That's all very interesting, but did you have a question? Make sure you provide enough information when you ask your question to increase the likelihood of somebody being able to answer it.
In reply to Andrew Beckett:
In reply to Ivan13:
You could try using spectre's stb analysis to measure the loop gain; you just need an iprobe or 0V vsource in the feedback loop to act as the measurement device for the stability analysis. I'm afraid I'm unfamiliar with the specifics of SRAM design to be able to give you anything more specific. The analysis will then allow you to plot the loop gain whilst keeping the loop closed without disturbing anything.