GPS WAAS Antenna Location : Masthead or Sternrail?

[boden36]

Ours is steel boat, on sternrail off centre to stb. No problems. Inside in a steel boat is a no go.
On top of the mast violates the KISS principle.

Regards,
Richard.

[Wotname]

Originally Posted by Dockhead

GPS receivers (rarely antennae) don't use coax. They are connected to your network by either an NMEA0183 cable or to the N2K backbone.

No problem with that at the masthead, but a RPITA (that's a Royal PITA) to put it there and run cable, and to get to it in case you need to service it. A horrible place for it in my opinion.

I have mine enclosed within the instrument case above my scuttle, with nothing but a thin sheet of fiberglass and a windshield between it and the sky. Easy to get to, simple cabling since the network is already right there, and perfect reception. No need to go up to the masthead.

Clearly we are singing on different hymn sheets

My GPS receiver is connected to it's antenna via coax and a couple of GPS receivers I recently installed also required coax.

Presumably you are referring to a GPS receiver with an internal antenna and whose output is then networked to whatever other fancy pants () electronic equipment that is on-board.

Alas, apart from a small basic Furuno MFD, I don't have any other nice electronic nav equipment to play with

Yes, I concur with the difficulties of mast head antennas etc but I don't really believe (yet) the argument that it results in poor GPS performance.

[gbgreen59]

[QUOTE=boden36;1665189]Ours is steel boat, on sternrail off centre to stb. No problems. Inside in a steel boat is a no go.
On top of the mast violates the KISS principle.
QUOTE]

A good experiment to look into multipath is to record horizontal position data for 2 to 3 hours (over night would be really good) with the boat, ideally, sitting still (that is the hard part). Do this experiment away from other sailboats. Look at how much the position varied over the period that the data was collected. The horizontal position variation should be at 3 meters or less (no WAAS) at about 70% of the time (RMS/1 sigma numbers). Anyone can do this experiment. Also, a picture of the antenna installation would be good too. It would be interesting if many people did this experiment and reported their findings here.

Another thing to do is to set a handheld GPS to record its position. Then run it to the top of the mast using the main halyard. Let it sit there for a few hours the pull the data and plot. GPS data can be plotted on Google Earth. Compare the handheld position with the GPS position recorded from your current GPS antenna installation...this would be some really fun data for people to post. Will provide a lot of insight into GPS performance at various installations for everyone's benefit .

My boat is on the hard for the winter, so I can't help with these experiments until spring. But I'll run a few experiments next spring with the hand held in several locations around my boat to see how things vary. Proof is in the pudding, correct?...talk is cheap.

[gbgreen59]

I apologize for butting back in again...I am hopeful that the information is useful....

On another thread, it is mentioned that a particular receiver provider went from 12 channels to 66 channels recently. As there are usually no more than 12 GPS satellites visible at one time, why the extra channels? Well, it depends on the software written around those channels. But, the better receivers use those channels to fight multipath (amongst other things). The "primary" channels track the satellites in view (up to 12ish). Then the "alternate" channels continually search for multipath signals. Assuming the decision logic is good in the software, the extra channels will minimize the effects of multipath.

[DotDun]

Originally Posted by gbgreen59

Here is an example of DOP issue. My brother has his GPS antenna (built into the plotter) on the console of his bay fishing boat. With he and I standing in front of it (it is at about chest level), we were cruising down a canal between rows of houses on each side with houses on stilts (Galveston area). He showed me that the GPS had us passing along the houses and not in the narrow canal. Plus, he pretty much blamed the GPS being bad. But what was happening is that he and I were blocking a good portion of the sky and the houses were blocking most of the low elevation satellites. Hence, the DOP was poor and the position had a bias in it. I have other interesting examples of DOP issues also.

This may not be a DOP issue. I've witnessed several chart/map registration errors.

[gbgreen59]

I am very sure that this was a DOP issue. DOP is a parameter that is based on the number of visible satellites and how they are distributed around the sky. All of the low elevation satellites were being blocked, plus my brother and I were blocking low and high elevation satellites over about 1/3 of the sky. Low elevation satellites are very important to a good DOP.

Position error produced by a poor DOP is evidenced as a bias in the position solution. The chart would have been off by over 20 meters.

I have often wondered if some of the error attributed by charts could have been GPS antenna installation problems. Can you give more information on how the chart error was confirmed? How large is the error that you have observed in the charts? How did you measure the error in the charts? It would be interesting to explore that too.

I know that storms can change things and markers can move. But things that are fixed like reefs, harbor entrances, etc should remain the same. What is the error being observed in fixed objects on charts and how was the error measured?

