Sorry for coming in so late on this "juicy" post but I've just gotten back home from 4 days in the hospital getting a gall bladder out and a hernia repaired. Feeling much better already.
I've seen a number of these kind of upgrade situations over the years and agree that this job could be done well with a modern sensorless vector (yes, I would definitely use ABB's ACS800 drive) and a good TENV squirrel cage induction motor (that would be a Marathon Black Max). As you can tell, my general line here is very similar to what kamenges has been suggesting.
Getting to sizing specifics, you have to start with load speed and torque requirements. I believe that a top speed of 2100 rpm has been determined. Torque requirements are a bit unclear but 145 in-lbs has been mentioned. That is likely peak torque with some data presented that indicates a 6/1 ratio in input current peak/steady state. That would make steady state torque around 25 in-lbs. From Big John's picture (what a luxury!) I see a 143/5TC motor with a backside encoder and what appears to be a fairly open space environment around the motor (again, what a luxury!).
How here we come to a critical point. That encoder might must be a speed feedback device for the control speed loop or it might be that PLUS the position detection device for the positioning controller. If it is the former, it can be removed from the system. If it the latter, it must be reused or replaced in some way in the new system.
To size the motor, the peak torque is going to control. At 145 in-lbs, that equates to 12.1 ft-lbs. An induction motor is conservatively capable of 200% peak torque so that would be 6 ft-lbs continuous. Using the formula HP=Torque x rpm/5250 and solving for Torque, that would be a 2hp four pole motor. Checking my Marathon motor data sheet, I see that a 2hp four pole Black Max motor (part number Y551) is in a 145TC frame, is dual voltage wound at 230/460V, and has a backside mounting face for an encoder. That would be my motor choice.
The drive sizing is a bit more delicate (is that a technical term?). Clearly, it has to be 230V three phase input but if I choose a 2hp normal duty or variable torque drive I'll only get 6 x 110% peak torque. If I choose a 2hp heavy duty or constant torque drive I'll only get 6 x 150% torque. Breakdown torque in the selected motor and in general, all NEMA B induction motors occurs at about 240% nameplate amps so we need a drive capable of 6 x 240% = 14.4amps short term (five second rating would be enough here but using the 1 minute rating wouldn't hurt anything except add a little cost). Checking my drive rating data sheets, that would be a 3hp heavy duty rated 240V drive. It has a 1 minute rating of 13 amps and a 5 second rating of 24 amps). This drive is not going to need any kind of encoder to achieve the performance levels needed so that is not an issue. Braking is an issue probably so, since this drive has a built-in brake chopper, all that would be needed is a properly sized resistor. Checking my resistor sizing chart, a 22 ohm resistor with at least 285watts rating is needed for full drive capacity braking 3seconds out of every 30 seconds. That's probably more than is needed because you won't be stopping that hard. But the duty cycle is a bit higher so I wouldn't go under the 285watt rating. Resistors are cheap anyway!
We have some control issues to deal with. First is the positioning. The existing controller sounds like it works well so why not keep it. I'm guessing here but likely the existing encoder on the motor backside drives the positioning controller so why not keep it? It sounds like Big John has the ability to do some machining in his facility so why not build an adaptor flange and coupling for the old encoder and mount it on the backside of the new Marathon motor.
Next, retain the old positioning controller if it is a separate piece or, if it is buried in the existing motor amplifier/drive package, just retain the whole thing and disconnect the amplifier/drive power circuit keeping the old positioning circuitry intact. If it fails in the future, you can deal with a new positioning encoder/tach at that point. (Actually, as we speak, ABB is beginning to offer a positioning front end for the ACS800 but I have to be careful just how much I say about that on here).
The next issue to deal with is the communication signals between the old controller and the drive. I don't see any information about that in this thread so far but that shouldn't be a big deal. Basically you've got Run/Stop, Forward/Reverse, and an analog speed signal of some sort. Or, if the Run Speed is always the same, you can program the Fixed Speed on top of the Run signal and always run at the same speed. If reverse is a different fixed speed, you could program that on top of the Reverse signal. My point is that the inputs of the drive are highly configurable and, if that doesn't do it, a signal conditioner between the positioner and the drive surely would.
The one remaining issue here is to get a system properly programmed and tuned up to take advantage of its capabilities. None of the above will make anybody happy lacking that. I would suggest someone that carries an ACS800 Startup Certification, as a minimum. Let's see here---Poplar Bluffs AR, Kalamazoo MI, I could do that! But then so could a lot of other people.
Good luck!