# Controlling the frequency of AC used to energize an electromagnet

I am trying to build an electromagnetic pendulum which essentially has a pendulum arm with a permanent magnet at its end which is 'kicked' by an electromagnet placed at the mean position to drive the pendulum without letting it damp. Now I was wondering how I can change the frequency of the AC to correspondingly change the magnetic field strength and obtain different amplitudes doing so. I came across VFDs and this build http://blog.hardcore.lt/mic/archives/011040.html.

I am fairly new to electronics. I am more of a hobbyist. This seems way too complicated, at least for me and i can't help but feel whether I am overthinking this. I already have my reasons for not using a 555 timer and just DC pulses to energize the electromagnet so that won't do. In case VFD is the best way to do this, can I please implore any one of you to dumb this build down for me or give me a simpler way of doing it which I am guessing is possible since I am not really using this for an AC motor?

• do you want bi-directional magnetic fields, or will the magnetic field always be in one direction with varying strength (e.g. do you need both positive and negative voltages across the electromagnet)? – helloworld922 Jun 1 '16 at 18:33
• Yes. That is my reason for not using just timed DC pulses. I want bidirectional fields flipping at the frequency of the AC. – K.C Karthik Jun 1 '16 at 18:39
• I would think that you use a magnetic switch (reed or Hall) to sense the pendulum passing and then pulse the electromagnet. This will automatically compensate for varying pendulum length and frequency. – Transistor Jun 1 '16 at 20:56
• But really cmon. Doubochinski's pendulum. Bethenod's pendulum. Or any Argumental pendulum. Google it! PLEASE!! There is even one of many technical papers on the subject - Cintra and Argoul 2016 - nonlinear argumental oscillations: a few examples of modulation via spatial position. This will tell you how you can build 6 variations of such a pendulum! – K.C Karthik Jun 12 '16 at 5:35
• It is easy enough to drive an electromagnet with a power amplifier (or a "power oscillator", for that matter) Yes, you appear to be over-thinking this. It appears to be a wacky laboratory curiosity. Is there any benefit to this experiment beyond the ability to say that you reproduced the phenomenon? – Richard Crowley Jun 12 '16 at 22:42

If it is difficult to answer your question that is because it is a puzzle to understand it in the first place. A pendulum is a RESONANT device in which the freqqquency of its swing is controlled principally by its effectiove length. The only effect of your electromagnetic "kick" is to replace thev energy lost in air-resistance to its movement. Any attempt to affect the dynamics of your pendulum must rerquire an enormous amount of power. Whatever may be your purpose you might be more successful with a conical pendulum.

• The aim is to not change the frequency of the pendulum but change the frequency of AC and I don't wish to change the dynamics but merely drive it with the magnetic field to overcome the drag resistance, friction etc actually. Google Doubochinski's pendulum and you'll find what you need – K.C Karthik Jun 8 '16 at 10:56

According the Wikipedia, Doubochinski's pendulum does not vary the ac frequency. Instead, the starting conditions dictate the steady-state amplitude.

If you merely want to keep the pendulum swinging uniformly using an electromagnet, the simplest way is to drive the electromagnet with a negative-resistance circuit. This will automatically sense the swinging of the pendulum and adjust the current appropriately to keep the pendulum swinging uniformly.

A fine and elegant early example of this is a tunnel diode motor (also a tunnel diode pendulum) that appeared in C. L. Strong's "The Amateur Scientist" column in "Scientific American" magazine for October 1965 on page 112. Besides the swinging magnet, it uses just a coil of wire for the electromagnet, a tunnel diode for the negative resistance, two resistors to set the appropriate bias voltage for the tunnel diode, and a 1.5 volt battery. The tunnel diode and swinging magnet apply a combination of dc and ac to the coil. (Mr. Strong says he also successfully made a pendulum with the tunnel diode powered with a battery improvised with blotting paper, moistened with saliva, and sandwiched between a nickel and a penny.)

These days, tunnel diodes are scarce, but the tunnel diode can be replaced with a simple negative resistance circuit made of a couple of FETs or a couple of bipolar transistors with a few resistors. Google "lambda circuit" for various examples. I have had good results using either the tunnel diode or a lambda circuit. You can merely change the bias voltage to change the amplitude of the swings.

If you are set on demonstrating Doubochinski's pendulum, you can use most any signal generator, as suggested by another responder. As long as you are using a typical pendulum, hanging from a low-friction pivot or on a thread, and swinging in air, there will be very little energy loss per swing, and a signal generator should provide plenty of power. If you don't already have a signal generator, you could use an el cheapo model such as a "1HZ-500KHz DDS Function Signal Generator" that you can get from a number of on-line vendors. It is settable in 1-Hz steps, giving good frequency resolution at all but the lowest frequencies. You said you don't want to use a 555, but if you don't object to an IC, you could use one of the resistance-controlled clock chips, such as from Linear Technology. With any of these schemes, if you want to drive a huge pendulum and you find you do need more power, you could merely add a buffer amplifier chip having a low-impedance output, or you could use an audio amplifier from a sound system.

