I haven't had good luck building physical sine wave generators. I want to make a fixed 60Hz sine wave generator.
Giving up on doing it myself, ebay has lots of boards with those out-of-production function generator ic's on them. But to do the distortion trim on them I would need instruments that I'd rather not buy.
It occurs to me, though, that if I follow the sine output of the chip with a regular opamp 12db/octave lowpass filter set at 60Hz, it would clean up a lot of the wave's distortion. Harmonic distortion is harmonics, right? If you filter out the harmonics you filter out the distortion, right?
I expect a lot of attitude from the ideaphobes here. It always happens. Knock yourselves out. If anyone has anything helpful or constructive or experience to share, that would be pretty keen.
Giving up on doing it myself, ebay has lots of boards with those out-of-production function generator ic's on them. But to do the distortion trim on them I would need instruments that I'd rather not buy.
It occurs to me, though, that if I follow the sine output of the chip with a regular opamp 12db/octave lowpass filter set at 60Hz, it would clean up a lot of the wave's distortion. Harmonic distortion is harmonics, right? If you filter out the harmonics you filter out the distortion, right?
I expect a lot of attitude from the ideaphobes here. It always happens. Knock yourselves out. If anyone has anything helpful or constructive or experience to share, that would be pretty keen.
There're many ways to clean-up a sine wave:
1) Resonant filters.
2) Sinewave shaping circuits.
3) Notch / comb harmonic filters.
4) Harmonic cancellation methods.
Each method has its advantages and also some disadvantages. Filtering the second harmonic using, say, a 12dB/ octave slope would give you much less than 12dB suppression.
My recommendation would be to use (4) that is, cancellation of selected harmonics using trignometric properties.
1) Resonant filters.
2) Sinewave shaping circuits.
3) Notch / comb harmonic filters.
4) Harmonic cancellation methods.
Each method has its advantages and also some disadvantages. Filtering the second harmonic using, say, a 12dB/ octave slope would give you much less than 12dB suppression.
My recommendation would be to use (4) that is, cancellation of selected harmonics using trignometric properties.
Last edited:
You could make a second-order peaking low-pass filter with a fairly high Q, the second harmonic then gets suppressed almost 4Q times. A state variable filter could do the trick.
Forget about function generators for sine waves.. Get a precision oscillator.
You can also use elliptical filter + 2T RC bridge [ https://www.patreon.com/posts/iellipticheskii-75204168 ] [ https://www.patreon.com/posts/khroniki-xxi-8-75057332 ]
For handling 60Hz using filtering techniques you'd also need a gyrator.
https://en.wikipedia.org/wiki/Gyrator#Simulated_inductor
https://en.wikipedia.org/wiki/Gyrator#Simulated_inductor
As you want 60 Hz and live in the USA, why don't you use a mains transformer and a filter?
To get ultra-low distortion sine waves, you really do need a sine wave oscillator stabilized by a very slow feedback loop. A low-distortion sine-wave oscillator has enormous Q.
Ed
Ed
Last edited:
Well, an oscillator is really what I want. Show me one that I can build and it will actually work.
The problem is that the AC line amplitude and frequency vary. Ultra-low distortion means stable as a rock.
ETA: Yes, the OP needs to adjust the design for 60Hz.
Ed
Originally, the thread starter wanted to use a function generator, apparently an old-fashioned one with an XR2206, and a filter. I would expect the mains frequency to be more stable than a typical XR2206 function generator.
The OP has not stated how low the distortion needs to be. The oscillators I am referring to have distortion lower than the THD wars.
Ed
Ed
@newvirus2008 - The high Q filter stabilized by a slow feedback loop works because an oscillator supplies its own input. Perfect output -> perfect input. The slow feedback loop makes only minuscule adjustments.
Ed
Ed
(Here) is a self contained quartz crystal controlled oscillator that runs at 6.000 Megahertz and has a frequency stability of ±100ppm. For USD 2.45 in quantity=1.
Connect the crystal oscillator's output to a digital frequency divider you've built out of counters and gates, which divides by one hundred thousand. Now you have a 60.00 Hertz square wave with a frequency stability of ±100ppm.
Put that 60.00 Hz square wave through a cascade of analog filters such as mentioned in post #3 above. Done. A frequency accurate, low drift, independent of AC mains, 60.00 Hertz sine wave.
Connect the crystal oscillator's output to a digital frequency divider you've built out of counters and gates, which divides by one hundred thousand. Now you have a 60.00 Hertz square wave with a frequency stability of ±100ppm.
Put that 60.00 Hz square wave through a cascade of analog filters such as mentioned in post #3 above. Done. A frequency accurate, low drift, independent of AC mains, 60.00 Hertz sine wave.
Yeah, you're right. I saw somewhere that the standard for the power companies, which few of them meet, is two to five percent distortion. I'm going to use this oscillator to run a turntable. Those are the numbers I have to beat. I think that if I can get under one percent, I'll be fine.
I haven't seen much to dissuade me from the cranky old function generator and a low-pass filter. Oscillator circuits that I see strike me as a little unstable. I'm sure as heck not going to build a circuit board the size of a sheet of typewriter paper. This is not a Dreamland project.