# Designing a polarity-reversing electromagnet with energy recovery

#### Hamzaha

Joined Feb 28, 2020
13
Hi there,

I'm looking to design an electromagnet that changes its polarity by the trigger of a pulse signal, by considering the lowest energy drop, I am looking for a sub-system to reserve the energy of the electromagnet to re-use it.
• If there are known circuits performing these requirements, what are they called?
• If not, what are the suggested yet recommended composed circuits to make it up?

Thus the high-level elements I've are:
- Electromagnet, with a purpose to change its polarity.
- Sensor signal that's a pulse signal, consider a push button.
- Sub-system to reserve the electromagnet energy when flipped of polarity, which my question primarily aims to find.
- Control system to reverse the electromagnet's polarity, as an H-Bridge.
- The system voltage is 5-15VDC.

Note that utilizing an MCU is available to be utilized in the desired sub-system if needed, but anyway it's preferred if it can be substituted easily with electronics.

With Thanks,
Hamza

#### Papabravo

Joined Feb 24, 2006
21,225
How do you imagine that energy is "reserved" in an electromagnet? How do you imagine that such "reserved" energy can be used? I think you are having some kind of "fever dream", or you have discovered some new physics.

#### Hamzaha

Joined Feb 28, 2020
13
How do you imagine that energy is "reserved" in an electromagnet? How do you imagine that such "reserved" energy can be used? I think you are having some kind of "fever dream", or you have discovered some new physics.
Hi Papabravo, ofcourse there will be energy losses, but as a coil it will be energized, and further a current flow in the de-energizing phase (to change the magnetic polarity).
Ofcourse this reserved energy is lesser than the energy used to energize it. We're on the same physics until now

#### crutschow

Joined Mar 14, 2008
34,392
Diodes across the H-Bridge switches to protect them (intrinsic body diodes if using MOSFETs) will recover the inductive energy back into the power supply capacitors when the inductor is turned off.
For maximum power return, you can use Schottky diodes across the switches.
You should need nothing more.

#### Papabravo

Joined Feb 24, 2006
21,225
So, what you describe is similar to what happens in an inverter that converts DC power to AC power. The inductor may or may not have a core, and the core may or may not be able to function as an electromagnet. Is that necessarily relevant for your intended purposes?

Joined Jul 18, 2013
28,669
What is this particular application?
Applications that use Electro magnets are often required to carry out a reverse polarity in the course of operation, one such common apps are lift magnets, where small parts are used in the lift process, and if reverse (De-Mag) is not implemented, they remain attached rather than released.

#### Hamzaha

Joined Feb 28, 2020
13
Diodes across the H-Bridge switches to protect them (intrinsic body diodes if using MOSFETs) will recover the inductive energy back into the power supply capacitors when the inductor is turned off.
For maximum power return, you can use Schottky diodes across the switches.
You should need nothing more.
Thanks, I'll carry a further research in this direction. I'll further inquire if got a concern.

So, what you describe is similar to what happens in an inverter that converts DC power to AC power. The inductor may or may not have a core, and the core may or may not be able to function as an electromagnet. Is that necessarily relevant for your intended purposes?
I've highlighted this electromagnet to do the job:
https://www.aliexpress.com/item/1005005988571046.html
Could there be any problem to not function as a magnet?

What is this particular application?
Applications that use Electro magnets are often required to carry out a reverse polarity in the course of operation, one such common apps are lift magnets, where small parts are used in the lift process, and if reverse (De-Mag) is not implemented, they remain attached rather than released.
Yes it's similar to that, but I also need further to reverse its polarity periodically (by a pulse signal).

#### Alec_t

Joined Sep 17, 2013
14,312
I also need further to reverse its polarity periodically (by a pulse signal).
How much energy do you think will be needed to carry out the reversal, and where will that energy come from?

#### Hamzaha

Joined Feb 28, 2020
13
How much energy do you think will be needed to carry out the reversal, and where will that energy come from?
I don't have calculations right now, but ofcourse the same energy to carry the first polarization.
So what I see the cycle is: there's a polarization phase, energy recovery phase (less than the polarization energy), reverse polarization phase, energy recovery phase.
Both polarization phases are triggered by a signal.
But since the energy recovery phases do not have a dedicated signals or intended timing, I think the signal would trigger the de-energization phase with recovery phases and to carry on flipping the polarization phase.

