Does an AC Magnetic field induce movement in a DC Magnetic Field?

[Dinoduck94]

Summary: Can a rotating AC Magnetic field induce movement in a static DC Magnetic flux?

I'm designing a control panel, and the customer has asked us to reduce the EMC as much as possible; there are no drives, or other noise creating devices, just AC circuits.
I thought a good starting point would be to calculate the amount of induced voltage on the communications cable from the AC circuits.

My question is in regards to the DC circuits in the panel; It's a 2 bay panel, left bay is for Power and right bay is for Control - so the Control bay has a load of DC circuits within it. I know that DC circuits create a static magnetic field and as such do not induce voltage on other cables (apart from at start-up), but would the rotating flux from the AC circuits induce movement in the static DC Magnetic flux?

Additionally, if it does, is there any known mathematics or studies been done to investigate this?

Thanks in advance.

 

[mfb]

Mechanical movement?

DC cables have forces between each other. A current flowing through a magnetic field leads to a force. That still applies to AC, here the force will vary with time.
Keep in mind that DC can still be switched, by the way.

 

[Dinoduck94]

Mechanical movement?

DC cables have forces between each other. A current flowing through a magnetic field leads to a force. That still applies to AC, here the force will vary with time.
Keep in mind that DC can still be switched, by the way.

Hi mfb, I think you misunderstand my question - sorry if I didn't make it clear.

The AC circuits will have a continuous current going through it, the waveform will be sinusoidal - the magnitude of the magnetic field follows this waveform, thus creating a "rotational" movement of the Magnetic field - the rotational affect is more prevalent when it is 3 phase, but even as single phase, the magnitude still changes in relation to the waveform.
The DC circuits, will also have a continuous current and as such will generate a static magnetic field (not a moving one like the AC circuit).

In order to induce a voltage on another cable the cable needs to move within the magnetic flux - as the magnetic flux is moving on the AC circuit we don't need to move the cable, the movement of the flux simulates this action.
As the DC magnetic field is static, there won't be any induced voltage on another cable from it's magnetic field.

My question is more on the interaction of magnetic fluxes; whether a magnetic field with a varying magnitude can clash with the flux of a nearby static magnetic field and cause it to no longer be static; and thus induce voltage on another cable.

Thanks

 

[mfb]

I'm still not sure what exactly is supposed to move. Magnetic fields don't "move" in space unless you have extremely high frequencies.

Magnetic fields don't affect other magnetic fields. You can analyze each source of the magnetic field separately. A time-dependent magnetic field will induce voltages in cables in general.

 

[Dale]

whether a magnetic field with a varying magnitude can clash with the flux of a nearby static magnetic field and cause it to no longer be static

The fields are linear, so they obey the principle of superposition. The combined field is simply the sum of the two source fields.

 

[Dinoduck94]

I'm still not sure what exactly is supposed to move. Magnetic fields don't "move" in space unless you have extremely high frequencies.

Magnetic fields don't affect other magnetic fields. You can analyze each source of the magnetic field separately. A time-dependent magnetic field will induce voltages in cables in general.

Say there is 100A on this cable, then following the sinusoidal waveform of the current it will go from 0A>100A>0A>-100A>0 over one cycle as shown below:

(Credit for the picture to: https://www.researchgate.net/figure/Alternating-Current-Waveform_fig1_317165946)

This creates a pulsing magnetic field; for a 50Hz circuit then that's 100 pulses a second (1 for each peak).
On the above chart you can remove the word 'Current' on the 'y' axis and replace it with 'Tesla', showing the magnetic flux density increasing as the current increases.

It's this constant increase and decrease of the Magnetic flux density, that simulates movement.

To induce a voltage you need to move the cable through the magnetic field, but as the magnetic field is pulsing (going from a high Tesla down to a low Tesla and back again) this simulates the movement and thus induces the voltage on a stationary cable.Thank you for answering my question though; so magnetic fields do not interact with each other? So this pulsing magnetic field will not interact with the flux in another magnetic field?

 

[Dinoduck94]

The fields are linear, so they obey the principle of superposition. The combined field is simply the sum of the two source fields.

The sum of the fields would be sqrt(a^2+b^2). But my question refers to the interaction of a pulsing magnetic field with a nearby static magnetic field.

 

[Dale]

But my question refers to the interaction of a pulsing magnetic field with a nearby static magnetic field.

