There are several topics up on the Lynxmotion forums on how to debug USB/Serial communications to the SSC-32 or other processors, such as: http://www.lynxmotion.net/viewtopic.php?f=2&t=4702
If that topic does not help, I think I would start by breaking down the problem to see where the issue is.
First thing I would do would be to double check the jumpers on the SSC-32. Things like what baud rate jumpers are in place. Also do you have the TX/RX jumpers in place to use the RS232 connector? Is the VS1=VS2 jumpers in place?
If you have not done so, I would download the program: Lynxterm (http://www.lynxmotion.com/images/files/firmware/ltrm111.zip
I would then configure lynxterm to your Serial port number and baud rate. Make sure the baud rate matches what you have the SSC-32 baud rate jumpers to. Then click on connect.
Hopefully it connects... Then in the terminal window type: ver<cr>
Hopefully it will return the version string like: SSC32-V2.07Alpha1A-EGP
If this worked than we know that we have the correct com port, baud rate and power to the logic side and the processor is still working.
If this does not work, I would personally probably unplug all of the servos from the SSC-32 and try again and see if that works...
Also: when you first supply power to the SSC-32 does the LED come on? If so when something tries to communicate with the SSC-32 does the LED flash? If so it says the SSC-32 is getting some input, does not say it is valid input or even the right baud rate, but it at least is seeing something.
If you get the "VER" test to work, I would probably click on the "All=1500" button and see if all of the servos go to their center point. If that works it looks like there could be an issue with FlowArm which is a flowstone program... Unfortunately I have not done much with flowstone and don't have my Arm currently setup to hook up to a PC...
If none of the above works, I would probably start hardware debugging. Things like what voltage is going into the voltage regulator and what is coming out. Is that voltage making it to the Atmega chip... But I am pretty sure you know better than I how to do this.