quarter page)

Great Design, Troubleshooting, and EMC/ESD Seminars!
Learn and have fun, all at once at a beautiful, historic venue 22 miles from Las Vegas.

...

High Frequency Measurements Site Index

 

Webinars on Targeted Topics That Can Avoid/Fix Design Problems!
...

 

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

If you want to learn more about ferrite material types check out http://palomar-engineers.com/ferrite-products/ferrite-cores/ferrite-mix-selection .  Another source of information is an Analog Dialog article "Ferrite Beads Demystified" at https://www.analog.com/en/analog-dialogue/articles/ferrite-beads-demystified.html .  And one about core shapes at https://www.murata.com/products/emc/ferrite/basic/selection .

 

I recently got interested in ferrites as a way to isolate the noise of micro controllers and digital logic from the power leads to analog components in a mixed signal environment.  I used a small series inductor and shunt capacitor in my last project involving analog switches and became concerned with the power capability of the small inductors.  After I close out my current project, I plan to give ferrites a try is a test circuit with some concern with possible oscillation caused by the inductor/capacitor combination.  Much to be learned. 

 

Alan 

 

Jim,  Thanks for the guidelines.

Last Edited: Thu. Nov 15, 2018 - 10:54 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

mannen wrote:
like at the microcontroller or whatnot..
or in the MCU.

Some AVR have slew limiting for IO [megaAVR 0-series, PORTCTRL.SRL; mega4809, tFALL = 1.3ns or 11ns slew-limited (typ, 5V, 20pF load)]

 

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Slew rate limiting can really reduce the EMI generated by changing signals on port pins. Sometimes, you really need port pins as fast as possible (SPI on a bus running at maximum transfer rate, for example), but then you have to be really careful about distance, layout, and grounding.

 

Jim

 

Until Black Lives Matter, we do not have "All Lives Matter"!

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

vertamps wrote:
Would a used spectrum analyzer  with an antenna be feasible?
Yes

Expensive instruments can be rented (short duration) or leased (long duration)

If add the tracking generator option to the spectrum analyzer (it's a part of a network analyzer) then can evaluate EMC.

vertamps wrote:
Or even those much cheaper USB analyzers for $20-80.
Likely no due to inadequate bandwidth (some satellite telemetry is at 40MHz ISM); those would be good for evaluating power supply stability (error amplifier's gain margin)

If can spend an order of magnitude more, an inexpensive network analyzer :

DG8SAQ VNWA 3 and 3EC Low Cost Vector 1.3 GHz Network Analyzer VNA

One of its modes is a 100MHz spectrum analyzer; good for most AVR, acceptable for XMEGA AVR.

 


Technical Tidbit - March-April 2014

by Douglas C. Smith

Technical Tidbit - March-April 2014
Troubleshooting Radiated and Conducted Immunity Problems in the Development Lab

...

Figure 1. Test Setup for Troubleshooting Radiated and Conducted  Immunity Problems
 

Abstract: Tackling radiated and conducted immunity problems can be difficult because of the high cost of the equipment and chamber normally used to perform these tests. An inexpensive test bench setup that can effectively find radiated and conducted immunity problems is presented.

...

Re: chamber - some perform pre-formal-EMC testing at a rural basement.

 

USB, 30MHz bandwidith with the BNC kit : Analog Discovery 2 [Reference.Digilentinc]

 

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

mannen wrote:
worst case senario you put a 0ohm ressitor or delete for production.
RC can be nearly as effective as LC with the added benefit of current limiting (when one mis-wires an input or output during provisioning)

mannen wrote:
It just feels like there should be design rules that one could look at. havent found it yet thou.
AN_1619 AVR040: EMC Design Considerations | Application Notes | Microchip Technology Inc.

via ATMEGA4809 - 8-bit AVR Microcontrollers - Microcontrollers and Processors

Each series of XMEGA AVR have a schematic checklist which shows the power supply filters (VCC, AVCC; inductor, ferrite bead)

 

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

mannen wrote:
I have put one on all VCC input lines ...
What nearly happened on USB VBUS during EMC testing a USB megaAVR :

https://www.avrfreaks.net/forum/ce-testing-costdifficulty#comment-667261

in CE testing cost/difficulty | AVR Freaks

 

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Just a little comment. SOme years ago, I attended a short seminar given by Dr. Henry Ott -- then with Bell Labs.  Bell was ver interested in cost effective control of emissions because of course they had boatloads of equipment all over the US. Dr. Ott presented a lot of information in a very practical manner such that I could at least appreciate the issues and how you might go after them.

 

Amazingly, he has been active and giving seminars -- but that's about to end. On January 1st  2019 he will hang it up.

 

His website will still be active http://www.hottconsultants.com/

 

He has a recent book about EMC engineering. I have no idea what's in it because that's not an issue for retired folks like me. :)  However, I suspect that it has a great deal of useful information. Caveat emptor, but it might be worth a look.

 

hj

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Thank you very much for all the information =)

especially from Jim. that was interesting to read.