It is linear. I already said that.

 

[berkeman]

Summary: Can a rotating AC Magnetic field induce movement in a static DC Magnetic flux?

I'm designing a control panel, and the customer has asked us to reduce the EMC as much as possible; there are no drives, or other noise creating devices, just AC circuits.

"Reduce the EMC" is a bit ambiguous -- did this customer specify which types of EM Interference (EMI) they want to minimize? Is it truly just AC Mains frequency interaction between high-current conductors? Or are they worried about switching transient noise or some other source of digital noise?

What are they worried about "noise" from the panel interfering with?

 

[Dinoduck94]

"Reduce the EMC" is a bit ambiguous -- did this customer specify which types of EM Interference (EMI) they want to minimize? Is it truly just AC Mains frequency interaction between high-current conductors? Or are they worried about switching transient noise or some other source of digital noise?

What are they worried about "noise" from the panel interfering with?

Hi Berkeman,

The customer seems to be overly concerned regarding EMC. The panel is entirely mild steel, so any external frequencies aren't going to be a problem; the only source of noise would be AC Mains frequency interaction with nearby communication cables, within the panel.

I know steps that I can take to reduce the EMI caused by the AC cables such as segregation; distance; bundling the 3 phase cables in a delta configuration etc. (The customer's spec says we can't shield the comms cable due to transfer speeds)
I wanted to provide an excel sheet to the customer to show: this was the calculated induced voltage on the comms cables before any action was taken, and this is the predicted reduction based on these steps - so that they have confidence in us in regards to this issue.
I'm making the excel sheet as a supporting document for future projects too, hence why I'm going into it with such detail.

So I'm able to calculate the Inducted Voltage on a comms cable, from a nearby AC circuit (3 phase or single, aligned or bundled); I was just unclear on the interaction between AC magnetic flux and DC magnetic flux: whether they would just gloss over each other, not interact and therefore the induced voltage would be calculated from the AC magnetic flux; or whether the fluxes do interact, which would cause the DC flux to pulse with the AC flux and therefore the inducted voltage would be the sum of the 2 fields as worst case scenario.

Thanks

 

[berkeman]

The customer seems to be overly concerned regarding EMC. The panel is entirely mild steel, so any external frequencies aren't going to be a problem; the only source of noise would be AC Mains frequency interaction with nearby communication cables, within the panel.

(The customer's spec says we can't shield the comms cable due to transfer speeds)

It still kind of makes no sense (not your fault). I work a lot with communication cables and systems, and I've yet to see any interference at AC Mains frequencies in communication systems. The one exception is RS-485 type systems where no transformer coupling of the data signals is used. Then the common-mode ground voltage between two different RS-485 transceivers can push the differential comm signal outside of the common-mode voltage range of simple RS-485 transceivers.

The other potential problem that I already mentioned is if there is high-frequency switching noise injected back into the AC Mains by whatever is getting powered by the cables. Then you need to look into the CMRR of the comm transceivers, and also look into whether the noise sources are passing the Conducted Radiation limits set by the FCC (and other agencies in other countries).

What is their data network? 10BASE-T? 100BASE-T? Do you know what PHY circuits are being used on their comm nodes? And there are definitely shielded comm cable options at higher frequencies. The shielded cable needs to be designed correctly (so Zo and other parameters are within spec) -- you can't just wrap a shield around an unshielded comm cable.

 

[Dinoduck94]

I work a lot with communication cables and systems, and I've yet to see any interference at AC Mains frequencies in communication systems.

Hi Berkeman, thanks for your response,

I can't say I've experienced it either, but it's "Electrical Engineer 101" when you design a Low-Voltage Control panel; AC frequencies must be segregated from Communication systems so as to avoid inducing voltage on to them - that induced voltage would then become electrical noise on the system leading to errors, malfunctions and worst case is equipment failure.

The other potential problem that I already mentioned is if there is high-frequency switching noise injected back into the AC Mains by whatever is getting powered by the cables. Then you need to look into the CMRR of the comm transceivers, and also look into whether the noise sources are passing the Conducted Radiation limits set by the FCC (and other agencies in other countries).

The 'Conducted Radiation limits' aren't that much of a concern with this application, but it is something I have looked at in regards to this.

