Some additional thoughts on your EPS problem.
I don't have the service manual for the 1994 and aside from the change to a single control unit / power unit, there were some other changes and I think some of the mid to later EPS na1 cars got a rack change-out to deal with some problems. All I can refer to is my 1997 service manual and in the 1997 service manual the diagnostic flow chart for code 22 (same chart for 22 and a bunch of other codes) does not point a finger at the EPS rack - unless the rack motor has an open circuit. Your rack seems to work unless you run it up against the stops so you do not have a permanent open circuit.
The description of code 22 in the service manual is opaque. It just refers to a lower current limit with no description of what that means. Pure speculation is that this means lower than expected motor current which would jive with all of the diagnostic tests listed for code 22 which seem to look for bad electrical connections. When you did your electrical connection checks, did you check the large cable connection from the control unit to the rack drive motor which would include checking for damage at the terminal plug and the EPS rack connection?
The off-the-wall suggestion I have is that since the error only occurs when the rack is run up against the stops is that perhaps there is some burning on the commutator or the brushes which cause a reduced current at that position. Why this would occur is a question since the controller has a current limit function that should reduce this risk. However, if the rack was run up against its stops repeatedly, sustained and hard then the current limiting function might have been inadequate resulting in commutator damaged - which does seem like a reach to me.
Honda treats the rack as a non repairable item and in the '97+ service manual does not provide any details on the rack. However, I am thinking that it might be possible to check for damage to the motor by measuring the resistance of the motor (Honda does not provide a spec for this). To do this, set up the following test conditions
- electrical connections to the rack disconnected,
- the front end of the car jacked up so that the front wheels are off the ground
- battery disconnected,
- ignition key in position to allow you to turn the steering wheel and
- ohmmeter connected across the rack motor terminals.
Measure the drive motor resistance in the dead ahead position. Turn the steering wheel so that the wheels turn from dead ahead to the limits in either direction and watch the resistance value. There will be slight fluctuations as the brushes cross commutator segments. If you get a big increase in resistance particularly at the rack limits then that may be an indication of commutator damage or brush damage at those limits. If the motor resistance does not change over the range of the rack travel then the problem may not be with the rack. Note that this test presumes that when depowered the drive motor remains coupled to the rack. If there is some type of clutch which allows the motor to freewheel if there is no torque coming out of the motor then this test will not work. Since Honda offers up no details on the mechanical guts of the rack I can't be sure. I am more familiar with column style EPS units and in those units that I have played with the motor remains coupled when depowered.
If you do detect damage to the motor that might be quite repairable. There are specialists who do commutator repair and in the best of circumstances your problem might be just damaged brushes.
However, the first thing to do would be to check the service manual applicable to the 1994. I am offering up my comments based on the 1997 service manual which may not be 100% applicable to the 1994 system. Clearly the description of error code 22 and the diagnostics for code 22 changed between 1991 and 1997. If the 1994 system is significantly different than the 1997 system then my test suggestion may be totally out to lunch.