I can get close Average STFT's but the range within a given cell seems awfully wide, for example say at 2800 rpm and a given manifold pressure, a specific cell can vary from +15 to -10, it is not uncommon for the range to generally be 10-20 percent plus and minus even if the average is within a percent or two...Is that abnormal?

Same if i track it with MAF tuning for a given frequency, generally a plus AND minus say 10 for a given frequency cell that may have an AVERAGE of less than 1 percent error. Seems to me I've seen folks mention a plus or minus of 1 or 2 percent, I'd be happy with 5 percent.

In tracking the 02 (mv) sensors bank to bank and graphing the oscillation i see sometimes each bank totally in sync, then at times its like one is slower than the other and will be totally offset, ie .1mv vs .8mv at the same location bank vs bank. I can not find a pattern and it seems erratic, for example at a steady state throttle, they will track extremely close for a while, then totally shift where they are opposite oscillations with one high the other low. Naturally when they track closely my fuel trims for bank 1 vs 2 are similar, when the 02 readings flip flop, of course i get a 5-6 percent fuel trim difference between the banks.

I did have one bad stuck sensor replaced a very short time ago. Anyway is this a sign of perhaps the other sensor going bad or??

Doing only closed loop STFT on narrowband, LS3/480, headers, no cat.