Looking into the CMRR of the transceivers and the PHY circuits of the nodes goes beyond typical practice for Control Panel Design, and it is something I was never taught (on the job, or in education); but thank you for mentioning them, I have some researching to do - there are clearly some gaps in my knowledge.

 

[berkeman]

Are the comm circuits RS-485 maybe? That would indeed lead to these concerns becuase of ground shifts and so-so CMRR or RS-485...

 

[Tom.G]

It has been asked a few times and you really do need to tell us what the communication technology is. For instance if it is a fiber optic link there will not be any interference, whereas if it is single-ended 3.3V CMOS logic there will be grief.

Anyhow, the usual baseline approach is to twist the signal wire with its respective return wire in the data cable, often from 6 to 20 twists per foot. That way both Magnetic and Electric noise coupling largely cancels out in the two wires. Of course the same holds for the wiring that generates the interfering signal, although at 100A you won't be twisting those wires very tight! Every bit helps though.

By the way, shielding the data cables helps quite a bit with Electric field noise pickup but does practically nothing for Magnetic pickup.

The above applies to both Digital and Analog communication. The following applies mostly to Digital but is also useful for high speed Analog implementations.

As to the requirement of un-shielded data cables due to speed, the customer may be shooting himself in the foot on that one. It is possible to get up to 100MHz on just a twisted pair if you have good transmitters and receivers, a moderately quite environment, and error correction. If the data delay really is critical then twisted, shielded cable is the way to go (or a fiber optic link).

The shield not only helps keep out interference, but primarily it creates a line with controlled impedance throughout its length. Without the shield any change in nearby material will change the cable impedance at that point. The signal will be slightly reflected backward at that impedance change, and be reflected forward again at the previous change.
(For a worst-case demonstration, take a 10-20ft. length of rope and tie one end to a door knob. Take the other end in your hand and walk away until the rope sags 1 to 2 feet in the middle. Now rapidly shake the rope up and down once, and watch the wave reflect off the far end and come back to your hand.)

This decreases the Signal-to-Noise Ratio (SNR) at the receiving end, and increases the Bit Error rate (BER). To get back to reliable communication, you then must decrease tha data transfer rate.

The other requirement for truly high speed communication is that the output impedance of the transmitter must match the characteristic impedance of the cable, and the far end of the cable must also be terminated with its characteristic impedance. (That rope & door knob example above.)

All of that said, it is hard to imagine anything operating at powerline frequency requires very high speed, after all at 60Hz powereline frequency the wavelength is about 3000 miles.

Anyhow that's a bit of background and my 2-cents worth.

Cheers,
Tom

 

[Dinoduck94]

Hi Tom,

The communication is via Ethernet Cat5; however I feel like the conversation is veering off of my original question.

The original question was referring to the interaction of the magnetic flux generated by 2 separate fields: one pulsing and one static; rather than what can be done to reduce EMI on communication systems. Above, mfb answered it by saying that fluxes to do interact, however if I'm honest I do not believe this to be correct; due to the push and pull effect that we experience with magnetism.
Don't get me wrong, both yours and Berkeman's input is definitely appreciated, it just doesn't address the initial question.

However further from your point about twisting the wiring that generates the interfering signal, there is interesting research that has taken place to try and put some mathematics to it (how many twists per meter, for what benefit kind of thing), if you're interested the papers are named below:

'Magnetic field Reduction of twisted three-phase power cables of finite length by specific phase mixing' authored by Matthias Ehrich and Lars Ole Fichte at the Helmut Schmidt University in Germany.

'Models of Cable Bunch formed by twisted three-phase cables' authored by Matthias Ehrich and Lars Ole Fichte at the Helmut Schmidt University in Germany.

'Reduction of Power System Magnetic Field by Configuration Twist' authored by Per Pettersson and Niclas Schonborg from the IEEE in Stockholm, Sweden.

 

[Dale]

Above, mfb answered it by saying that fluxes to do interact, however if I'm honest I do not believe this to be correct; due to the push and pull effect that we experience with magnetism.

You can look at Maxwell’s equations and by inspection see that they are linear in the fields. Since they are linear in the fields then the total field from a pulsed source and a static source together is clearly equal to the field from the pulsed source alone plus the field from the static source alone. That is how linear systems work.

The only way for the fields to not simply add is if they don’t obey Maxwell’s equations. But from your description this is just an ordinary circuit. Those obey Maxwell’s equations.