Gain bandwidth product

Open loop gain of what?

It doesn't, necessarily.

John

Keep reading. To keep the hi gain amplifier from becoming an oscillator, the gain plot (Bode plot) should cross the 0db axis at 6dB per octave... a one pole low pass roll off rate... a capacitor across the feedback resistor of the gain stage...

It is not necessarily so. This effect (gain decreasing linearly as frequency rises) is caused by a single pole (one energy storage element with the energy supplied through a linear source, like a resistor) of a low pass filter. Many circuits have more than one energy storage element in them, and some may be in a high pass configuration, or in a resonant configuration. Those circuits do not have the gain decreasing linearly and in proportion to rising frequency.

But if an amplifier (such as an opamp) is being designed to have stable negative feedback programmed gain over a wide range of gain choices, it is very helpful to this end if the gain roll off with rising frequency, is dominated by a single pole, since this keeps the input to output phase shift near 90 degrees till the gain falls below 1.

Excessive phase shift around a closed loop can convert negative feedback to positive feedback. If this happens while there is still effective gain around the loop, the closed loop configuration is unstable and will sustain oscillations (a self reinforcing echo that creates copies of itself, that travel around the loop, instead of damping out).

I still have no idea what you, or the op, are hinting at. He mentions open-loop unspecified somethings, which don't oscillate and don't necessarily behave as he thinks wikipedia predicts. You seem to be talking about closed-loop amps, which can be stabilized in all sorts of ways, 6 dB/octave open-loop transfer function being one of the least interesting.

This is all too fuzzy.

John

Op-amps?

Because that's where the gain bandwidth figure product come from !

Eg if an op-amp has a gain bandwidth product of 1 MHz, the gain is 1 at 1 MHz,

There will be a maximum gain at low frequencies where it stops increasing though.

Look up dominant pole compensation for more details.

Graham

Who knows?

John