The energy would come from a power source, say a battery.

#### Alec_t

Joined Sep 17, 2013
14,312
Your pulse source could be, for example, a '555-based oscillator. The oscillator would drive an H-bridge (see post #4) coupled to the electromagnet. But why do you want to do this?

#### Ya’akov

Joined Jan 27, 2019
9,117
Hi there,

I'm looking to design an electromagnet that changes its polarity by the trigger of a pulse signal, by considering the lowest energy drop, I am looking for a sub-system to reserve the energy of the electromagnet to re-use it.
• If there are known circuits performing these requirements, what are they called?
• If not, what are the suggested yet recommended composed circuits to make it up?

Thus the high-level elements I've are:
- Electromagnet, with a purpose to change its polarity.
- Sensor signal that's a pulse signal, consider a push button.
- Sub-system to reserve the electromagnet energy when flipped of polarity, which my question primarily aims to find.
- Control system to reverse the electromagnet's polarity, as an H-Bridge.
- The system voltage is 5-15VDC.

Note that utilizing an MCU is available to be utilized in the desired sub-system if needed, but anyway it's preferred if it can be substituted easily with electronics.

With Thanks,
Hamza
Welcome to AAC.

Your application is a mystery and it necessarily leads to speculation. Knowing the application is always important to provide competent consulting but in this case I think it is critical.

You know what problem you are trying to solve—but we don’t, and that leaves us without the context needed to make good suggestions. Sometimes it will even turn out the questioner has turned their “solution” into an unrelated problem that, once solved, will have no effect on solving the problem that provoked it.

So, please, give us a description of the problem this proposed solution is intended to solve so our help can be focused properly.

#### Danko

Joined Nov 22, 2017
1,834
I'm looking to design an electromagnet that changes its polarity by the trigger of a pulse signal, by considering the lowest energy drop, I am looking for a sub-system to reserve the energy of the electromagnet to re-use it.
This system with ideal components, with once charged capacitor, will work as long, as you want:

If the circuit shown in post #12 is used then it will not be possible to reverse the current in the inductance.
...
AND, the waveforms shown do not reflect any level of reality relative to the associated circuit drawing.

#### Attachments

• 1.1 KB Views: 1
Last edited:

#### Tonyr1084

Joined Sep 24, 2015
7,886
Two questions: First;
why do you want to do this?
Second; how often do you want to reverse magnetic polarity?

#### MisterBill2

Joined Jan 23, 2018
18,440
If the circuit shown in post #12 is used then it will not be possible to reverse the current in the inductance. That is because the negative end of the voltage source is permanently tied to ground, and also the one end of the coil is also tied to ground. So the alternative to connecting the positive supply to the one end of the coil is connecting it to a direct short circuit.
So I see no reason to spend any more time on this project until we see a reasonable explanation about what the goal is and by what means it would be achieved.
AND, the waveforms shown do not reflect any level of reality relative to the associated circuit drawing.

Last edited:

#### Hamzaha

Joined Feb 28, 2020
13
Your pulse source could be, for example, a '555-based oscillator. The oscillator would drive an H-bridge (see post #4) coupled to the electromagnet. But why do you want to do this?
Welcome to AAC.

Your application is a mystery and it necessarily leads to speculation. Knowing the application is always important to provide competent consulting but in this case I think it is critical.

You know what problem you are trying to solve—but we don’t, and that leaves us without the context needed to make good suggestions. Sometimes it will even turn out the questioner has turned their “solution” into an unrelated problem that, once solved, will have no effect on solving the problem that provoked it.

So, please, give us a description of the problem this proposed solution is intended to solve so our help can be focused properly.
I want to achieve the highest effeciency possible of driving the electromagnet that can push and pull permanent magnets periodically, which explains the 'reversing' functionality.

