Hi Mick
Sounds like things are making some sort of sense at last. I think the reversal you have seen could be due to a slight offset error. Although this could be ajusted using the pot on the main PCB, I wouldn't bother as the CSMIO can set the overall offset for you at the press of a button.
Are you able to get a high enough motor speed with 9V input now, and is the motor running smoothly? If speed is still limited I would be inclined (maybe with Craig's approval) to try a small adjustment of RV4 which affects loop gain. Maybe a little iteration of the settings of the 3 pots RV1, RV2, and RV4 would give you the result you require.
I have always presumed the red is positive pin 1 on the CSMIO and the blue is negative pin 14 on the CSMIO.
Yes this is the correct connection. Remember to start with Kd = Ki =Kvff = 0, and gradually increase Kv. Start with Kv = 10 and increase in steps of 10x to maybe 100000. If the motor suddenly runs away, reverse the sense of the encoder feedback and try again. If the motor begins to oscillate, reduce Kv.
It's always safest to have your eStop act directly on the drive enable as well as going independently to the CSMIO so that Mach knows what's going on. At its simplest, you could just add another contact block to your eStop switch and wire this in series with the connection to the drive enable. Some folks use a Piltz (may be incorrectly spelt!) safety relay, but this isn't necessary.
Craig
I doubt if we'll ever agree, but here is my final take on things. The greatest advantage of the matched servo and drive is that the drive has knowledge of all motor parameters. This is especially important in the drive's current loop, less so in the velocity loop, and less again in the position loop. Providing the drive takes care of the current and velocity loops, all that's needed to implement the position loop is encoder feedback and the calculation of a few Z transforms. I'm sure that the fast processor within the CSMIO is perfectly capable of doing that. I use Chinese Kinco drives and servos as you know, and like the DMM these are packed with features and flexibility. I am not aware that any significant features are lost by setting analogue mode and using the IP-A for the position loop. The CSMIO offers all the tuning features you need, including the scope, in its plugin. Don't just take my word for it - you only have to look at posts from folks like Hood to see that the IP-A is capable of excellent results.
Allan