Tim Kerby wrote:

>Multipath should not be too much of a problem if the beam is collimated
>and the first ping is the one taken count of.

You are right Tim, but Sean was looking at a different scheme which had
the problem I posted.

The scheme was to zero a counter and then send a ping.  As soon as an
echo was detected, the counter was incremented and another ping was
immediately sent -- then another, and another...

Then after a set amount of time the number of pings would be counted to
determine the distance from an object.  The counter would then be zeroed
and the loop would begin again.  Thus, my caveat about multiple echo
returns.

From what little experience I had, the surest method to measure to the
closest object is to send the ping and measure the time for the first
return, as you suggested.  It is then wise to wait to send the next ping
until no echo could possibly return and be detected.  We did some trial
and error to find the appropriate time, but found that a large metal door
would return a detectable echo a great distances in a concrete corridor.
 We shortened the time between pings by zapping the AGC gain back down to
1:1 which killed any distant echoes.

Another idea we were going to try was to send out 3 pings, then we would
send out 4 pings, then 5, then 3, then 4, etc.  We were going to count
the number of return pings to see if we were looking at the current ping
or a previous one.  Unfortunately, I resigned before the idea was carried
out, so I don't know how practical it is.

Michael