To follow up a little; i think i found the issue; I'm using a usb-serial converter, and apparently, merely plugging or unplugging the serial end of the device into the brainstem also has the effect of "hiccuping" the DSATX, just as plugging the brainstem itself in does; causing it to reboot the DSATX microcontroller (Grounding issue maybe? One wouldn't think so as the usb, mobo, and DSATX should share a common ground, but maybe the innards of the USB-Serial dongle has some bizarre wiring...)
In my case, I had a usb-serial cable connected from the carputer mobo into the DSATX; when I turned the ignition, the DSATX would assert PS_ON and power up the mobo, which then immediately powers up its USB devices. Once the USB-serial adapter gets juiced up, it hiccups the DSATX, causing it to reboot and drop the power; the USB-serial device then powers down, and hiccups the device yet again. Now, since Vacc is already high, the DSATX almost instantly asserts PS_ON after coming back up, and the cycle repeats, until the DSATX just gives up entirely at some point.
From the serial console you can't really tell it's power-cycling because the micro-controller comes back up almost instantly, and just prints another status message to the screen, which is exactly what would happen had it not rebooted.
So, my next course of action is to try a plain serial connector, as I happen to have a spare serial header on the motherboard, to see if it exhibits the same behaviour. In the meantime, does anyone know if this is the correct behaviour for the DSATX?
Edit: looks like a couple other people have had similar issues with the brainstem; anyone out there using the USB brainstem; any better/worse than the serial?