How can magnetic fields be shielded from external sources?

[Dusan Stan]

TL;DR Summary

It seems I don't understand how magnetic shielding supposed to work. I tried shielding a wire, using some ferrites, but it doesn't work.

It seems I don't understand how magnetic shielding supposed to work. I tried shielding a wire, using some ferrites, but it doesn't work. I assumed the magnetic field will concentrate in the magnetic material, bypassing the meter magnetic loop, so less will be measured by the meter. I thought the ferrite is not proper for the 50Hz, so I tried a steel pipe I had laying around, but to no effect. What I'm doing wrong?

 

[zoki85]

Ordinary steel, unless it is very thick, is rather ineffective shield against low frequency magnetic fields. Assuming you don't have superconductive materials or sheets of mu-metal, you should try wrapping around with several layers of mutually isolated permalloy strips. The shielded object should be completely situated in a space enclosed by the shield

 

[Dale]

Summary:: It seems I don't understand how magnetic shielding supposed to work. I tried shielding a wire, using some ferrites, but it doesn't work.

It seems I don't understand how magnetic shielding supposed to work. I tried shielding a wire, using some ferrites, but it doesn't work. I assumed the magnetic field will concentrate in the magnetic material, bypassing the meter magnetic loop, so less will be measured by the meter. I thought the ferrite is not proper for the 50Hz, so I tried a steel pipe I had laying around, but to no effect. What I'm doing wrong?View attachment 258632View attachment 258633

You cannot magnetically shield a wire this way. By Amperes law if you have a nonzero net current going through a loop then you will have a nonzero magnetic field along that loop. No shielding can prevent that.

If you want to limit the magnetic field of a wire then you need to make sure that there is no net current through any loop outside the wire. You can do this with coaxial cable or twisted pairs.

Coaxial cable will completely eliminate the external magnetic field. Twisted pair will not eliminate it entirely, but will reduce it a lot. Plus it is less expensive than coaxial cable.

If you use twisted pair and need further suppression then your ferrites may help. But the first and most essential thing to do is to get rid of the net current.

 

[DaveE]

This isn't really magnetic shielding in the sense of reducing the magnetic field. Because you are measuring the current through one wire in your circuit, what you are doing is adding a filter component which alters the circuit to reduce the loop current. The ferrite beads or iron pipe are acting as an inductor which you have added in series with the rest of the circuit.

An inductor does have a magnetic field which is primarily contained withing the iron material. This magnetic field does store energy and create a back emf to oppose the current flowing in your wire, so in theory it could work. However effective filter design depends on all of the rest of the stuff in the circuit too. At low frequencies, like 50Hz, you will find this approach isn't effective since you will need a huge inductor, much, much larger than what you have tried here. Probably one too heavy to lift. However, I am guessing about the other components in your circuit; usually 50Hz is from a power source and that implies a very low impedance circuit.

True magnetic shielding usually is aimed at reducing the magnetic fields that propagate away from the circuit (like radio waves). These solutions almost always involve containing the circuit inside an enclosure that traps the field inside. Sometimes you will see very high permeability "walls" that aren't enclosures to divert the field away from sensitive parts, but these don't really contain the field they just move it away locally.

edit:
The twisted pair of wires that include equal currents is an exception to the containment I described. In this case the magnetic field produced reverses polarity with each twist, then from a distance away from the wires the fields cancel each other. This is a cheap and surprisingly effective technique. However, if you are only measuring the current in one of those wires you will see a negligible effect.

 

[zoki85]

The wire can be shielded in this way (more or less successfully) but the effect of shielding can't be measured by a clamp meter like shown in the photos. As matter of fact the clamp meter can react to local AC field and show readings even in the case of zero current through the wire. That's the main reason clamp-meters are recommended only for strong current measurements.

