Yes that is correct. You see, there are two clocks at work here. The scan clock and the channel clock. The period of the channel clock is the delay between channels also known as the "interchannel delay". This way, all the channels can be sampled as close to the edge of the scan clock as possible. Like you said, the rest of the 20us is just "free-time." Some people find this useful as it sort of simulates simultaneous sampling. If all the channels are sampled quickly at the beginning of the scan clock edge, for all practical purposes, they might as well have been sampled at the same time. However, some need to slow down the channel clock (or increase the interchannel delay) if they run into incorrect measurements due to settling time issues.
I hope thie h
elps. Let me know if you have any questions.
Russell
Applications Engineer
National Instruments
http://www.ni.com/support