this circuit when set up like RG's polarity reverser appears to provide the proper result.
I used an input buffer then 22k input and feedback R with the 1M bias R
well, I'm nearing completion of a certain circuit that employs this DCOA as a polarity reverser. I had this on the breadboard and things seemed to work just fine. Then, when I built it on a board, I was getting an oscillation. After going buggy looking for my mistake, I found the original article and realized that I had not included the bypass cap on the lower 10K R as noted. When I included this cap, one 'mode' was fixed, but then the circuit was misbehaving in its other 'mode'.
I added another bypass cap at the 6.8K R - now things appear to be fine.