Might be some SMART utility or Windows itself doing SMART checks every x minutes, preventing a spin down. Or the Intel RST driver issuing commands for who knows what behind your back. Without source code hard to tell what that switch is actually doing. If you can, try with a single really old 100-500 GB disk to narrow it down command-capability-wise.
This is difficult to achieve under Linux even. Under Linux I had to tweak smartd to check for sleep first, before a SMART-check command is issued. Otherwise all disks wake up again. Similarly there is a package on Linux called udisks2, which will poll the SMART status every 10 minutes, keeping disks awake. Gets installed with every desktop installation so a user can e.g. mount USB thumb drives without being root. Had to patch the package to increase the interval to 60 minutes. Then, I needed to use hd-idle to issue a sleep command to the disks when nothing happened within the last 30 minutes. Because another problem is that modern disks might not support the legacy APM set-sleep-timer-commands anymore, only EPC (Extended power Conditions). I tried both APC and EPC timers on my 14 GB HGST drives, so drives sleep by themselves after a while, without external intervention, but both methods were in my case inoperable. Hence hd-idle as the last straw. There's a software called "openSeaChest" which lets you experiment with EPC. All in all, pretty damn frustrating.
RAID 0 with 6 SATA disks btw is what I call audacious. One dead disk and all data, poof gone.