This system with ideal components, with once charged capacitor, will work as long, as you want:

View attachment 312565

Two questions: First;

Second; how often do you want to reverse magnetic polarity?
I see, a typical LC circuit but with a switch.
The reversing frequency is variable, I aim to minimize the idle state, so the triggering signal starts the reversing process, I think a switch over the coil that's closed is the way to go?

Context of the question:
Prior to asking the question, I've written down two ideas mainly depends on switches, but with knowing these will have some challenges and missing parts. However, reached to ask Google Bard which answered me that the functionality is already established in some known circuits: H-Bridge with Regenerative Braking, Bi-directional DC/DC Converter, Controlled Oscillator with Flyback Converter. Therefore I thought I was going to reinvent the wheel, thus wrote the question.
And after seeing the solution back to be as a simple H-Bridge with LC circuit, it encouraged me to share my previously written ideas.

My two ideas of circuits are as follows:
1)

NB: Ignore values

1. Starts with activating the first polarity of H-Bridge to complete a circuit (Switches SW1,SW4, and SW6).
2. On a certain time that capacitor's voltage is about to equal the source voltage and about to blocking the current therefore, open the previous switches and let the electromagnet continue to self-feeding by closing the short-circuit switch (Switch SW5).
3. On a trigger signal, close the inverse polarity switches of the H-bridge alongwith Switch SW6, and open switch SW5.
4. Continue as the point 2.
5. On a triggering signal, repeat the process.

2)

Here's another idea, where it goes in 4 major phases, presuming the circuit is kicked with conditions: (C1 charged by the battery as the polarity appears and the inductor is passing current clockwise (Left to Right)).
• Close switch SW1, resulting in LC circuit, Inductor discharges as current (Left to Right) --> Current stops with higher Vc1 --> Current flips in direction (Right to Left).
• When the capacitor voltage Vc1 decreases to 90% of its original voltage*, switch SW1 is opened and switch SW5 closes, resulting in L circuit with shorted ends.
Meanwhile switch SW2 is closed to fully charge Capacitor C1.
• On the triggering signal, opens switches SW5 and SW4, and closes switch SW3 to inverse the polarity, as explained in point 1 but the other direction, resulting (Left to Right) current flow.
• When the capacitor voltage Vc2 decreases to 90% of its original voltage, switches SW3 opens and switch SW5 closes, resulting in L circuit with shorted ends.
Meanwhile switch SW4 is closed to fully charge Capacitor C2.
• On the triggering signal, repeat the process.
/* this 10% (or any better number) due to the circuit losses, so feeding 10%Vc of energy each half-cycle.

Ofcourse both circuits would have missing parts in order to work as intended, especially the second one. I could be not knowing them currently.

To make it easier, we can assume the source voltage to be a 14.8V Lithium Ion battery pack (4*3.7V).
And for the signal generator, we can use a signal as 555-timer with a frequency controllable by a variable resistor.

#### ThePanMan

Joined Mar 13, 2020
792
Hamzaha your schematic doesn't make sense. How does SW1 interrelate to SW2? How does SW2 interrelate to SW3? How does SW3 interrelate to SW4? How does SW4 interrelate to SW5? How does SW5 interrelate to SW6?

Off the cuff, my thinking is a single DPDT switch would reverse the current through "THE RIG" (the "rig" being the LC circuit). Changing polarity to the rig using a DPDT switch would facilitate the whole switching operation. However, just what is the purpose of SW5 & 6? That's a lot of switching. Are those switches to be thrown by hand or by some program? If by programming then don't use switches in your schematic. Use BJT's or MOSFET's.

You were asked how often do you intend to switch polarity on the rig. You addressed it but didn't answer it. Switching a thousand times per second? Or 5 times per hour? Or other?

To be honest, the first thing I thought you were attempting was some kine of Over Unity device. That's a disallowed topic. I may be wrong but I don't readily see a use for such a project other than maybe attempting to get more power out than you put in. Can't be done.

#### MisterBill2

Joined Jan 23, 2018
18,440
"Pushing and pulling a permanent magnet which has a changing air-gap between the magnetic poles and the electromagnet will require much different current levels for similar amounts of force. Rotating a magnet inside a coil is similar to an electric motor. and the timing will need to be very close to recover any energy available from inertia.
So it does not seem reasonable that any recovered energy will offset the losses from switching. Really, it seems to be similar to an elementary DC brushed motor..

