Just for fun, I decided to characterize a few bulbs I have and t use one to build a wien bridge oscillator.

I have 3 small bulbs. They are, 5V 60mA, 6V 40mA, and 12V ("grain of wheat") bulbs.

I connected a voltmeter in parallel and an ammeter in series to measure the voltage/current over a range similar to what I might expect to find in this circuit.

As it turns out, the graphs look surprisingly similar, so here are the graphs of the bulb I eventually used (the 12V grain of wheat bulb).

First is the raw Voltage vs current graph:

Next, the more useful voltage vs resistance graph:

So, I build a circuit like this:

For my circuit R1 = R2 = 5k9, C1 = C2 = 47n, and R3 = 100R. The op-amp was a 741.

The frequency is calculated as being around 573Hz.

R3 was chosen so that (R3 + RL)/RL woud be greater than 3 with a cold lamp, and equal to three at a sensible amplitude (which turns out to be between .25 and .75 VRMS).

I picked 100R because that gives a gain of about 5 when cold, and a gain of about 3 with a resistance of 50 ohms (or just over 0.5VRMS)

And so I built it:

Not really exciting. Note that you can't see any glow from the bulb.

And the output looks like this:

The upper trace is from the the oscillator, the bottom one from a signal generator at close to the same frequency (around 575Hz).

The top trace is about 4V peak to peak. That suggests the voltage across the bulb is about 1.333 V ptp. And that's about 0.47VRMS, which is a little lower than expected.

This image was taken after the oscillator stabilised. After power is applied, it can take 10 seconds or so before the output settles down. The instability is in the form of amplitude variation.

I don't have anything to measure the distortion of this sine wave, but It looks pretty good to me, and much less than 60%!