Op-Amp vs Crystal Oscillator: 40kHz Sine Wave

[digitalblggr]

Is it better to use an op-amp oscillator or a crystal oscillator to produce 40 Khz sine wave?

 

[Bob S]

See simple Wein sine-wave oscillator at

http://hobby_elec.piclist.com/e_ckt18.htm

Bob S

 

[digitalblggr]

I know I built that one and got around 40 Khz. I wasn't too happy with distortion and precision. I was considering building a quadrature oscillator, but wanted to know what people thought about using crystal ones.

 

[vk6kro]

In the above circuit, if you take the output only from pin 1 of the crystal oscillator inverter, it should be a fairly good sinewave. Its frequency will be accurately controlled by the crystal.

It is unconventional to to take the output from what is an input, but the waveform here will be better than at the output of the inverter.

If you did get some distortion, you could filter it to get rid of harmonics.

There is no need to wire up the circuitry involving the other 5 inverters if you don't want to.

 

[Phrak]

Actually he is asking for a sinewave output, vk6kro.

http://www.discovercircuits.com/PDF-FILES/125ksine1.PDF" one is claimed to have 0.01% distortion. It can be driven with the front end given by vk6kro. It's a bit busy on parts, and the amplitude has some dependence on the value of the capacitors in the low pass filter.

 

[vk6kro]

Actually he is asking for a sinewave output, vk6kro.

As I mentioned, pin 1 of that chip will have a sinewave on it.

 

[Okefenokee]

Doesn't anyone make a simple chip for this sort of thing? I've noticed that a common question is about getting a sine wave in the audio or ultrasonic frequency range. There are plenty of chips for digital clocks that could be used to create accurate sine waves in the higher frequency ranges (100's of kHz to 100's of MHz). The answers here tend to be about making one's own generator.

Maybe someone should create a start-up company that makes integrated, all-in-one, easily programmable, and accurate sine generators for lower frequencies.

 

[Studiot]

Doesn't anyone make a simple chip for this sort of thing?

The problem is that there are so many low frequencies!
Users mostly also want variable, ramp or special frequencies, rather than just a few fixed ones as you tend to get at higher ranges.
For mass production these days it is simple (and cheap) to use digital frequency synthesis which is why the 8038 and XR2206 went out of production.

 

[Okefenokee]

For mass production these days it is simple (and cheap) to use digital frequency synthesis which is why the 8038 and XR2206 went out of production.

Yeah, but not everything is about mass production. What about hobbyists and developers? A small breakout board with an output which could be tuned from 10Hz to 100kHz would invaluable. One could design something that could be tuned by adding highly accurate resistors or capacitors.

 

[sophiecentaur]

Yeah, but not everything is about mass production. What about hobbyists and developers? A small breakout board with an output which could be tuned from 10Hz to 100kHz would invaluable. One could design something that could be tuned by adding highly accurate resistors or capacitors.

But I'm afraid that practically everything IS about mass production. If it weren't for mass production there would be nothing for hobbyists to use (at a price they could afford). You basically have to grin and bear it and build around what's available - you have absolutely Zero consumer clout to make the manufacturers produce stuff especially for you. Just like in the Valve days, aamof.

This is why developers do so much development with simulations and emulations.

 

[Phrak]

Actually he is asking for a sinewave output, vk6kro.

As I mentioned, pin 1 of that chip will have a sinewave on it.

I read that as '1 pin'. I'm not at all clear how stable the amplitude of any given crystal, or variation in amplitude over a part numbers is from a driven crystal. Do you know anything about that?

-------------------------------------------

A pic is fairy easy to program. An output driven as a pwm and combined with an operational amplifier and some multiple of 40 KHz oscillator oscillator will give a clean stable source.

 

[vk6kro]

I read that as '1 pin'. I'm not at all clear how stable the amplitude of any given crystal, or variation in amplitude over a part numbers is from a driven crystal. Do you know anything about that?

-------------------------------------------

A pic is fairy easy to program. An output driven as a pwm and combined with an operational amplifier and some multiple of 40 KHz oscillator oscillator will give a clean stable source.

I have used this circuit as a local oscillator for receivers (although not at 40 KHz).
The input to a CMOS gate is very high impedance, so you have to take the output to some device that is also high impedance. To maintain the sinewave, this would probably mean a FET either as a source follower or a common source amplifier.

A I recall, the sinewave here could be very clean looking on an oscilloscope, although this may not always be the case, I guess.

The signal level is large (several volts p-p) and absolutely steady as long as the load does not vary. And, you can't match the precision and frequency stability of crystal control.

 

[Mike_In_Plano]

Man, there are a bunch of ways to tackle this. I've even used a square wave and followed it with a second order, underdamped filter (frequency at the peak) and gotten something pretty true and cheap.

I've gotten the ultimate purity out of wein bridges, but at a high price. The gain control either introduces distortion, or has so little effect that the circuit is tweeky.

You might try a digital divider driving a 4 bit R-2R A/D and filtering it with a cap at the end. It's surprising how far even a few bits can push the distortion down. If your a purist, make it 6 bits.

Loading low frequency crystals isn't much fun. It doesn't take much to swamp the oscillator. Definitely a place to use an FET follower and be careful with the leads.

 

[Phrak]

I have used this circuit as a local oscillator for receivers (although not at 40 KHz).
The input to a CMOS gate is very high impedance, so you have to take the output to some device that is also high impedance. To maintain the sinewave, this would probably mean a FET either as a source follower or a common source amplifier.

A I recall, the sinewave here could be very clean looking on an oscilloscope, although this may not always be the case, I guess.

The signal level is large (several volts p-p) and absolutely steady as long as the load does not vary. And, you can't match the precision and frequency stability of crystal control.

Thanks for the info. I've been curious ever since looking for way to construct a precision sine wave frequency and amplitude reference that wouldn't require tweeking.

 

[Studiot]

Catch 22

One of the big problems with constructing a low distorion oscillator is knowing or proving the level of distortion.

This is why many follow a well trodden design path.

It really depends upon what other equipment you have available to make this assessment.