 

[DaveE]

the clamp meter can react to local AC field and show readings even in the case of zero current through the wire

Yes, many of these meters are unreliable. But the good ones actually work quite well. Fortunately it is really easy to do a cursory test to spot the bad ones. This isn't the best instrument to get as a cheap ebay brand. This is a case where you get what you pay for.

As you said, you can look for zero reading when there is no wire inside, hold the closed meter as close as possible to the measurement location and check for zero (or an error small enough to be ignored). You can also try reversing the polarity and look for the same answer. Many clamp-on meters say the wires should be near the center of the loop, but you can easily test yours to see if this matters. Whenever possible compare to other measurements, like a different brand meter or know currents. Overall, be skeptical and do a constancy check; the good meters don't change their answer when you move stuff around. The DC clamp-on meters are especially prone to errors, you should almost always check the reverse polarity on these.

Keep the contact surfaces between the core halves clean and make sure the clamp is 100% closed.

 

[alan123hk]

It seems I don't understand how magnetic shielding supposed to work. I tried shielding a wire, using some ferrites, but it doesn't work.

Quote from the link below "First, one important point must be clear: Magnetic shielding does not block a magnetic field. No material can stop the lines of flux from traveling from a magnet's north pole to it's south pole. The field can, however, be redirected"
https://www.kjmagnetics.com/blog.asp?p=shielding-materials

Because the magnetic field cannot be blocked by covering the current source with ferrite, you can try to completely shield the coil current sensor of the clamp meter with a high permeability material.

Since the magnetic field can be redirected (but not eliminated) by this measure, the magnetic field will no longer pass through the coil current sensor and no induced voltage will be generated.

Of course, this is just an interesting experiment and has no practical use , but anyway, I think it might work.

 

[Baluncore]

The clamp contains a magnetic core that is mechanically closed about the wire. The flux through that core is due to current in the wire, and is measured in the meter.

When you put a ferrite core about the wire you give the flux an alternative path that does not pass through the clamp meter circuit. That is why the reading falls slightly. The beads threaded on the wire are providing a magnetic short circuit for the flux, so it does not go through the sensor in the meter.

 

[alan123hk]

In my opinion, wrapping the copper wire with ferrite core in principle cannot prevent the magnetic field from entering the coil current sensor of the clamp meter, even if the ferrite core can fill up all the space around the clamp of the meter.

The two reasons why it is basically invalid are : -

1. This violates Ampere's law (unless an equivalent reverse current is generated and flowing through the clamp to cancel the magnetic field).

2. Try to analyze it by applying the concept of magnetic circuit.

Modeling the wire with a single turn coil with a constant current I, so MMF =NI =I

Partitioning the loop area (At) of the single turn coil into 3 different portions , As, Ac and Af, where

As = cross sectional area of the air space
Ac = cross-sectional area related to the clamp of the meter
Af = cross-sectional area of the ferrite core
so that At=As+Ac+Af

According to the magnetic circuit theory, we can reasonably approximate the total reluctance (Rt) as 3 reluctances in parallel, namely Rs, Rc and Rf, where

Rs = effective reluctance of the air space
Rc = effective reluctance of the clamp of the meter
Rf = effective reluctance of the ferrite core
1/Rt = 1/Rs+1/Rc+1/Rf = total permeance of the magnetic circuit

,which should inversely proportional to the respective As, Ac and Af.

Therefore, after the ferrite core is inserted, the inductance of the single-turn coil will actually increase, and the magnetic field in space will actually change as well. However, regardless of the ratio between Rs, Rc and Rf, the magnetic flux (Φ) go through the coil current sensor will be still Φ= NI/Rc, and Rc should remain approximately the same. In other words, we can still expect clamp meter providing effective reading.

 

[hutchphd]

which should inversely proportional to the respective As, Ac and Af.

The permeance is also proportional to the length of the magnetic path so the effect of the beads will be relatively greater. But the conclusion(s) still hold . In fact with the wire in the center this problem can be solved exactly which I will do if I get a chance...

