I would concentrate on the MSO MSF as others are probably using the same motor in a hoist situation and it is working for them.
What times are you using on the Brake engage time eg 150ms?
What is the full motor part number - including the brake voltage etc
What is the ratio from the motor to the z axis
Do you have enough voltage on the brake at the motor ? I had one servo brake/motor that failed over time due to this. - sorry dmargineau seems to have this one covered.
As a trial - Can you use an external switch (output) to turn on /off the brake
with the MSO MSF (obviously do not move until the brake is lifted)
Get a trend of the servo Torque and Brake operation and Action Status
How did you tune the Servo?
Which version of RS5000?
Can you post the program? (Zip) If not then the just the Axis is better than nothing.
dmargineau Motion UM001 pdf - search "Brake Engage Delay Time"
command reference (Servo Action Status) versus Drive Enable is a bit confusing especially when added with a brake - if during the brake engage Time a load is on the servo the servo can move?
Michael, I have long ago given up on RA Literature Library contents as being user friendly and void of any personal interpretation inteventions requirements...
...They are getting better though...Maybe one of these days, they will contain only "black and white" statements...
It looks like the OP has done his homework and knows his way around motion applications.
We still have to learn about the history of the at hand system and, at least in my head, the original lack of a brake interposing relay, lit up a warning light bulb...
I am not sure if the application's functionality has been fully explored within the design stage; and I have a hunch that the x and y axii are identically "operated" even if, obviously, the z axis does a lot more "work" and its "reaction" needs to be as close to instantaneous as possible.
A brake will need an X amount of time to engage (Off command timing + internal springs tension/friction characteristics) and an Y amount of time to disengage (inductor magnetical characteristics + command/supply voltage application timing).
An electro-mechanical brake interposing relay has more On/Off timing lag than a solid state one and probably the latter will reduce the load drift
some; as you have proposed, a "faster" relay command source could also help since the application is very "close" to be perfectly functional.
The increase in load drift might have been induced by the mechanical gearing "loosenes" acquired during normal functionality; a new gear reduction provides more internal friction than a "broken-in" one hence the comissioning tunning and drive sizing do not quite apply anymore, especially if the "financially efficient" path has been originally chosen...
As stated within my previous post, the application might be slightly "undersized" for a perfect, long term functionality; a larger brake, providing high quality controls and the most expedite commands timing have been implemented, will engage faster (and disengage slower!)thus the drift could be further reduced.