Arduino based anchor watch

[noelex 77]

Sounds like a great project.

Unfortunately, the anchor position will not be correct. The anchor alarm needs to record the position of the anchor not the position of the GPS antenna when the anchor is dropped.

If the GPS antenna is near the stern, as it is on most boats, the error is significant. As the boat moves around the swing circle the apparent distance will vary by 2x this error, making it difficult to be sure if the anchor has not moved.

There are various solutions to this problem. The best is to use a heading sensor to automatically displace the centre of the alarm radius. This approach is complex and would require communication with the NMEA 2000 network. A simplier solution such as the facility to manually move the centre of the alarm radius may be preferable, at least at first.

[Quebramar]

Originally Posted by noelex 77

Sounds like a great project.

Unfortunately, the anchor position will not be correct. The anchor alarm needs to record the position of the anchor not the position of the GPS antenna when the anchor is dropped.

If the GPS antenna is near the stern, as it is on most boats, the error is significant. As the boat moves around the swing circle the apparent distance will vary by 2x this error, making it difficult to be sure if the anchor has not moved.

There are various solutions to this problem. The best is to use a heading sensor to automatically displace the centre of the alarm radius. This approach is complex and would require communication with the NMEA 2000 network. A simplier solution such as the facility to manually move the centre of the alarm radius may be preferable, at least at first.

Would the following idea reduce the error whilst avoiding operating complexity?
Pressing the button would generate recording the position twice at a certain time interval (immediately and 15-20 sec. later maybe) so that the movement of the boat/heading is captured. If one assumes that this heading would be going away from the anchor point, then one could hardcode subtracting from the first position the distance bow-gps antenna in the right direction based on the heading info.
Although not 100% accurate it could substantially reduce the double error mentioned by Noelex

[Anders]

An adjustable offset from where the GPS antenna is to the bow would be a great feature.

[goboatingnow]

Yes I should have mentioned that the setup includes a correction for the gps antenna offset, this offset is used to adjust the anchor point

Correcting the boats position for the antenna is a little more complex as you really need heading , again I have a sat compass on the 2K but it’s not my intention to require the whole nav system to be powered up

The gps antenna is under the deck approx mid ships

[noelex 77]

Originally Posted by goboatingnow

Yes I should have mentioned that the setup includes a correction for the gps antenna offset, this offset is used to adjust the anchor point

Without heading information you cannot adjust the anchor point. You can adjust the anchor radius by the distance between the bow and the GPS antenna, but this does not fix the error I am referring too.

Without correcting this error the anchor alarm will not work satisfactorily if the boat moves around the swing circle. You will end up with false alarms or have to select a large alarm radius, which decreases the sensitivity significantly.

You need a means of at least manually adjusting the drop point that will be recorded when you push the button (unless the GPS antenna is at the bow). Most of the phone apps will do this.

[goboatingnow]

I really don’t want to have to do anything other then push a button

So I’ll investigate adding a heading sensor. Flux gate modules are cheap

[noelex 77]

Originally Posted by goboatingnow

So I’ll investigate adding a heading sensor. Flux gate modules are cheap

That is ideal.

As well as automatically setting the correct anchor drop position eliminating the large error I am referring too, incorporating the heading information also gives more accurate real time information on the anchor position. This secondary improvement in accuracy is only small (typically only a few metres) so it is not essential, but it does help pick more rapidly and accurately if the anchor is moving slightly backwards in the gusts.

These small anchor movements often precede a full blown anchor drag.

It is also valuable while monitoring the anchor to be aware of the boat's heading. When dragging, most sailboats no longer point into the wind. The bow will fall off typically 30° or more and stay there, providing a second clue that the anchor is not holding.

Thus an anchor alarm that correctly incorporates heading information will place the centre of the alarm radius in the correct position without manual intervention, correct the GPS position for the true anchor to boat distance and also show if the boat has adopted the typical dragging attitude.

It is a pity more equipment manufacturers and software developers do not offer this option. Vesper is the only company that I am aware of. This makes the Vesper anchor alarm significantly better, although frustratingly while this feature worked perfectly on my Vesper Vision, the updated Vision 2 loses heading information over a 180° arc for unknown reasons.

[goboatingnow]

Thanks. It sounds like a good idea all round. Easy to add to the Arduino

I understand the issue in correcting the anchor drop point with heading but I don’t see the gain in having heading when it swings ( leaving aside the bow falling off alarm )

I mean once I set the max radius the boat must swing within that anything outside that is a drag in reality I don’t see heading helping here. It doesn’t really matter which part of the boat goes outside the radius.

For example here in the Ionian we can get very calm nights , boats end up everywhere within the swing radius ( in fact this can lead to bumps ) it’s not a drag if the bow falls off once I’m within the max radius.