 

[alan123hk]

The permeance is also proportional to the length of the magnetic path so the effect of the beads will be relatively greater

Permeance is equal to uA /L, where
u = permeability
A = cross-sectional area
L = length

Depending on the geometric irregularities of the magnetic circuit path, determining the effective values of A and L can be very difficult, especially for the air space portion, because its magnetic field diverges into the huge surrounding space, the effective area may be very large, so the permeance may be much greater than expected

Anyway, I just hope to make the situation conceptually clearer, more detailed, and easier to understand by applying the magnetic circuit theory.

 

[Baluncore]

So, what is really going on in the OP's circuit ?

1. The clamp meter.
The meter shown in post #1 measures AC current only, so it is almost certainly a current transformer with a mechanically closed magnetic path and a low resistance burden on the multi-turn secondary. A clamp meter appears as the clamp inductance, inserted in series into the wire, but short circuited by the internal transformed secondary burden.

Doing the numbers; If the clamp had a 40 turn secondary, with 0.1 ohm burden, then for 400 amp in the primary it would have 10 amp in the secondary, dropping 10 A * 0R1 = 1 volt across the burden. That is what the instrument measures, scales, and displays as “400”A.

The 0.1 ohm load would be transformed to appear as a series drop of 1V / 40 turns = 25 mV in the single primary "turn".
The equivalent parallel resistance is 25 mV / 400 amp = 0.0000625 ohms.
Note how the square of the turns ratio transforms the secondary burden, to 0.1 ohm / 40² = 62μ5 ohm in the primary.

The meter appears as the clamp inductance in parallel with a 62μ5 ohm series resistor. The maximum voltage drop in the wire being “measured” is 25 mV. All without cutting the wire.
That 25 mV drop reduces the available load voltage and so reduces the current drawn in the circuit while the clamp is in place. For a 230 V circuit that will be a current reduction of about 25 mV / 230 V = 0.01%.

2. The ferrite beads.
Threading the wire through non-conductive ferrite beads increases the self inductance of the wire, and so reduces the AC current flowing in the circuit. But the inductance of the non-conductive bead is not short circuited by any burden as is the clamp meter inductance.

Each ferrite bead adds another series inductor (probably less than 1μH) to the circuit. The beads appear to be small and made from a ferrite designed for low loss at MHz frequencies. Since the bead reactance is X_L = 2·π·f·L; those beads will not be significant at 50 Hz, but their inductance will oppose the flow of RF along that wire.

Doing the numbers, 10 beads at 0μ5 H each makes 5 μH.
At 50 Hz we have X_L = 1.57 milliohm in series with the wire.
At 1 MHz we have X_L = 31.4 ohm in series with the wire.

3. The steel pipe.
Threading a short steel pipe onto the wire will result in bad 50 Hz effects since it is a conductive solid steel core magnetised by the current in the single wire. The metal sleeve will get hot due to eddy currents flowing in the surface. That is why you must NEVER pass a single AC wire through a conductive aperture. You must pass all the return circuits through the same aperture to cancel the induced magnetic field, and so prevent inductive heating and energy loss. That is true for single and three phase AC circuits.

 

[hutchphd]

In my opinion, wrapping the copper wire with ferrite core in principle cannot prevent the magnetic field from entering the coil current sensor of the clamp meter, even if the ferrite core can fill up all the space around the clamp of the meter.

After more thought I realize this is absolutely true. The magnetization field H is caused by real currents alone (not "magnetization currents"). To the degree that the ferrite bead has no real induced current and does not change the source current it will have no effect on H at the clamp and hence on the reading. Only at high frequencies will there be these effects become considerable. Sorry I was a bit muddled earlier.!

 

[alan123hk]

To the degree that the ferrite bead has no real induced current and does not change the source current it will have no effect on H at the clamp and hence on the reading. Only at high frequencies will there be these effects become considerable

I agree with that.

