Spooky? Electronic Anchor Watch Confusion

[myocean]

How can this happen? Or - what has happened? What needs to be changed?

We stay at anchor with out 50 ft catamaran with a 35 kg Bugel-anchor. A wide sandy bay, 15-20 knots of wind, wind surfers around are enjoying it. We stayed here two nights already. During the night the wind calms down to 5 knots and less.

We do monitor the anchoring situation with our Watchmate AIS which features a very convenient anchor watch feature. Giving an alarm when a set maximum distance from the starting position is exceeded.
When we set the anchor it did not get a grip immediately. So the starting position was finally up to 68 meters away from us. About 25 meters of chain are out, the depths is about 5 meters. With less wind the boat moves towards the anchor. 68 or 69 meters have been the maximum distance displayed since two days, so we set the alarm to 70 meters.
After a nice morning on the beach with the kids we stay aboard for the afternoon and suddenly the alarm sounds. Curiously we watch the watchmate screen. The wind direction has changed a bit, max be 10-20 degree? The distance displayed goes up to 72 meters, 73, 75, 76, 80 and so on. We are moving!?! A clear track is visible on the graphical plot.
Time to react! We do not check the movement by taking additional bearings but start the engines. The ropes holding the anchor chain between the two bows seem to have less tension than maximum. It is well possible the the anchor is drifting.
We decide to take some tension from the chain using the engines. We do this for about 30 seconds and hope the anchor with get a grip again. Without the motor thrust we clearly feel the anchor holding us again. We try again and go forward and than backwards with the engines and feel a clear jerk stopping the boat. Nice!
Looking again on the AIS we see the distance being 119 meters now. So we drifted 50 meters, wow!
About 20 minutes later I look back at the AIS screen and can't believe what I see: The wind is still the same from South-East and we are now back at 68 meters from the starting position, just as nothing has happened. A comparison with the plow on OpenCPN confirms the track visible on the AIS screen: We moved away from the original position and back again.
Dos anybody understand this?
Was this a glitch in the GPS signals - so much? A problem with the AIS receiver (integrated in the AIS), Do I need a new anchor and need to get rid of any kind of sea monsters which occasionally move it back to it's original position?

We have tried to understand if we have really moved. Looking at the shore and two other boats in the bay we are not 100% sure but it is very well possible that we are where we have been all the time. I can imagine that people without an electronic anchor watch would not have noticed nothing.

What shall I do to be safe in future? I think the safest but not simplest solution is to always have a second anchor. But what else can be done?




Attached Photos:
-anchor watch screen after the engines were off again, new alarm distance set to 120 meters (circle)
-anchor watch screen some time later, track further away clearly visible
-Track on OpenCPN (using the same GPS signals, coming from the AIS)

[Vasco]

Next time take some bearings and plot them on a chart. Who can understand all that electronic stuff.

[noelex 77]

I am not used the wachtmate,but with 25m of chain out your swing radius should only be 25m + the distance to the GPS unit + a small margin for error.
Say 25+ 12+ 8 say 45m. This should be zeroed from where the anchor is set.

If I understand correctly your anchor dragged 40m before setting. An anchor that takes this long to set set is not reliable. A modern new generation anchor will set in less than a couple of meters. The older generation anchors take longer, but more than 10m and it should be picked up for a second try.

If you have 25m of scope out a reading of 119m is of no value.

[noelex 77]

Originally Posted by Vasco

Next time take some bearings and plot them on a chart. Who can understand all that electronic stuff.

If the wind/tide direction is constant transits are the most accurate. Even a few meters of movement backwards is apparent.

The advantage of GPS anchor alarm is that it always keeping watch day and night, good and poor visibility. It will also still work with a change a wind, or tide.

I have a GPS alarm on almost 365 days a year 24 hours a day. With a well set alarm the false + are almost nil. Until you pull up anchor and leave without cancelling it. Not that I ever do that

[bdbcat]

myocean...

What you have seen is normal GPS signal geometry dilution just before the GPS receiver chooses another set of satellites to compute a fix. Or external GPS interference from terrestrial sources. Or an invisible meteor that disturbed the ionosphere and scrambled the GPS signal propagation model.

In any of the above cases, there is little to be done in real life anchoring. I use 0.050 NM for my anchor alarm, about 100 metres. It signals about once a month in daily anchoring. Always at night

Good Luck, and thanks for using OpenCPN

Dave

[noelex 77]

Originally Posted by bdbcat

myocean...

What you have seen is normal GPS signal geometry dilution just before the GPS receiver chooses another set of satellites to compute a fix. Or external GPS interference from terrestrial sources. Or an invisible meteor that disturbed the ionosphere and scrambled the GPS signal propagation model.

In any of the above cases, there is little to be done in real life anchoring. I use 0.050 NM for my anchor alarm, about 100 metres. It signals about once a month in daily anchoring. Always at night

Good Luck, and thanks for using OpenCPN

Dave

0.05nm, or almost 100m is to coarse if you have 25m of rode out.

Mine is currently set for 0.04 Nm and I have 50m+ of chain out.

[Hudson Force]

I have had momentary false GPS positions received on a couple of occasions and one of these did give me a false anchor drag alarm....



In the image above, the line at the cursor was never transitted by my vessel. For a brief moment my position was falsely indicated at a point to the lower left of the screen and then immediately back to another false point at the acute angle to the upper right on the same line as the cursor and then, all within a moment, back to the area of the boat icon.

