LTspice – Instability of Fully Differential Op-Amp

Yes - sometimes the terms are somewhat misleading. In general, we have three different gain conventions for feedback circuits:

• Open-loop gain of the amplifier (alone): Aol,
• Closed-loop gain of the amplifier with feedback: Acl,
• Loop gain Aloop, which is the gain of the complete loop (to be measured or simulated after breaking the loop at a suitable point). Hence, the loop gain is the product $$A_{loop}=(-A_{ol}*\beta)$$ with beta = feedback factor
  Note that this definition includes the inverting sign at the opamp input.
• Based on these conventions the closed-loop gain is $$A_{cl}=(A_{ol}*\alpha)/(1-A_{loop})$$with alpha = forward factor, if existent, otherwise unity

Note that stability margins (phase margin, gain margin) are defined for the loop gain Aloop only. For the purpose of determining the margins, the loop must be opened at a suitable point (opamp output or inv. input) for injecting a test signal. I hope this clarifies something.

It is the purpose of the following example to demonstrate the meaning (and the correct sign) of the term"alpha":

Example:

(a) Non-inverting amplifier (R2=feedback resistor): alpha=1, beta=R1/(R1+R2);

(b) Inverting amplifier: alpha=-R2/(R1+R2), beta=R1/(R1+R2).

Here is a high-speed opamp (OPA211, 35MHz) with Rout of 25,000 ohms (edited up from specsheet 25 ohms). To cause oscillation, we add 1uF Cload. Rout+Cload are INSIDE the feedback loop, thus providing another 90 degrees phaseshift. Result? Notice the 25,000 ohm Rout drops the Fring by sqrt(1,000).

For more understanding, read up on Barkhausen's 2 criteria. They relate to the Nyquist plot/point crossing.