The ferrite core acts as a lossy inductor connected in series with the wire, so its impedance increases together with the frequency, which prevents high-frequency current from passing through the wire. In addition, high-frequency current can be dissipated as hysteresis loss. The purpose is to suppress high-frequency noise.

If the current is weakened or blocked by the ferrite, the reading on the clamp meter will decrease or become zero. However, if I understand correctly, the OP is attempting to shield the wire with ferrite cores so that the magnetic field generated by the AC 50Hz current flowing in the wire cannot reach the clamp meter, and the current can continue to flow.

I am just trying to describe why it doesn't work.

 

[hutchphd]

It doesn't work because 50 Hz generates no appreciable current in the ferrite and so, even though it gets a large internal field , does not affect the field outside the bead according to maxwell. It doesn't work because it doesn't work. (?)

 

[alan123hk]

It doesn't work because 50 Hz generates no appreciable current in the ferrite and so, even though it gets a large internal field , does not affect the field outside the bead according to maxwell. It doesn't work because it doesn't work. (?)

Inserting a ferrite into a single-turn wire is just like inserting a rod-shaped ferrite into a multi-turn air coil, so that the inductance increases. When the impedance or the working frequency is large enough to reduce the AC current to a certain extent, the clamp meter will show the reduction in current (assuming the clamp meter can respond to such high frequencies ).

A strong magnetic field is indeed generated inside the ferrite. However, even if we consider the strong magnetic field in the ferrite can be regarded as an equivalent circular current, but this equivalent circular current is confined inside the ferrite, regardless whatever its normal vector, it can not form a reverse current to cancel out the current in the wire, so the field outside is not affected.

Someone may ask a question, as the magnetic field can be redirected, can the field be redirected or shielded by wrapping up the entire wire loop with a very high magnetic permeability material, so that all the field is concentrated only inside the high magnetic permeability material ?

The answer is still negative in theory, and the reasons are similar as before : -

1. This violates Ampere's law

2. Again using the concept of magnetic circuit theory and modeling the wire as single-turn coil

The high magnetic permeability material path is only one branch of the parallel circuit , it can not eliminate the magnetic flux flowing in other parallel branches (including the branch of the clamp meter) because all of them are directly connected to the the single-turn coil magnetomotive force (MMF) source.

Just like electric circuit, if a resistor is connected across an ideal voltage source, we can not change its electric current by add another resistor in parallel with it.

 

[DaveE]

The high magnetic permeability material path is only one branch of the parallel circuit , it can not eliminate the magnetic flux flowing in other parallel branches (including the branch of the clamp meter) because all of them are directly connected to the the single-turn coil magnetomotive force (MMF) source.

Just like electric circuit, if a resistor is connected across an ideal voltage source, we can not change its electric current by add another resistor in parallel with it.

I think your circuit analogy is wrong. It is actually more like a current source (the magnetic flux generated in the wire). Inserting high permeability materials as a "wall" (i.e. with enough air gap to avoid a significant change in the inductance seen by the circuit) doesn't reduce the amount of magnetic flux generated, but it does redirect that flux through the added material. More like parallel resistors connected to a current source. This technique is only effective nearby the added material. You are correct that it will not effect the flux in areas away from the added material. It is a local effect.

 

[DaveE]

The ferrite core acts as a lossy inductor...

Yes! At high frequencies this is often much more significant than the increased inductance added to the circuit. This explains why you will often see powdered iron used instead of ferrites in EMC filters.

Of course, as has been said, it's all meaningless at 50Hz. At very low frequencies the only magnetic structures that matter are really big and heavy (unless you count Faraday cages as magnetic structures).

 

[alan123hk]

I think your circuit analogy is wrong. It is actually more like a current source (the magnetic flux generated in the wire)

I am not against your idea, in fact I think your argument is reasonable to some extent .

