TomT said:
And there is absolutely no advantage to more than 5 satellites in terms of accuracy...
True, which is why I said a 12-channel, not a 50 or 60 channel. Generally, the more channels the receiver has, the faster your cold-start time.
My estimation was based somewhat on the
Copernicus II GPS module.
Actually, the time accuracy is at best 14ns. Still pretty damn good.
I like 30, lets use it. 30MPH speed, 30seconds sample, 30ft for accuracy, 30 nanoseconds accuracy. Sound fair?
30MPH for 30 seconds is 1,320 feet traveled. 30 seconds * 5280 feet per mile / 60 minutes per hour / 60 seconds per minute * 30 seconds. Or you could just know it's exactly 1/4 mile.
With a 30-foot accuracy error, this means the stop and start can be a total of 60 feet off. This means distance can be anywhere from 1260 to 1380 feet. Pretty big difference for the sample, but lets roll with it.
30ns start and 30ns stop come to a total of 60ns. This is a range of 29,999,999,940 to 30,000,000,060 nanoseconds. This error is only 0.0000002% of the example. This is pretty much negligible.
So, taking the worst case scenarios, we have 1260 feet traveled in 29.999999994 seconds, and 1380 feet in 30.000000006 seconds.
The first scenario comes out to 28.63636364209091 MPH
The second scenario comes out to 31.36363635736364 MPH.
Obviously, this isn't as accurate as I first thought. 30 seconds is also a pretty short sample time. Unfortunately, longer times might actually increase inaccuracy due to conditions such as elevation change (even inches will add up), wind resistance, even subtle engine power differences. Even cruise control won't keep a perfectly stable speed for minutes.