#### Hamzaha

Joined Feb 28, 2020
13
Hamzaha your schematic doesn't make sense. How does SW1 interrelate to SW2? How does SW2 interrelate to SW3? How does SW3 interrelate to SW4? How does SW4 interrelate to SW5? How does SW5 interrelate to SW6?
I think the issue is resolved in:

Are those switches to be thrown by hand or by some program? If by programming then don't use switches in your schematic. Use BJT's or MOSFET's.
Yes, the switches here are abstract, they are mostly being MOSFETs.
However, just what is the purpose of SW5 & 6? That's a lot of switching.
SW5 is to enable the electromagnet being continuously polarized by the continuous current.
I've reviewed the schematic and intended operation, SW6 is redundant where deactivating the H-bridge does the job.

Off the cuff, my thinking is a single DPDT switch would reverse the current through "THE RIG" (the "rig" being the LC circuit). Changing polarity to the rig using a DPDT switch would facilitate the whole switching operation.
You were asked how often do you intend to switch polarity on the rig. You addressed it but didn't answer it. Switching a thousand times per second? Or 5 times per hour? Or other?
It's a variable frequency starts with 10-1000 Hz, so the switches need to be transistors.

To be honest, the first thing I thought you were attempting was some kine of Over Unity device. That's a disallowed topic. I may be wrong but I don't readily see a use for such a project other than maybe attempting to get more power out than you put in. Can't be done.
I've read about that topic, some believe in it, but I've not seen a working one.

"Pushing and pulling a permanent magnet which has a changing air-gap between the magnetic poles and the electromagnet will require much different current levels for similar amounts of force. Rotating a magnet inside a coil is similar to an electric motor. and the timing will need to be very close to recover any energy available from inertia.
So it does not seem reasonable that any recovered energy will offset the losses from switching. Really, it seems to be similar to an elementary DC brushed motor..
That's a good point I've missed to consider, it's true that it needs to be solved.
I think there's 3 possible solutions:
- Mechanically limiting the other part from going away, brake or a mechanism as a spring to return it back.
- Control timing: Increase pulling time.
- Power: Increase pulling current.

Per mechanical limiting, A spring cannot be effective for different frequencies, but another mechanisms might do the job, I'll seek to solve it mechanically, because I think it'll make the circuit more complex and harder to implement.

Yeah, thanks for mentioning that. A DC motor is similar to what's intended.
Is it energy efficient, or considered to be? if so, how?

#### MisterBill2

Joined Jan 23, 2018
18,440
About 40 years ago my employer at the time produced an assembly machine that used a magnetic pickup for a rather small part. It used an electromagnet to pick up the part from a non-metallic nest, and then moved the part into position where it should be released. BUT the residual magnetism did not decay fast enough for the production rate. So we had to add a very brief burst of reversed polarity to cancel that. That short burst was delivered thru a capacitor, about 0.47 mFD, I think. Then the pickup magnet would retract and the revers capacitor was discharged by a short circuit, until the next cycle.. That is the other way to produce the reverse polarity burst. Not any power recovered, but the problem solved.

#### Irving

Joined Jan 30, 2016
3,868
Yeah, thanks for mentioning that. A DC motor is similar to what's intended.
Is it energy efficient, or considered to be? if so, how?
A good brushless motor running at its optimum design parameters is 70 - 80% efficient, but that wouldn't be the case here...

Incidentally, I was intrigued by your thinking, but I think the fallacy in your argument is the 'recovery' from the reversed connection. For any real-world electromagnet actually doing 'work' the losses are far greater than you might anticipate.

What you propose seems similar to the flywheel buses used in Scandinavia and elsewhere: in those buses a heavy flywheel is spun up to a high speed using an electric 'motor'. The flywheel then drives the wheels of the bus either directly, or using the motor as a generator driving electric motors at the wheels. The latter, though sounding less efficient, turns out to be much better & easier than a mechanical coupling, having around 55% efficiency overall.

Last edited: