Here are my teams experimentaly determined numbers for RPM.
http://chiefdelphi.com/forums/showpo...0&postcount=13
It sounds as though we did basically the same thing you are doing last year, and it worked wonderfully for us. I don't think its possible that your issue is being caused be the CIM motor differences themselves. More likely are uneven chain tensions, bent axles, or a too-tight center distance on one gear set. Before you put the motors in, did one side turn more easily than the other? Is one side now noticably more difficult to backdrive by hand than the other side, when the system is not even turned on? (if you didn't remove the internal pins in the drills, maybe don't try that)
I can say that, as a driver, adding the CIM motors to the drill motors made a huge improvement to both the pushing power and manuverability of our six wheel drive robot. I am a big fan of quad motors drives because they are within the reach of almost any team and are still competetive with more complicated shifting drive trains. I would not give up on this experiment, especially given than its the off season.