On Sep 2, 2009, at 4:54 PM, Kirkwood wrote:
> I was very careful NOT to mention accuracy in my response to Malcolm.
> Accuracy is a completely different issue. You not only have the wire
> to
> consider but the latency of the photo cell and the speed of light.
> If the
> lights are 1000ft apart (across the track) you have a 1000NS delay for
> arrival to the photo cell. We also don't know the precision of the
> surveyed
> mile. It was not my intent to open up another can or worms! My
> point was if
> you only have 0.001sec resolution you are introducing diminished
> precision
> when rounding to mph.
The 1000NS sec delay would only b e in the instance that the vehicle
crosses at maximum distance from the light-emitting beam, right? If
the vehicle was right by the receiver the delay would be much smaller,
I think. But realistic traps are narrower -- more like 100 feet,
taking a full order of magnitude off that number.
The precision of the surveying has always been a question in my mind
-- although today's high-end surveys are very accurate, I'm told by my
surveyor friend. But -- his new-end GPS surveys, which he reports are
within a half-inch over ten miles -- are within the half-inch error
when compared to old chain-type measuring. How precise was that?
My proposed system would not be affected by speed of light (other than
the delay in the photocell beam break) -- because each timing event
would include clock information -- so delay in arriving at the
computer is not of an issue. And wouldn't many of the errors be
consistent - and therefore subject to being tuned out of the system by
measurement and then correction in all computations?
Jon
_______________________________________________
Support Team.Net http://www.team.net/donate.html
Land-speed mailing list
http://autox.team.net/mailman/listinfo/land-speed
|