Heck of a way to run an oscillator?

The amp just needs to be in its linear region to begin with

NT

"skeptical engineer"

You need to post a link to stuff like that.

.... Phil

That was probably a gain block fed back by the SAW device. And yes, this is a classical way of building an oscillator: the whole thing will oscillate at the frequency where the phase shift due to the whole loop is a multiple of 360 deg, provided the gain there is slightly more than unity. Startup can be explained by noise triggering the circuit's free response, which is an exponentially growing sine wave. Eventually, nonlinearities will set in, distorting the waveform and changing the oscillation frequency. As the SAW device is high-Q, with a steep phase slope around the intended frequency, this usually turns out to be a pretty good and simple oscillator. Just watch out for not having excessive gain.

Pere

OK - right you are.

This is posted at Mouser, but it is an Epcos doc. The question is, would an op-amp (generic amplifier) work in this circuit?

jb

Interesting. Relating to the SAW design itself, this brings up the issue of high frequency 'sound' waves in various media. At sound speeds of 3x10exp3 M/s, a gigahertz f wave will have a wavelength of 3x10exp3/10exp9 = 3 microns. in quartz.

That's not an opamp but a generic amplifier, and only meant as illustration, as an amp "building block". Yes, a fast enough opamp can do this but that isn't necessarily the wisest and most frugal approach. Cheapskates use logic parts for that because you get a six-pack for a nickel :-)

OK thanks. - What logic part would produce a sine wave output? jb

One that has no extra gain than what is required to start, so it's not constantly slamming against the amplitude limit. One that limits its amplitude into a resistive load. Or one fitted with an AGC circuit. It helps if the clipping is somewhat soft. Look at the 74HCU04 inverters for a start.

nd they just take an op amp with the SAW filter in the feedback loop. They explain that the op amp amplifies noise, but preferentially at the filter a llowed frequency. 2 components.

s is a classical way of building an oscillator: the whole thing will osci llate at the frequency where the phase shift due to the whole loop is a m ultiple of 360 deg, provided the gain there is slightly more than unity.

se, which is an exponentially growing sine wave. Eventually, nonlinearitie s will set in, distorting the waveform and changing the oscillation freque ncy.

nded frequency, this usually turns out to be a pretty good and simple osci llator. Just watch out for not having excessive gain.

of high frequency 'sound' waves in various media. At sound speeds of 3x10 exp3 M/s, a gigahertz f wave will have a wavelength of 3x10exp3/10exp9 = 3 microns. in quartz.

As usual, Wikipedia hasn't got it entirely right. There were acoustic micro scopes that went up into the GHz region, though thye may not have been comm ercially successful.

A 'U' inverter? Something like a 74HCU04?

According to Barkhausen, you just need an active element with sufficient gain and a frequency selective component in the (positive) feedback path.

When first switched on, the amplifier input related thermal noise is amplified, Part of it is going to the frequency sensitive network, which attenuates the lowest and highest frequencies, leaving a broad peak around the resonant frequency. This quite broad peak goes into the amplifier input, being amplitude to a larger amplitude than the thermal noise. After a few iterations, the lower and higher noise sidebands are attenuated and the peak amplitude at resonance is increasing, finally creating a narrow spectral line.

To sustain oscillation, the amplifier gain must be greater than the losses in the feedback loop and the phase shift must be such that there is a positive feedback.

Of course, the amplitude can't grow forever, but is limited by the amplifier output voltage swing. Alternatively, the amplifier gain needs to be reduced to keep the voltage within the amplifier output swing in order to generate a clean sine wave, such as in the Wien bridge oscillator.

Even if this Barkhausen explanation assumes a linear amplifier so that the thermal noise gets amplified in the first place, in practice, a class C amplifier might be usable, since the switch on transient will contain frequency components that will be amplified and looped back.

While such circuit would start reliably with a battery power supply and a power-on switch, the same circuit might not start, when plugging a power supply into a mains socket, due to the slow DC voltage build up and thus no high frequency components.

There are app notes on the Epcos sight that show a one ( UHF) transistor stage with many parts.

You ARE dreaming.

... Phil

Phase gain is a deciding factor..

"Clipping" is for digital guys. AGC is necessary for low distortion. ...Jim Thompson

Yes, AGC is how it's done. Or follow it with a resonant circuit or LC lowpass filter.

Filters, unless they're dead-on, introduce phase shift. ...Jim Thompson

For most oscillators that act as master or are in similar functions it does not matter.

Also, easy to lock up if the oscillation gets quenched.

Self-excited class C oscillators are rarely class C under all conditions, but nonetheless are often succeptible to quenching, which is irritating from time to time...

Tim

Usually, however, clipping/clamping/limiting pretty much invariably produces the lowest phase noise. Simple principle, is the more devices, the more noise.

I have actually spent some time running sims on just about every reasonable oscillator topology you can imagine for phase noise. One transistor wins.

Kevin Aylward B.Sc.