Seems to me that you have several issues that could and will be causeing this problem. . . First reduce the microstepping to 2000. Not much to gained above this it will also ease the PP work load. . . . Also good idea to run the driver test in Mach folder to test how well your PP works.
Then re tune your motors and see how it performs.

Next 24V is way too low for these motors and will restrict rapid speeds, this would be the first place I would turn to. . . . You could easily double this figure with these motors (If the drives can handle it)

Regards daisy chaining and grounding etc it's not really rocket science but is important if you want to restrict potential problems.
Basicly when running power from the power supply you want to run a separate wire to each drive not 1 wire to first drive then jump or "daisy chain" from this to next drive.

Think of the PSU as the centre of a star with all single wires running from it out to each component that needs to takes power from it.

The same applies to the ground wires, run each ground back to 1 central ground point. It's a good idea to have single ground connector with multiple connections, basicly you dont want multiple ground points around the machine other wise you can introduce ground loops which can cause noise issue's. Thou I'm pritty sure this is not your problem.

Their is another potential problem that can cause what you are experiencing. . . Mid band resonance.!!
This is caused by resonance from the machine which at a certain point in the accelleration curve interfears with the pulse stream going to the motors causing all sorts of random issue's.! rough sounding motors and poor performance as well has stalling as the speeds increase are just some of them.
Some drives have compensation for this built in some don't.! . . Not sure if yours do or not.?

If I was to place a bet on your problem I'd put money on it being down to 24V supply. . . And possibley too high on the stepping.

Also what Kernal speed are you running Mach at.?