At 9:16 you calculate the resonant frequency of the ferrite bead with input and output capacitor. You say that the capacitors are in series. But it seems to me that looking from the load, C2 should be in parallel with a series connection of the ferrite bead and the capacitor, right?
Can you make a "crash course video" on how to use the .measure command as you showed @10:50 ? I was not a ware of that functionality in LTSpice, not that I use it too much either, but would like to know these kind of features. I really like how you explain things, that is why I ask.
@ I am talking about the measurement setup. The LtSpice measurement requires setting up criteria to search for the target as oscilloscope requires triggering in order to capture waveforms. If you are familiar with the at, when, to and all other statements used as measurement utility in Ltspice, they are there to emulate oscilloscope triggering.
This is another great instructional video on practical power electronics topics, thanks professor. I was talking about this topic with younger engineers last week, we had issues with voltage ringing caused by resonances from ferrite beads and low value ceramic capacitors affecting power integrity on powe rails. Sharing this video with my team!
I have often found that trying to use datasheet parameters for ferrites doesn't get the simulation as close to the observed performance as I would like. I presume this is mostly down to board parasitics. I generally make provision for a ferrite and then try a few different options - it's not clever, or intellectually edifying, but it gets me to a result quite quickly. Making ferrite provision also gives you a nice way to isolate different parts of the circuit during development - which is reason enough, regardless of the potential filtering benefits (one can always fit a 0R link if the ferrite doesn't help).
@@AllTheFasteners I have seen a lot of 000R resistors placed on FB places on HP printers and other products. They might be there because different requirements in different areas of the globe 🌎
Great video, Professor! I have a question that I hope you can address in your next video about ferrite beads. When designing the layout for a MOSFET gate drive, there's a rule to make the loop as small as possible to minimize inductance. However, there's also a ferrite bead placed on the gate drive loop, which acts as an inductor at frequencies like 10 MHz. How do these two considerations work together in the design?
That's a good point. Gan Systems used to suggest adding a ferrite bead to the drive, but I don't think they do anymore. This was deemed useful for suppressing very high-frequency oscillations. In any event, they had a resistor in series that damped the loop.
Could you do a follow on video to this where you talk about the saturaton effects of ferrite beads. Most manufacturers do not provide this information about their beads, with the exception of Laird (used to be Steward). Since ferrite beads are quite useful at supressing the high frequency noise caused by harmonics of the switching regulators, a little more info on how and when to use them would be nice. Because ferrite is such a non-linear material, a lot of engineers (who should probably know better) tend to over simplify it, avoid accurate analysis, and then sprinkle them like pixie dust to solve EMI/EMC problems. I have been guilty of that, but would like more understanding so i don't repeat it.
@@sambenyaakov Yes, that is a very good video! I guess the takeaway from these two videos is that one needs to understand the ferrite fundamentals, and then test and iterate to make sure that it is really solving the issues and not causing more. And then don't let purchasing make a part substitution after the fact unless it is tested.
I thought you were going to go into saturation, this is 0805 part which will be almost fully saturated at 1A. So the problem will not exist in the real world, of course the bead will not be an improvement in that case too.
I did not find the info on the DC dependence for this bead. However it is specified to 1.6ADC while the intended use was below 200mA. See also ua-cam.com/video/F-wSu162tCo/v-deo.html
You can't do that. The EMI lab (should) take photo's of the customers setup as evidence of how the product passed. If any competitor was to query the results the photos would show the common mode chokes in place.
hmm so basically unless i need to filter 10 20 50MHz and up i should stick to normal inductors This material remanded me of a simulation where i wanted to low pass filter a 3v3 PWM and when i tried a ferrite i ended up with 10-15V spikes somehow It was great that i simulated that mess and not fries MCUs .
@@jaro6985 I understood that it's on the input! I understand that beads (As all components) should be used with care, but every board I have encountered has MANY of them.
@@sambenyaakovThen can this situation be mitigated by shifting a resonance frequency to a sub-hertz range with a bigger inductance and a bigger capacitance (of course with keeping X high for f
Can you make a specific video highlighting Frenetic software please?
Will see
Thanks for this video. But I have one question: How can I show delta diagram ( @11:01 ) in LTSpice
You use the step and meas commands.
At 9:16 you calculate the resonant frequency of the ferrite bead with input and output capacitor. You say that the capacitors are in series. But it seems to me that looking from the load, C2 should be in parallel with a series connection of the ferrite bead and the capacitor, right?
It is assumed that the injection is of current not voltage
Can you make a "crash course video" on how to use the .measure command as you showed @10:50 ? I was not a ware of that functionality in LTSpice, not that I use it too much either, but would like to know these kind of features. I really like how you explain things, that is why I ask.
The concept is similar to oscilloscope measurement. You can familiarize yourself by reading the help document.
@@getusel An oscilloscope does not step parameters and plots them, no.
@ I am talking about the measurement setup. The LtSpice measurement requires setting up criteria to search for the target as oscilloscope requires triggering in order to capture waveforms. If you are familiar with the at, when, to and all other statements used as measurement utility in Ltspice, they are there to emulate oscilloscope triggering.
This is another great instructional video on practical power electronics topics, thanks professor. I was talking about this topic with younger engineers last week, we had issues with voltage ringing caused by resonances from ferrite beads and low value ceramic capacitors affecting power integrity on powe rails. Sharing this video with my team!