Perhaps you could explain more

[noelex 77]

When a gust hits the boat it will yaw and thus is no longer pointing directly at the anchor. If we assume the anchor does not move and the rode rode does not stretch the distance from the bow roller to the anchor will not change.

However, the distance between the stern of the boat and the anchor will change. If you draw a diagram the effect is easier to visualise.

The effect is only minor and in practice there will be other variables such as the stretch in the snubber and the inaccuracies inherent in the GPS position, but plotting the position of the bow rather than the stern (or other location on the boat) removes one variable that erroneously indicates the distance between the centre of the anchor alarm radius and the GPS antenna is altering

With heading information incorporated into the anchor alarm the position of the bow, where the rode is attached, can be measured rather than the position of the actual GPS antenna, which is often some distance from the bow.

Eleminating this minor error is much less important than preventing the much larger error that is induced if the alarm radius is centred on the position of the GPS antenna when the anchor was dropped, but incorporating heading information will automatically correct both problems.

[goboatingnow]

Ok thanks
I agree the heading info to offset the initial drop position error is valuable

The yawing isn’t a problem. In the deep anchorages here in Greece there’s considerable springing back forth on say 40-60 metres of chain the boat often drifts up to the anchor point , a gust and she’ll go yaw and drift near broadside ( almost as if she was dragging ) then her head will come up and she’ll lie head to wind.

Hence alarms on yaw itself would be too spurious.

Given that the alarm radius will be greater then the setting distance , I don’t see the issue of the gps antenna position having any real significance in determining dragging. ( in effect a fixed offset to the radius would mostly take carr of that issue.

I mean do I care if the bow or the stern goes over the radius !

Having said that excessive yawing would be useful to detect

[goboatingnow]

Thinking about it. It may be better to add the canbus interface and nmea2000 decoding. Heading information would be available during the anchoring process as at that time all my electronics are on

Later the lack of heading won’t make that much of a difference

[StoneCrab]

Don't let the naysayers discourage you. Especially those who say that there are other available options like on a smart phone. We all learn from new attempts, that is how things evolve.

This could be a great project for a parent to pass on to a child at junior high age. (after you've done the hard work) Arduinos are fun and easy to configure. Nothing like a practical application to get a kid's juices flowing... or mine for that matter.

My son and I have done some low energy projects for ocean drifters. We started with an Arduino but the power consumption was too high. There are other processors available with lower power consumption. We worked out the application on the Arduino and then went shopping for the best platform. The drifter had to be solar powered with a limited footprint. You wouldn't have the same limitations but since you mentioned power consumption, I wanted to comment on that.

[Jack C]

Depending on your boat length, the location of the GPS antenna becomes important. Its important to mark the anchor position. Marking the anchor location 30' back from its actual position means a position error of 60' when the boat swings 180 degrees? Jeff Siegal, the Active Captain developer wrote about this a few times.
I once anchored on 40' of chain in the crowded small anchorage at Cape May in a storm. Boat length 39'. Here marking the anchor location accurately was important.Had 2 boats on a shared anchor drag towards me because the occupants were asleep.
The were oblivious to horns. I had to motor over in my Portland Pudgy and bang on their hull to wake them up.


My primitive but effective anchor watch back then was a hand held GPS located with me in the V berth with waypoints set around my position and waypoint approach alarms set. I compensated for the anchor position error by using the rode length as an offset.
The track feature showed me the boat movement with the change in wind and current. I set an egg timer for 30 minutes.

[noelex 77]

Originally Posted by Jack C

Its important to mark the anchor position. Marking the anchor location 30' back from its actual position means a position error of 60' when the boat swings 180 degrees?

Exactly. This is responsible for many of the false positive alarms. Pity, when it is easy to fix. You don't need sophisticated integrated electronics with heading information (although this will give a slightly better result).

Good anchoring gear is the most important factor, but anchor alarms provide a valuable back up. Unfortunately, many cruisers set these alarms incorrectly.

[goboatingnow]

Originally Posted by StoneCrab

Don't let the naysayers discourage you. Especially those who say that there are other available options like on a smart phone. We all learn from new attempts, that is how things evolve.



This could be a great project for a parent to pass on to a child at junior high age. (after you've done the hard work) Arduinos are fun and easy to configure. Nothing like a practical application to get a kid's juices flowing... or mine for that matter.



My son and I have done some low energy projects for ocean drifters. We started with an Arduino but the power consumption was too high. There are other processors available with lower power consumption. We worked out the application on the Arduino and then went shopping for the best platform. The drifter had to be solar powered with a limited footprint. You wouldn't have the same limitations but since you mentioned power consumption, I wanted to comment on that.

Indeed. I’m a retired embedded systems engineer so I can move platforms ( STM32 , ATSAMD etc ) to suit the ultimate platform.