• The steady state amplitude is brought about by velocity modulation by means of the frequency of AC depending on the phase relationship. You see it says a larger array of amplitudes become available with the frequency? That's what I want. How I will get it is by the initial conditions yes, but to be able to maintain the pendulum at some particular amplitude I need to be able to give different frequencies of AC. Supposing I want 59 degrees amplitude, I need to let go at 59 degrees but in order to be able to keep it there, I need to give 50 Hertz AC. I cannot go for less and expect the same amp. – K.C Karthik Jun 12 '16 at 5:46

Keeping a pendulum going is not about just producing the right drive frequency. The phase of the drive signal is critically important.

The point of a pendulum is usually to free-swing as much as possible. These devices are also designed to loose very little energy each swing. Put another way, they are very high Q filters. As such, you don't want to, and don't need to, drive the pendulum all the time. You only need to give it just enough boost each cycle to match the little bit of energy lost that cycle. This is easiest done with a small pulse once or twice per cycle. A electronic analogy is a class C amplifier.

To add mechanical energy, the pendulum needs to be pushed in the same direction as it is moving. Since it is moving most at the bottom (middle) of the swing, this is the logical place to add the energy. A simple way to do that is to turn on a magnet shortly before the pendulum reaches the bottom, then turn it off when it does reach the bottom. This is assuming a simple magnet that attracts the pendulum.

The tricky part is knowing when to turn the magnet on. You could use some kind of position sensor, like a light beam. However, these tend to take significant power. A cute trick is to use the electro-magnet in the bottom as a sensing device during one half-swing, then the driving device during the other. Since the pendulum period can be accurately known, measuring its phase once per cycle is enough to tell you where the pendulum is at all times.

You can even measure the period between detections to correct for minor variations in the pendulum's frequency. I did that in a for-fun project that used the pendulum period to measure temperature (the pendulum rod was copper tubing). It worked very well, and the thermometer was amazingly sensitive and accurate.

• The point is to not just build a driven oscillator although the driving is really important to me. Google Doubochinski's pendulum. It's a non linear oscillator which is what I'm trying to build here. The frequency of AC modulates the velocity of the pendulum to get discrete amplitudes. I shouldn't or need not influence the phase since the pendulum apparently, does it on its own, adjusting it's phase to get accelerated in the positive half cycle, decelerated in the negative half cycle. if there whole number full cycle, they won't get accelerated but anything else and the net accn is non-zero. – K.C Karthik Jun 8 '16 at 11:32
• @K.CKarthik, If you are making Doubochinski's pendulum (never heard of it) then you should put that in your question. Is there some reason you can't use a commercial signal generator? – George Herold Jun 8 '16 at 13:22
• It was a deliberate choice to not name it for a reason you yourself are a prime example of. Names can only confuse or deflect. The objective has been outlined but obviously not well enough and I'll take the blame for it. A signal generator would be perfect if I could only amplify power? How should I go about doing that? – K.C Karthik Jun 8 '16 at 13:39

There are any number of ways of driving an electromagnet from a variable frequency. If you want to just use a bench signal generator, you would need a simple AMPLIFIER to drive the electromagnet. This is exactly how speakers, headphones, earbuds, etc. operate. Depending on the impedance of your electromagnet, a common audio amplifier would possibly be a simple and cost-effective solution.

To amplify the power of the output from a signal generator you simply add a power-amplifier: i.e one which affords the ability to raise both the voltage and the current simultaneously. Are you asking for advice on how to buy, or how to design, such a unit?

• In a manner of speaking, yes. How should I put together say a signal generator and a power amplifier circuit together? A lot of cheap Audio power amplifiers I have checked out seem to have a limited power out put range but if I built an electromagnet accordingly, will I be able to use the audio power amplifier? – K.C Karthik Jun 12 '16 at 18:38

... can I please implore any one of you to dumb this build down for me or give me a simpler way of doing it ...

simulate this circuit – Schematic created using CircuitLab

Figure 1. Simple experimental setup.

I would be inclined to build a test setup as shown in Figure 1.

• Find a signal generator app for your mobile phone.
• Connect the headphone jack to the input of a suitably powerful audio amplifier.
• Find / wind a solenoid with enough resistance to meet the minimum required by the amplifier - typically 4 to 8 $\Omega$. If you feel the amplifier is complaining about the highly inductive load you can add a few ohms series resistance to damp it down.
• Set the signal generator app to 'sine' output and turn up the volume gradually monitoring coil and amplifier temperature for problems.

You may wish to measure inductor voltage or current to ensure consistency as you vary the frequency.