[gbgreen59]

Here is another example of DOP being important to the GPS solution. I pack a handheld GPS in my shoulder pocket when I go snowboarding. Then, I plot the tracks on Google Earth when at home. The plot on Google Earth shows me getting off the lift, very accurately, when at the top of the mountain. However at the bottom where mountains block a lot of the sky, my entrance to the lift is scattered over 20is meters and is biased to the parking lot that is near the lift (never shows me at the lift). This wasn't a problem with Google Earth's accuracy of where the bottom lift is. It is an issue of poor DOP.

There are other examples of people asking in frustration about their GPS accuracy. I explore the circumstances with them, and find that the sky was being blocked by something based on what they were doing...such as mountain climbing, driving through downtown Chicago, walking in a deep ditch...etc.

[DotDun]

Originally Posted by gbgreen59

I have often wondered if some of the error attributed by charts could have been GPS antenna installation problems. Can you give more information on how the chart error was confirmed? How large is the error that you have observed in the charts? How did you measure the error in the charts? It would be interesting to explore that too.

I know that storms can change things and markers can move. But things that are fixed like reefs, harbor entrances, etc should remain the same. What is the error being observed in fixed objects on charts and how was the error measured?

I have (4) sets of chart/maps on my Furuno. (2) vector, (1) raster, and (1) satellite photos. In my slip on a 150' wide waterway, with the boat not moving and the GPS on for hours (reporting DOP <1), each chart will show the boat in a different location. The largest difference being one side of the waterway to the other, a difference of at least 100'. This is a chart/map registration issue, not a GPS location accuracy issue.

I believe map registration errors are more common with raster and sat photos (scanned images) than vector. But it's unclear to me how vector positioning data is determined when it comes to land based objects (like edges of waterways, large polygons, etc.).

[gbgreen59]

Here is another example of GPS in a high multipath environment.... I carried a handheld GPS with me when at Bush Gardens, Tampa several years back. I got out the GPS and stood next to a roller coaster that has big metal posts for structure. Plus, there are metal buildings scattered around. The "sky view" display showed that all visible satellites were being tracked at a very good signal level. Yet, the position solution was wandering around by up to 50 meters. This is characteristic of a high multipath environment.

[gbgreen59]

Originally Posted by DotDun

I have (4) sets of chart/maps on my Furuno. (2) vector, (1) raster, and (1) satellite photos. In my slip on a 150' wide waterway, with the boat not moving and the GPS on for hours (reporting DOP <1), each chart will show the boat in a different location. The largest difference being one side of the waterway to the other, a difference of at least 100'. This is a chart/map registration issue, not a GPS location accuracy issue.

This is excellent data, and a good assessment, thanks. Do you remember how much the GPS position varied over the two hours? It would be good to know that. Also, where is the GPS antenna located on the boat? Not that it changes your assessment, I am just curious.

[DotDun]

Originally Posted by gbgreen59

This is excellent data, and a good assessment, thanks. Do you remember how much the GPS position varied over the two hours? It would be good to know that. Also, where is the GPS antenna located on the boat? Not that it changes your assessment, I am just curious.

GPS is on the stern rail (actually dinghy davits). Location is slight urban canyon with house ~35' high within 50' on one side.

2 minutes after booting recorded position with 8 sats, 2 hours later the recorded position with 11 sats had moved 5 feet. 4 hours after start the recorded position with 11 sats had moved 6 feet from the original fix and 6 feet from the interim fix. Hence, the GPS was walking around within a 6 foot circle.

HDOP was .9 - 1.0 and VDOP was 1.7-2.1

FWIW, the most accurate chart is the sat photos, it shows the boat right where is should be. The next is Jeppesen C-MAP vector chart which is ~30' off. The worse are the NOAA 'free' charts, both raster and S57 vector are off by ~80'-100'.

Granted, there always have been and always will be accuracy issues with 'measuring' location (vs. surveyed). But, in my experience, the chart/map registration have more/larger errors than the GPS fix (given the GPS has reasonably good view of the sky). I haven't tested while purposely trying to mess with the RF (sealing a GPS in a metal garbage can). There are companies building receivers that have claimed success at <-150dBm levels.

[gbgreen59]

DotDun, I really appreciate this information. Sounds like an excellent experiment with a good baseline for making measurements. 6 feet of position variation is really good too. Obviously, the DOP was good and looks like multipath was minimal too.

[The Smokester]

Question: How is the DOP displayed by a GPS representative of the position accuracy? Does it represent something related to the standard deviation of time series sampled positions--in which case it may not include the biases from restricted satellite views mentioned above.

[DotDun]

Originally Posted by The Smokester

Question: How is the DOP displayed by a GPS representative of the position accuracy? Does it represent something related to the standard deviation of time series sampled positions--in which case it may not include the biases from restricted satellite views mentioned above.

The dilution of precision or DOPP figure is used to give a simple characterisation of the geometry of the satellites being used for a fix.

GPS Accuracy | Errors & Precision | Radio-Electronics.Com