Comment by dis-sys

yes, 1m for every 3.3ns error, we all know that.

the problem is those GPS vendors only provide long term average error for their products, there is upper bound guarantee at all.