Im no expert on this subject but I believe if you put a aprox 1.8k resistor from the input to the neg voltage it will bias the opamp into class A....There is a rather long thread about this somewere on the forum that deals with Biasing the OPA627 into class A...do a search and I"m sure you will find it....
I took it a step farther and already built a pair of JFET Cascodes as in: See below. Also you are right on with the 1.8 K resistor. That will give me around 8mA from my 15 volt rails.
Been at Tangent already (long ago) and built an adapted JFET Cascode he suggests in his article using a pair of 2n5484s and I made a pair of them. I also checked out some of the other articles (hyperlinks) to ensure that I adapted it correctly to my situation (analog output stage of CD player). I never found an example of exactly what I did anywhere. I even emailed back and forth with Warren. (He doesn't like helping newbies other than his writings, he feels the rest is up to them, even legitimate questions)
One well respected guy here also took a look at Tangent's article and didn't agree with everything there (too many assumptions).
It seems CarlosFM and PPL actually go into more depth about this situation and op amp in particular. So before I do it blindly, I want to ask them. Most people here suggest Tangent Soft or reading Walt Jung. Others say jumping the negative rail to the output of the op amp with a resistor or cascode knocks it out of balance. PPL is still here, I have to join another site to ask Carlos. But I want to challenge them on this (not in a bad way, just to understand what they know and why they say it is Ideal before doing it to mine). I already fried 2 OPA627s in the process the first time playing with this idea.
Thanks guys for helping me try to get to the bottom of this.