Magamp oscillator (2023 Update)

On a sunny day (Tue, 3 Nov 2015 11:45:33 +0100) it happened Piotr Wyderski wrote in :

Transductors.. have gain, so why not?

Solid state diode -> tunnel diode, not much else needed.

Magnetic amplifiers are conventionally effective in an AC environment

- the power source is already 'oscillating'. Such a circuit can be made to be unstable to oscillate around the original conditions.

With a DC supply, you'd need the initial transient application of power to 'start' the circuit, which, from a theoretical point of view, is cheating. If forced to stop cold, it would not restart oscillation without temporary removal or reapplication of power.

The suggestion of a tunnel diode is one that introduces an external 'starting' oscillator. Remove the magnetics and it will still oscillate. There may be other ways to externally induce a magnetic circuit to oscillate from a static DC condition. In some environments this might be satisfactory.

RL

Hence my question, whether it would be possible to build an all-magnetic circuit.

Best regards, Piotr

so derive your ac with a spinning magnet/coil.

NT

Yes, an Alexanderson alternator would be an option. :-) But how about a circuit without any moving parts. Doable?

Best regards, Piotr

If you mean a DC powered circuit that oscillates using only passive parts and diodes, no relays or such, I've never seen it done. Except tunnel or some other negative-resistance diode.

I'm not sure if you can call a magnetic amplifier a passive part. It can exhibit admirable power gain, but needs to be powered by AC, for example:

The guy claims it has the gain of 2000. So recently, I have started to wonder if you can use such a device as a DC-powered oscillator. I agree, the answer "yes" would be a surprise, but maybe it is "yes" indeed?

Best regards, Piotr

On a sunny day (Tue, 3 Nov 2015 15:31:04 +0100) it happened Piotr Wyderski wrote in :

DC motor... :-)

And, in fact, it is a correct answer. Yes, it can be done that way. Now try without movables or argue it's impossible. :-)

Best regards, Piotr

You can also make an amplifier or an oscillator using varicap and rectifier diodes, but it also needs an AC pump.

Since nobody has done what you suggest in about 200 years of tinkering, I suspect it can't be done.

I challenge you to find the word 'oscillate' or 'oscillator' in any text devoted to magnetic amplifiers or saturable inductor circuits.

Even one for motors and generators....come to think of it.

You may find scattered references to 'over-unity' devices, usually ignoring the energy input to the practical circuits.

It's not blindness or some kind of intellectual tyranny - concepts of gain, control theory and damping are readily applied, but 'passive components'...are....just....that. You need a start-up mechanism that can be seen by the reactive circuit elements, even at micro or nano fabrication scales, for the DC-only application.

Once it starts, it might be carefully maintained through the non-linear magnetics' opposing switching actions, the contribution of capacitive reactances and probably some non-passive elements like diodes to steer current.

And then, if youre very skillfull, it would do what most other oscillators hope to do - almost nothing else.

RL

You don't need to challenge me, I have exactly the same experience after nontrivial amount of googling. There's nothing about it. Which is the reason I asked here, the people inhabiting this newsgroup know a lot about all kinds of crazy things.

Best regards, Piotr

Just use a mains transformer.

If you must run it off dc for some reason, thermal devices can oscillate to produce ac - getting the frequency up will be fun, but just about doable.

I don't know whether you'd consider an arc as a moving part, they can oscillate.

NT

If it can oscillate it exhibits gain. If it exhibits gain it's active.

NT

If you've got gain and time delay, you can have an oscillator. I very much expect the reason no-one does it is that it's about as practical as mountin g a gun on top of your head and trying to fight a battle. Possible yes, but sensible no, there are far better options.

NT

Well, tunnel diodes work.

Cheers

Phil Hobbs

Maybe, if you can find something with an "s" or "n"-shape knee in its reluctance curve. Metamaterial perhaps?

Mark L. Fergerson

I think a DC motor has a commutator, so there's a switch involved. Easier, would be a common doorbell buzzer (again, magnetically operated switch).

It's necessary to have feedback; a magamp has a DC control input, so you need to get from AC output, through some kind of rectifier, with a phase shift, to the input. I nominate resistive heater/thermopile DC generator for the rectifier. That way, everything but the phase shifter can be welded-wires construction. Hey, it's gotta be RUGGED, to be impressive!

That's already an oscillator! Done!