Audio Oscillator

I'm looking to build a self-contained, adjustable frequency, adjustable amplitude audio oscillator.  I've done some searching on the forum and haven't come up with useful circuits (plenty of references to circuits that work, but nothing that looks like it has been built).

I did find a project on ESP that appears to do what I want:

http://sound.westhost.com/project22.htm

Has anyone built this circuit?  It has an option for square wave using a Schmitt trigger but i'd also like to support triangle waves...it seems like this should be pretty simple by just running the square wave output through an integrator but thought I would open this up for suggestions.

Matador, a NE555 Timer should do the trick. (search for "555 Audio Oscillator")

/R

It is far easier to make a square wave or triangle waveform, than a low distortion sine wave.

The classic old school function generators used a triangle wave as the basic frequency generator, then made square wave and sine waves from that triangle wave (google diode break amplifiers).  IIRC the old 8038 IC was popular for making cheap and dirty sine waves (emphasis on dirty).

If you want a clean enough sine wave to use with premium audio paths, you may want to use something like the wein bridge or phase shift oscillator with more sophisticated level control circuitry. This is the heavy lifting for low distortion. 

JR

> adjustable frequency ... running the square wave output through an integrator

As you vary frequency, the amplitude changes. 10V at 100Hz to 1V at 1KHz.

You MUST decide between "good sine" and "triangle with OK/no-sine" (or something complicated).

These days the only real call for "good sine" is distortion measurement. Below 0.5% THD, that's very specialized, and usually a joint-project with the analyzer.

Max RMS Power measures are not badly affected by "OK sine".

The root tool is variable integrator and Schmitt trigger. ONE pot. Gives triangle and square directly. The triwave can be shaped to "OK sine" several ways.

There's chips for this. XR-2206 is one, there's another (ICL8038? MAX038?). They are obsolete but can be found here and there, even as kits. PartsExpress has a $42 kit. eBay lists several Heath IN-1271 and an Elenco at similar prices. Also a heath IG-5218 which is true sine and reasonably low THD, nice machine.

There's also "DDS Function Signal Generator Source Module Wave 2MHz" with LCD readout and interesting specs for around $40.

JohnRoberts said:

IIRC the old 8038 IC was popular for making cheap and dirty sine waves (emphasis on dirty).

Click to expand...

I once made a variable speed capstan driver using an 8038 and a 100w MOSFET amp. The amp could have been bigger, but it was stable enough.

Thanks for the feedback!

I had convinced myself that I needed sine, however when I sat and thought about it I think I really only need something that can inject a test signal (to check heartbeat), allow me to do things like check for distortion or clipping, and can provide something near an impulse response for checking tuning.  I think I can do all of those things with just triangle and square waves.

I might pick up one of those "DDS Function Signal Generator" boards on eBay as it seems to do exactly what I need, as far less cost that rolling my own.

I've said this before but computer sound files played through a sound card or MP3 player can cover a lot of ground.

JR

Actually, we have field apps guys that travel half way around the world to visit customers with product issues, that carry their cellphone and a 3.5mm -> RCA cable with them. Your right John, it's amazing how much you can achieve with such a simple system.

I often wonder about the out of band noise with such a system. If you wanted really clean sine waves, then you'd need to filter pretty strongly.

I've looked into some of the codecs uses in cellphones, and while the A/D section is typically reduced quality for voice input, the D/A sections are generally full audio resolution. Perhaps not up to big dollar test equipment, but unless the rest of the cellphone audio path is compromised, the audio playback could be very respectable.

Playing back a full resolution digital file could be reasonably clean, of course YMMV.

JR

JohnRoberts said:

I've said this before but computer sound files played through a sound card or MP3 player can cover a lot of ground.

JR

Click to expand...

I'll scope up a test oscillator out of my Fast Track Ultra's line-outputs and I'll post the scope shots.

I suppose I'll have to come up with a calibration/mapping between dbFS (what the SW meter shows) and Vrms on my particular sound card into a bridging input impedance?  I suppose once the scope is hooked up, I can just adjust the output slider until I see what I want on the scope (peak-to-peak or otherwise).

Building your own sig gen is one of the rights of passage for an analog designer. Just for giggles, here's a simple low-distortion sine wave generator I designed using one quad opamp. I wanted something pocket sized that only used one knob for tuning and would run off a 9V battery, but it could easily be modded for a proper bipolar power supply and more output.

Range is 20Hz to around 15kHz, but I did try it with another range (using 1n timing caps) and it worked fine up to 80kHz. Maximum output about 2.5Vp-p, constant to within 0.3dB. Distortion less than 0.5% in the higher range, reaching about 1.5% in the lowest range. Any LDR should work providing it has a dark resistance in excess of 100k.