Wimberleytech, thanks. I have the model for the RCA part, CA3046 and as expected the bandwidth is much higher than the live circuit (I don't see the rise time increase through the amp). I wanted to play with as you suggest CJE and CJC to see if by increasing them I can in effect better duplicate the live circuit. Adding 50pf just about anywhere will cause the problem I am seeing in the actual circuit.
I find it hard to believe the NPN array failed as well as it having such a high junction capacitance from the start. I mean it is always possible, as I've only ever seen these used in low frequency applications but then again, that would only be about 4 or 5 data points.
Thanks again,
Jerry