[SOLVED] How does the core material of a current sensor affect its operating frequency

[Shuo__]

Hello friends
I am designing a current sensor to detect common mode interference currents in the line, the detected currents are between about 1M and 10M. The design uses Toroidal Cores and has a turns ratio of 1 to 10.
I have tried several materials with different permeabilities and different volumes, but the tests show no significant difference in their bandwidths.
This is not consistent with the theory I know, because the low pass frequency of CTs is mainly determined by the secondary inductance and burden resistor, and the high pass frequency is mainly determined by the parasitic capacitance of the secondary coil and burden resistor.

Can anyone give me some advice on how to choose the right material for 1 to 10 megahertz?
Thank you very much!

Translated with www.DeepL.com/Translator (free version)

 

[FvM]

because the low pass frequency of CTs is mainly determined by the secondary inductance and burden resistor, and the high pass frequency is mainly determined by the parasitic capacitance of the secondary coil and burden resistor.

Suggest to exchange low- and high pass in this statement.

High frequency limit (low pass corner) is also affected by high frequency core losses, I presume.
Please look at the complex permeability curve of a low frequency ferrite (N87) compared to high frequency (K1).

 

[cupoftea]

Above a few MHz and you need NiZn ferrite, rather than MnZn ferrite.....but above 10MHz and any ferrite is swamped in core losses........or just behaves as if it were an air core. This is why common mode chokes in offline SMPS actually use MnZn ferrite and loads of turns....because theres no point in even trying to have a high inductance for the stuff above 10MHz......so you let it be effectively an air cored inductor for anything above 10MHz...........and its high inductance at 150kHz to 1MHz takes care of the "envelope frequency" of higher frequency conducted emissions...ie the frequencies are too high for the ferrite, but the kind of "envelope of pulses" containing those high frequencys can get filtered, as the actual envelope is lower frequency.

Its kind of bizarre, but we just dont have high permeability cores for the >10MHz emissions.

Ditto your own case.

 

[Easy peasy]

https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&cad=rja&uact=8&ved=2ahUKEwiYkPX43Ib8AhW94TgGHbC9A1kQFnoECBgQAQ&url=https%3A%2F%2Fwww.fair-rite.com%2F61-material-data-sheet%2F&usg=AOvVaw1xF9m8K8NK7PS5FDqRMTQG

 

[cupoftea]

..thats very interesting, i had no idea there was NiZn that could go up to >25MHz. Looks like this is for Shuo__.
Strange they didnt specify/bother with it for common mode chokes.
Also strange, our EMC test house guy told us "ferrite components are no good at >30MHz".
But Shuo__ will be able to use it well.

 

[Shuo__]

Suggest to exchange low- and high pass in this statement.

High frequency limit (low pass corner) is also affected by high frequency core losses, I presume.
Please look at the complex permeability curve of a low frequency ferrite (N87) compared to high frequency (K1).

View attachment 180286 View attachment 180287

Thanks for the suggest, you are right, the two frequencies I described should be exchanged.

Suggest to exchange low- and high pass in this statement.

High frequency limit (low pass corner) is also affected by high frequency core losses, I presume.
Please look at the complex permeability curve of a low frequency ferrite (N87) compared to high frequency (K1).

View attachment 180286 View attachment 180287

Hi, thanks for your reply. You are right, the high frequency limit and low frequency limit in the description should be switched.
I have never paid attention to the complex permeability before, and that might be the problem.

"The region in which μ' decreases sharply and where the μ" maximum occurs is termed the cut-off frequency fcutoff. This is inversely proportional to the initial permeability of the material (Snoek’s law)" (Source:TDK General – Definitions)

So there is a trade-off between the frequency of the core material and its permeability?
The K1 material like you mentioned has a very high frequency, but its permeability is only μi=80. I currently use materials like T38 (from TDK) and 3E6 (from Ferroxcube), which have a permeability of μi=10000, but their complex permeability starts to drop at 0,1 megahertz.
So is that the reason why their high frequency limits behave similarly, even though they are different materials?

--- Updated Dec 20, 2022 ---

Above a few MHz and you need NiZn ferrite, rather than MnZn ferrite.....but above 10MHz and any ferrite is swamped in core losses........or just behaves as if it were an air core. This is why common mode chokes in offline SMPS actually use MnZn ferrite and loads of turns....because theres no point in even trying to have a high inductance for the stuff above 10MHz......so you let it be effectively an air cored inductor for anything above 10MHz...........and its high inductance at 150kHz to 1MHz takes care of the "envelope frequency" of higher frequency conducted emissions...ie the frequencies are too high for the ferrite, but the kind of "envelope of pulses" containing those high frequencys can get filtered, as the actual envelope is lower frequency.

Its kind of bizarre, but we just dont have high permeability cores for the >10MHz emissions.

Ditto your own case.

Hi, thanks for your reply.
I have reviewed several papers and they all recommend the use of high permeability materials for current sensors because this ensures a large enough magnetization inductance to obtain a low low frequency limit and to maintain a small PE (phase error).

"The higher the permeability, the more the cores are used in high-frequency transformers and adjustable inductors. Permeability helps to keep the magnetizing current at required levels while using in transformer applications. Also, high magnetic permeability in combination with low electrical conductivity helps to prevent eddy currents in the ferrite cores." (Source: cosmoferrites)

But I noticed that the complex permeability of the high permeability materials started to drop before 1 megahertz. And NiZn cores have very high frequencies, but they all seem to have very poor permeability.
So there is a trade-off between the frequency of the core material and its permeability?

If I don't care about the low frequency limit, because I'm interested in CM interference current frequencies above 1 megahertz. In that case I can choose to use a material with low permeability and high frequency right?

--- Updated Dec 20, 2022 ---

Thank you for the link, I am now looking at NiZn material

 

[mtwieg]

I've done a lot of hand-made current transformer for frequencies between 1-300MHz. The overall difficulty is mostly a function of the required bandwidth (in decades/octaves). One decade of bandwidth (like 1MHz-10MHz) is pretty easy, doesn't require much optimization. Many core materials are workable, including NiZn, MnZn, powder cores, even air core at higher frequencies and sizes.

Generally the easiest way to get wider bandwidth is to reduce the burden resistance. So long as you don't need very high sensitivity, this is easy.

If you want a few decades of bandwidth or more, things get much more difficult. Pearson builds some CTs with six or more decades of bandwidth.

Another thing worth considering is core saturation. This can happen if significant currents below your frequency of interest are passing through the sensor (especially DC). Also consider a faraday shield if you cable has large common mode voltages on it. Pearson has some excellent app notes on this and other design considerations, I recommend browsing all their literature.