I've just looked at the PMDX-126 manual.
The fault input should be left floating (aka unconnected) so everything gets enabled. To trigger the fault input/disable outputs, it has to be connected to 0V/Gnd.
For that to work with your drives, they need to be left in the default alarm output (active on alarm i.e. the + & - get connected on alarm), and then wired in parallel to the Fault input (all ALM+ to Fault input, and all ALM- to 0V/Gnd).

Regarding daisy chaining the Pul/Dir, I would personally do that at the BOB end, and use twisted pair/shielded wiring to minimise the risk of interference. The Enable signal is not as sensitive to noise, so I wouldn't worry about simply daisy chaining the Enable wiring across all the drives, if that was easier for wiring.

From what I understand from the manual, the Fault input once pulled low, disables everything, including the E-stop signal to the PC. The Error input doesn't disable anything, but triggers the E-stop input to the PC for 1 second, so is reliant on the PC disabling things.
The benefit of using the Error input, is if the drives allow re-enabling after a fault without needing their power supply cycled (i.e. turned off until the drives fully power down, and turned back on again), then resetting the E-stop via Mach and re-activating the drive enable signals should mean everything powers up again without the need for a power cycle. I know EM drives have the option to re-enable simply by disabling/enabling the enable input without a power cycle, but it has to be set via software.
The disadvantage being it is reliant on the PC handling the E-Srop.