Using THIS enclosure and THIS drive, I tested several machines on USB speed here.

my MBP, left port: 16.5MB/sec
my MBP, right port: 34MB/sec
new MBP ports: 36.5 and 37MB/sec
fairly new MP rear ports: 37MB/sec

This, compared to 12-16MB/sec on a PMG5.
Also I've seen the ports on the front panel of a MP or PMG5 to be capped at FW400 speeds. (yes, MP has dual FW800 ports, and they both run at 400)

Plugging in the elgato tv to video on the MP, any port, while testing, resulted in speeds dropping to 26MB/sec.

Lessons learned: the computer can make a big difference. The port on the computer makes a big difference too. And using the other ports at the same time also can hurt. I know the isight uses the left usb bus on my MBP but I was surprised that affected my speeds when the isight wasn't even active.

I will be testing esata in a few days with THIS expresscard34 adapter. (3.0Mbps) With this drive installed in the MP on sata it did 86MB/sec so I don't know yet if that's the drive's limit or the sata limit. I'd expect 3.0Gbps to practically (9 instead of 8) work out to 340MB/sec (3.0*1024/9) but then I will also be wondering what the speed of the expresss34 slot is on my MBP.

Also if we followed my "divide by 9" rule, the expected max for fw800 would be 88MB/sec (which is SPOT ON btw), fw400 would be 44MB/sec, and USB HS (480) would be 53MB/sec. That really shows comparatively how poor USB efficiency is for data transfer. (70% of expected, based on claimed bitrate) USB is abouts FW330 if you wanted to compare, and that assumes ideal bus conditions which can have a devastating effect on speed.
I work for the Department of Redundancy Department