DC Oscillator

[jonagik]

Hi,

If I wanted to have an oscillating, sinusoidal current in a DC circuit (for example a circuit with a battery, a resistance, and some other topology in series), how would I do this?

The first thing that pops to mind is the incorportation of an LC resonator. However, I can't see how to use this to achieve a sinusoidal current.

Any suggestions?

Thanks in advance

[billhowl]

An oscillator consists of an amplifier and a feedback network. Now, let us see which basic components are required to obtain oscillations.
'Active device' either Transistor or Op Amp is used as an amplifier.
'Feedback circuit' with passive components such as R-C or L-C combinations .
To start the oscillation with the constant amplitude, positive feedback is not the only sufficient condition. Oscillator circuit must satisfy the following two conditions known as Barkhausen conditions:
1. The first condition is that the magnitude of the loop gain (Aβ) must be unity. This means the product of gain of amplifier 'A' and the gain of feedback network 'β' has to be unity.
2. The second condition is that the phase shift around the loop must be 360° or 0°. This means, the phase shift through the amplifier and feedback network has to be 360° or 0°.

http://wikieducator.org/Sinusoidal_Oscillator

[Soeren]

Hi,

If I wanted to have an oscillating, sinusoidal current in a DC circuit (for example a circuit with a battery, a resistance, and some other topology in series), how would I do this?

What frequency and what voltage/current levels are you after?

[jonagik]

Power across the device wouldn't be too high (is there something I'm not seeing regarding the need for current/voltage values?). I'm looking for variable frequency.

I assume that there is an implementation of an LC resonator circuit I'm not seeing? Perhaps one where the dampening effect of resistance in the circuit is offset by a logic circuit which tops up the lost power?

[Soeren]

Hi,

Power across the device wouldn't be too high (is there something I'm not seeing regarding the need for current/voltage values?).

You tell me, I don't know what you need it for.

I'm looking for variable frequency.

What range?
There's a world of difference if we're talking say 1..100Hz or something in the MHz range.

I assume that there is an implementation of an LC resonator circuit I'm not seeing? Perhaps one where the dampening effect of resistance in the circuit is offset by a logic circuit which tops up the lost power?

I doubt that, but if you can specify what you need, in terms of what ranges of voltage current and frequency you need, it would be possible to help you.

[jonagik]

Thanks for your help.

Would you be able to explain the factors influencing design (or point me in the direction of a source) of a DC oscillator for the various frequency ranges?

Let's say I want a tunable oscillator in the low MHz range. How would you suggest I implement this?

[jonagik]

If a capacitance, inductance, and voltage source are placed in series (assuming no resistance) the current is a sinusoidal function of time.

Of course, if you add an equivalent series resistance (in account for resistance in the wires) there is a dampening effect. I assumed this dampening effect makes such a device impractical as it is imprecise. However, I wonder if, since the resistance is so low, it can be treated as zero thus making such a topology practical?

[jim5192]

I went down this path last year for my Circuit Design II project last semester. It's a waste of time unless you know how to manipulate nonlinear feedback (which is a grad school topic). The reason is that  in order to get a sinusoid generator to be stable requires exactly perfect component values in the Weiner Oscillator. Not only is this time-consuming/impossible to achieve, but once you achieve these values in the lab in a single sitting thevalues change with ambient conditions (temperature, humidity, etc.) over time, and your circuit may not respond the same the very next day. To achieve a stable, reliable circuit that results in sinusoid production requires nonlinear feedback so that imperfect component values don't effect the qualitative steady state response of the circuit, a sinusoid.

However, I was able achieve a tunable square-wave oscillator in the lab in 10 minutes using the Schmitt Trigger Oscillator.

Wikipedia provides enough information on this: http://en.wikipedia.org/wiki/Relaxation_oscillator

This circuit doesn't produce a square wave but you can use this to get oscillation and then actively filter out higher harmonics with a Chebyshev II Filter or a Butterworth Filter. I was able to achieve oscillation at 69 Hz and change. (only 1% from the design value!), which is tunable with one variable resistor. I'll note that my circuit oscillated very slowly, and you want a very high frequency, so you'd need to get some very high quality op amps with some high bandwidth in order to pull of active filter and even generation. Good Luck!!!!!

[Soeren]

Hi,

Would you be able to explain the factors influencing design (or point me in the direction of a source) of a DC oscillator for the various frequency ranges?

Let's say I want a tunable oscillator in the low MHz range. How would you suggest I implement this?

Life's too short for me to drag each bit of info needed out of you, so, unless you fill in the missing data, I'll rather go watch some paint drying.
I asked you before...
What range of voltages (i.e. min voltage and max voltage)?
What range of current (i.e. min current and max current)?
What range of frequency (i.e. min frequency and max frequency)?

And please... In numbers!
"Let's say low MHz" is not a specific range and says nothing of what you're trying to accomplish.

[jonagik]

Thanks for your response Jim.

Is a square wave or saw tooth wave acceptable in an RF transmitter circuit or does the current need to be approximately sinusoidal? I was operating under the assumption that for an RF transmitter the current needs to be appriximately sinusoidal, but if a square or saw tooth wave is acceptable that certainly makes things a lot simpler!

@Soeren my objective with this question is to learn about the concepts rather than produce an actual device, hence the lack of precision regarding figures.

[jim5192]

I couldn't say what type of signal you would need, but if you can generate a square wave (with the schmitt trigger oscillator), you can filter out the higher harmonics with filters.

I wish I knew more about the topic of RF circuit design. We have a class here that I was going to take but I won't get to. The book the class uses, which I purchased with the intent on ultimately taking the class, is extremely well written, and cheap. I recommend it if your goal is self-study.

David M. Pozar: Microwave Engineering (2nd Ed.)
ISBN: 0471170968

[jonagik]

Thanks for that - I'll take a look

[Soeren]

Is a square wave or saw tooth wave acceptable in an RF transmitter circuit or does the current need to be approximately sinusoidal? I was operating under the assumption that for an RF transmitter the current needs to be appriximately sinusoidal [...]

@Soeren my objective with this question is to learn about the concepts rather than produce an actual device, hence the lack of precision regarding figures.

Stating that you want to make a radio transmitter is quite an important bit of information, whether experimenting or not.

No, you cannot use a square, triangle, sawtooth etc. wave, as they would all have too much harmonic content. Likewise with  a sinosoidal wave, although in a lesser degree (depending on your description of sinusoidal).

Throwing words like "hf oscillator" after Google will net you more links than you will need for a lifetime, with lots of different configurations, voltage demands, topologies, crystal controlled or not etc.