| ★ wanayoo — archive 1999 http://geek.com/readercomm/081999/comm94.htm | Nouvelle recherche | Portail wanayoo |
|
|
|
|
||
 |
|||||||
 |
Reader Comments | ||||||
|
Resources
|
 |
Thoughts on GPS reset problems COMMENT From: Nick The GPS satellites were placed in service by the US military for the US military. The US military can turn them off and/or modify their operation anytime they want. Anyone that spends money on a GPS system should not assume they have any rights (as far as actual satellite operations go) to have a system that works for them. From: Bill T My Garmin GPS38 has the lock-on problem. ROM version 3.04. However Garmin support says it will lock-on if initiate is run many times?? They have a downloadable fix. ' My new Garmin GPSIII works fine. Great web site. Thanks. From: Jeremiah B I was not affected by the GPS Week Number Rollover (WNR). I am told by the factory that my GPS (Eagle brand) is both WNR and Y2K compliant, like most others currently available. I can answer the question as to why only count 1024 weeks. When the satellites send a signal, it has the date and time built into it. The handheld unit then calculates how long it took for the signal to travel from the satellite to the handheld receiver (in nanoseconds, because the satellites are only 9000 nautical miles away), and therefor your distance to the satellite, because the speed of light is known. Using distances to many satellites, it figures out where you are. The milliary wanted to keep the signal short, so that it does not take long for the hand held unit to receive the signal, and lock on to your position. Another reason to keep the signal short, was so that it did not take long for it to be sent, introducing more error. ex. It sends a signal at 8/20/99 18.20:12.123495847 UT but it takes an extra .0000012 seconds for the signal to actually be sent,because it is too long, introducing an error of .2232 miles (light travels at about 186000 mi/sec) an error this big is unacceptable). So, it only counts 1024 weeks, to keep the signal short, so that the whole process works fast, and no unreasonable error is introduced. UGEEK
RESPONSE Another reader sent in a response suggesting that the the 1024 limit was due to old 8-bit chips, and related to 2^10 in some manner. However, 2^8 is 256. There certainly could be some sort of limit in chip bits that explains such a theory, and I don't discount it. As for the last reason, that sounds reasonable as well. Perhaps it's a combination of both. I'd guess that anything over the 1024 limit was too slow using old silicon, as the last reader suggests. Today, chips are faster, and the slowdown of calculating larger numbers is next to nothing. Close enough to nothing that errors are much smaller, at least.
Home | Welcome
| Survey | Newsletters
| Site Map | Contact
Us |
|||||
