Bookmark and Share Email this page Email Print this page Print

GPS Coverage can experience short gaps

Jan 1, 2003
From Ocean Navigator #114
May/June 2001
A letter in your March/April issue (?Voyager reports GPS and e-mail now widely used,? Issue. No. 112) relates the use of GPS-directed autopilots by sailors in the Caribbean. This is a dangerous procedure as occasionally, for several minutes at a time, the GPS signal is lost, and the autopilot has no direction. I had this happen to me twice before stopping the practice, the last time crossing at night from Trinidad to Grenada in 20- to 30-knot winds. We were sailing hard on the wind when the GPS signal was lost. The combination of this wind and rather heavy seas (hard to estimate at night) created a very dangerous situation and resulted in a broken genoa lead block, ripped main, and some nasty nighttime repairs. Now I set the pilot on a course indicated by the GPS, but I set the autopilot and correct it for leeway manually. Incidentally, I have lost the GPS signal at least once on every 24-hour or longer passage.

Arnold Grubin lives in Southern California and has been sailing the Caribbean for the last 15 years, currently in an Amel Super Maramu - based in Tortola in winter and Grenada in summer.

Contributing editor Bill Brogdon, former head of the Coast Guard's office of navigation, responds:

I have seen GPS receivers lose lock for a few minutes on a number of occasions. This may have been due to receiver problems or antenna shielding or something else rather than signal problems. In all cases I was using a relatively low-cost receiver.

I don't believe that using an external navigation system to direct the autopilot is an especially good idea. The reasons should be obvious, and one of them is the effect of short-term signal loss. Another is the zigzags from small signal errors. More important, you lose any way to measure set and drift. It's better to set a course in the autopilot and let 'er go. That way you can find the course and speed made good on a specific course. You also can tell autopilot zigzags from signal zigzags. But there is a strong theory: if it's technically feasible, it must be desirable. This theory is nearly immune to performance facts.

Edit Module