Do I need to bias down sine wave?

I have soldered a simple module, that produces a sine wave. It works and outputs about 600Hz sine, biased up to 2.5V with amplitude about 0.5V. It is good for playing with arduino ADC applications, but mostly I want to use it as source to test different amplifiers, like transistors- or LM386- based ones. Do I need to (and how could I) bias this sine down for two and half volts, to 0, to get the "true" input signal for amplifiers?

That would depend on the amplifier. Many are AC coupled, which means they have a capacitor in the input, and reject DC components.

A simple capacitor would remove the dc bias

Allan

Yes, I have decoupling capacitor at the output, so everything will be okay and I don't need to shift it down? Some amplifiers (like JFET inputs) requiring very low voltage at their inputs, will not so high amplitude impact or even burn it?

The capacitor "blocks" DC, so that "shift's" the DC (and the average) down to zero. To made sure there is no "floating" DC or "bias leakage" through the capacitor, add a resistor to make a [u]high pass filter[/u]. (DC is zero Hz, so it's blocked by a high-pass filter.)

Some amplifiers (like JFET inputs) requiring very low voltage at their inputs, will not so high amplitude impact or even burn it?

You're not going to damage an FET with a couple of volts, but if it's a high-gain microphone preamp or phono preamp that normally works with a few millivolts, and that's connected to a power amplifier, you could blow a speaker or something.

megavoid:
Yes, I have decoupling capacitor at the output

No, you have a coupling capacitor or DC-blocking capacitor. A coupling capacitor couples signals from
one part of the circuit to the next.

Decoupling capacitors prevent/reduce coupling between adjacent parts of the circuit by holding the
supply solid (so it doesn't carry signal/noise to the next thing it powers).

Your oscillator will be sensitive to the load impedance, you might do well to buffer its output with a follower,
then its less likely to react to the load (say by stopping oscillating).

Thanks everyone for answers! Great idea to add emitter follower, I'll do this if oscillator become unstable.

Or an opamp follower...

If you're testing harmonic distortion to really fine resolution, you will need more filtering of the sine wave because a simple transistor circuit like that is too non-linear.