[noelex 77]

Originally Posted by CaptForce

I have had momentary false GPS positions received on a couple of occasions and one of these did give me a false anchor drag alarm....



In the image above, the line at the cursor was never transitted by my vessel. For a brief moment my position was falsely indicated at a point to the lower left of the screen and then immediately back to another false point at the acute angle to the upper right on the same line as the cursor and then, all within a moment, back to the area of the boat icon.

Gps has got more accurate over the last few years. If you mount a modern GPS above deck, with a good view of the horizon the instance of false positions that would trigger a correctly set anchor alarm should be very rare.
I almost never cruse built up areas, but there are instances of GPS jammers occasionally causing problems.

This is example of the semi-circular trace (over a couple of days) showing the accuracy that is possible in this case with a superior DGPS signal. Even a unrealistic, and inappropriate 4-5m drag alarm would not have been triggered.

I almost never get a false +, but my anchor alarm will alert me to less than a boat length drag. It works constantly to alert me of a problem.

If you anchor in front of me it would be great if I could convince you to do the same

[Vasco]

Originally Posted by noelex 77

If the wind/tide direction is constant transits are the most accurate. Even a few meters of movement backwards is apparent.

:

Yes a transit is the easiest way. I always do it when I anchor and then again when the tide changes if it's tidal.

[Hudson Force]

Originally Posted by noelex 77

......... I almost never get a false + .........................

Sure, I agree, but you used the word "almost". The event that I cited is one of three that I have had in the past eleven years. Still, if I'm awakened by my anchor alarm I check to see if this is the fourth!

[barnakiel]

Normal behaviour. You also get a false alarm immediately after your GPS loses grip and then re-sets.

It is either hardware related or else something related to the constellation, can't say. But we did see the same results before with our GPS handhelds set for anchor watch or simply left on while at anchor.

PS We use depth alarm, when applicable, just to make sure. Maybe using a different GPS unit for OpenCPN will help you get the record right in the future.

b.

[S/V Alchemy]

I have seen enough screwiness with GPS data (hey, honey, we only have the Number 3 up and we're doing 62 knots SOG!) over the years, if only intermittently, that I would widen my anchor watch circle a bit and take bearings as has been suggested. Basic pilotage is not always appreciated these days, but KNOWING by eye when the electronics are plainly in error is a comfort and a necessity, given that a drama-filled anchor haul and a panicky repositioning based on a bad fix could cause you more grief in a crowded anchorage than just a simple visual check that the boat hasn't veered out of line with a landmark.

Add to that the idea that in five metres, you can free-dive to see if the anchor is nicely set (particularly if you are hanging around on the hook for a few days), and the anchor watch feature, while useful, is put in proper perspective.

[jeffrobbins]

As mentioned above... when the constellation changes, and more or less satellites come into view, the positional error increases. When this happens, a GPS is often able to reduce the error. But the scenario described shows how a GPS responds to significant errors as it can take time for it's correction algorithms.

The WatchMate anchor watch algorithm attempts to further limit the effect of transient GPS errors but this must be balanced with the need to be sensitive to actual anchor drag situations. Errors on the side of false-positive are obviously favored.

The amount of error is typically related to the number and location of the satellites which are being received. For example, if the satellite constellation changes such that they are not as well positioned in the sky, the errors will increase. Or if your GPS antenna is blocked from a particular direction it may work fine until some or most of the visible satellites are in the blind spot.

Have a look at the GPS-Status page and see what the HDOP is. That number is essentially an indicator of the expected horizontal error and the lower the better. If you are seeing a high HDOP it may be that your GPS antenna (either internal or external) doesn't have an unobstructed view of the sky.

Finally, sometimes corrections are made to the satellites themselves and this can have a temporary effect. I don't have any information to indicate that is what did, or did not, happen in your case, but for an example see HDOP Advisory (Sept 11, 2009)

[Viking Sailor]

Also, it looks like we are in for some rough space weather...

Space weather may reduce GPS accuracy — Noonsite

NOAA / NWS Space Weather Prediction Center

[S/V Alchemy]

Originally Posted by jeffrobbins

As mentioned above... when the constellation changes, and more or less satellites come into view, the positional error increases. When this happens, a GPS is often able to reduce the error. But the scenario described shows how a GPS responds to significant errors as it can take time for it's correction algorithms.

The WatchMate anchor watch algorithm attempts to further limit the effect of transient GPS errors but this must be balanced with the need to be sensitive to actual anchor drag situations. Errors on the side of false-positive are obviously favored.

The amount of error is typically related to the number and location of the satellites which are being received. For example, if the satellite constellation changes such that they are not as well positioned in the sky, the errors will increase. Or if your GPS antenna is blocked from a particular direction it may work fine until some or most of the visible satellites are in the blind spot.

Have a look at the GPS-Status page and see what the HDOP is. That number is essentially an indicator of the expected horizontal error and the lower the better. If you are seeing a high HDOP it may be that your GPS antenna (either internal or external) doesn't have an unobstructed view of the sky.

Finally, sometimes corrections are made to the satellites themselves and this can have a temporary effect. I don't have any information to indicate that is what did, or did not, happen in your case, but for an example see HDOP Advisory (Sept 11, 2009)

The contrast between you and Craig Smith is amazing. Both of you advocate (or once did!) on behalf of your product, but only one of you seems to anticipate the sort of questions an actual customer might ask.