However, if we still use the framework of magnetic circuit theory to analyze, magnetomotive force NI should play a role in the equation analogous to the voltage.

https://en.wikipedia.org/wiki/Magnetomotive_force

I admit that when the ferrite is added, the space magnetic field will be changed or twisted to a certain degree, and the field will certainly be redirected some way by the ferrite. It's difficult to obtain detailed and accurate field distribution without in-depth calculations or simulations. Please note that my main point is that after the ferrite is added, all the field will not concentrate in the ferrite, and the field in the air space will not be eliminated.

In fact, by the Ampere'law, no matter how high magnetic permeance of the added ferrite, if we perform the line Integration around the different paths in the air space, the integral result is still the same NI, which proves that the magnetic field is basically retained and dispersed to the entire air space.

Inserting high permeability materials as a "wall" (i.e. with enough air gap to avoid a significant change in the inductance seen by the circuit) doesn't reduce the amount of magnetic flux generated, but it does redirect that flux through the added material

Yes, it does not reduce, but does inserting ferrite increase the total magnetic flux output from a fixed current source ?

 

[hutchphd]

For the concentric case: The magnetic field B inside the ferrites is increased. The field in vacuum is unchanged. The magnetization field H is caused only real currents and is unchanged by the presence of the ferrites. The outside ferrite is unaffected by the inside one (unless there are real induced currents) .
In this case the field theory is simpler then torturing magnetic circuit theory to fit.

 

[Baluncore]

The outside ferrite is unaffected by the inside one (unless there are real induced currents) .

There should be no induced current associated with the inner ferrite beads.
But the presence of the current clamp meter changes things.
If the current clamp is a current transformer, then there is an induced current. If the clamp is a Hall effect sensor in an air-gap, then maybe there is no significant induced current.
https://en.wikipedia.org/wiki/Current_clamp

 

[hutchphd]

The presence of the ferrite in the clamp certainly makes the clamp current higher because the B field inside the clamp is much enhanced by the magnetization M inside the sense coil (or the hall effect junction). But the bead on the wire will have no effect on any of this ...the magnetizaton field H at the sense coil will be the same and the current sensed will be unaffected.

 

[Dusan Stan]

Lesson learned. After more experimenting: - different frequencies and measurement methods; - even filling most of the space of the clamp with leftover transformer I cores, the measured current stays the same.

The idea was to find a way to prevent generation of external magnetic field without cancelling the current, but apparently, this is not possible.

Thank you all for your input.

 

[alan123hk]

Lesson learned. After more experimenting: - different frequencies and measurement methods; - even filling most of the space of the clamp with leftover transformer I cores, the measured current stays the same.

Yes, once the current-carrying wire is inside the closed clamp of the meter, there will be impossible to block or redirect the magnetic flux to prevent it from passing through the sensor inside the clamp of the meter, even if the entire space around is filled with high magnetic permeability material.

This is a reasonable conclusion drawn directly from Ampere's Law.
Certainly the use of magnetic circuit theory is like a personal preference and is not necessary. In fact, the magnetic circuit theory is also derived from Ampere's Law, which is mainly to facilitate the analysis and design of engineering practice.

 

[zoki85]

Lesson learned. After more experimenting: - different frequencies and measurement methods; - even filling most of the space of the clamp with leftover transformer I cores, the measured current stays the same.

The idea was to find a way to prevent generation of external magnetic field without cancelling the current, but apparently, this is not possible.

Thank you all for your input.

Glad you understood.
Shielding question can be separated to two cases: what is shielded from what?
First case: Can a wire be shielded from external AC magnetic field by placing it in appropriate enclosure?
Answer is yes, and for best result superconductive shield tubing should be used. Since I was initially unable to view your attached photos, and text alone appeared to ask about this case, my answer was given accordingly.
Ability to view the photos revealed the second case you were actually interested in: Can an external space be shielded from the magnetic field of current passed straight wire by placing it in appropriate tubing? Answer is no, regardless of what kind of tubing material you use.