The old hand controller (HC) will not work as it doesn't have Pulseguide support. On the ASCOM-talk yahoo groups, I suggested that the Celestron's ASCOM driver be changed to support a “timer” based simultaneous axis pulseguide. An updated Celestron driver was released last month to support the ability of PHD to implement pulseguide directly to the Syncscan controller.
However, it should be noted, that the EQDIR option should be more accurate due to less delays induced on the serial interface since the pulseguide commands are directly fed to the mount's stepper board controller. This should provide the ability to have better subpixel autoguiding
Without the EQDIR option (using the hand controller) there are two “cascaded” serial interface delays (very critical to mount's reaction time to the pulseguide corrections). One delay comes from the serial connection between the PC and the hand controller (HC). The other delay comes from the connection from the HC to the mount. Since Celestron's Pulseguide commands protocol is entirely different from the Mount's Stepper protocol, the HC does some sort of command translation. This adds to the delay. Pulseguide will be successful using this option (PC→HC→mount), however the accuracy in terms of subpixel guiding would be less.