Thanks. Ferrite beads are not a magic bullet!
Thank you professor, I was just about to use these parts in my board design 😊 This is incredible knowledge, thank you for sharing.
👍😊
Happy new year, Professor! Wish you all the best and good health 🎉🎉🎉
Thank you. Likewise.
An engineer once told us that you just sprinkle ferrite beads in there like pixie dust.
And?
@@sambenyaakov And then Peter Pan makes a visit, blesses your design and all is well. (He's critical of blind use of beads.)
@ 😊👍
I have often found that trying to use datasheet parameters for ferrites doesn't get the simulation as close to the observed performance as I would like. I presume this is mostly down to board parasitics. I generally make provision for a ferrite and then try a few different options - it's not clever, or intellectually edifying, but it gets me to a result quite quickly. Making ferrite provision also gives you a nice way to isolate different parts of the circuit during development - which is reason enough, regardless of the potential filtering benefits (one can always fit a 0R link if the ferrite doesn't help).
@@AllTheFasteners
I have seen a lot of 000R resistors placed on FB places on HP printers and other products.
They might be there because different requirements in different areas of the globe 🌎
Thank for input.
So why at stm32 schematic was used FB in past
In the past, what now? In any event, read the conclusion section.
Thanks
Thank you
Important topic - potential resonances from ferrites' inductive component is often ignored.
Yes, that's the key point to consider.
There are ferrite beads with peak impedance at 1MHz and higher saturation current.
I am interested what will happened with them
Can you point to an example of the types you meam?
Very practical! Thanks!
👍🙏
Great video, Professor! I have a question that I hope you can address in your next video about ferrite beads. When designing the layout for a MOSFET gate drive, there's a rule to make the loop as small as possible to minimize inductance. However, there's also a ferrite bead placed on the gate drive loop, which acts as an inductor at frequencies like 10 MHz. How do these two considerations work together in the design?
That's a good point. Gan Systems used to suggest adding a ferrite bead to the drive, but I don't think they do anymore. This was deemed useful for suppressing very high-frequency oscillations. In any event, they had a resistor in series that damped the loop.
Adding a ferrite bead to the drive path is asking for trouble.
Just put a resistor there.
Could you do a follow on video to this where you talk about the saturaton effects of ferrite beads. Most manufacturers do not provide this information about their beads, with the exception of Laird (used to be Steward). Since ferrite beads are quite useful at supressing the high frequency noise caused by harmonics of the switching regulators, a little more info on how and when to use them would be nice. Because ferrite is such a non-linear material, a lot of engineers (who should probably know better) tend to over simplify it, avoid accurate analysis, and then sprinkle them like pixie dust to solve EMI/EMC problems. I have been guilty of that, but would like more understanding so i don't repeat it.
Please see ua-cam.com/video/F-wSu162tCo/v-deo.html since that video was posted I had second thoughts about using FB in power electronics.
@@sambenyaakov Yes, that is a very good video! I guess the takeaway from these two videos is that one needs to understand the ferrite fundamentals, and then test and iterate to make sure that it is really solving the issues and not causing more. And then don't let purchasing make a part substitution after the fact unless it is tested.
@ Indeed
Please. I am interested in this, but the echo in the audio narration is too strong for me to listen.
Sorry to learn this. Try to lower the volume.
I thought you were going to go into saturation, this is 0805 part which will be almost fully saturated at 1A. So the problem will not exist in the real world, of course the bead will not be an improvement in that case too.
I did not find the info on the DC dependence for this bead. However it is specified to 1.6ADC while the intended use was below 200mA. See also ua-cam.com/video/F-wSu162tCo/v-deo.html
Degrading capacitor(s) would make a really bad joke by shifting system close to resonance.
Indeed
Use snap on ferrites to pass conducted emissions, then throw them in a drawer and forget about the issue completely.
Indeed
You can't do that. The EMI lab (should) take photo's of the customers setup as evidence of how the product passed. If any competitor was to query the results the photos would show the common mode chokes in place.
I certainly do not support this practice., but I know it's being done.
hmm so basically unless i need to filter 10 20 50MHz and up i should stick to normal inductors
This material remanded me of a simulation where i wanted to low pass filter a 3v3 PWM and when i tried a ferrite i ended up with 10-15V spikes somehow
It was great that i simulated that mess and not fries MCUs .
Indeed. Thanks for sharing you experience.
Very cool
👍🙏
Why is the bead inserted between the capacitors and not on either side of them?
You want low impedance = capacitor on your output, not a high impedance bead.
@@jaro6985
I understood that it's on the input!
I understand that beads (As all components) should be used with care, but every board I have encountered has MANY of them.
The capacitors are part of the units. And, putting an inductor before the decoupling cap is bad.
@@sambenyaakovThen can this situation be mitigated by shifting a resonance frequency to a sub-hertz range with a bigger inductance and a bigger capacitance (of course with keeping X high for f
@@sambenyaakov1. Is it a "real world" example?
2. How would you solve this problem with the F.B.?
Seems like for hobbyist use ferrite beads are rarely if ever necessary.
The FB ARE useful for filtering 30Mhz and up
👍🙏😍
👍🙏😊
It seems like youtube is missing something from long time....(Ans: Prof Sam Ben Yakov's Video😅)
Working on it!