Too much deviation

[RedHerring]

I have a problem with my Raymarine autopilot compass unit. It has too much deviation - up to 17 degrees. Two attempts at doing the full circle routine resulted in roughly the same outcome.

The boat (Ericson 35) has a fairly high bow. So, one possible explanation is that the wind is doing it. Only thing is, the last time I tried doing the full circle adjustment routine, the wind was in the "barely noticeable breath" category.

There is a shower sump pump installed about 60 cm (2') away from the sensor. That could be another culprit.

And also, I don't know if this sensor can be itself faulty in this peculiar way.

What would you do to solve this riddle?

[belizesailor]

I would check for sources of deviation first...like that sump pump. Address this issue and then re-swing.

Also to be near a shower sump pump it must be pretty low. In the bilge? Near the keel/mast step?

Wind is not a likely culprit as the software will tell you if you are turning to fast.

[model 10]

Seems like the shower pump would only have an effect if it was running. How about unscrewing it and setting it out in the open? You do have it mounted near the centerline?

[RedHerring]

> Wind is not a likely culprit as the software will tell you if you are turning to fast.

I'm theorizing that the wind may be making the boat turn faster going from upwind to downwind than the other way around. Basically, wishful thinking that I can solve it without having to dismantle anything

The sensor is placed under a settee in the main cabin, between a table and a forward bulkhead. The latter is also the compression post for a deck-stepped mast. The keel is lead-ballasted.

> You do have it mounted near the centerline?
Yup. That was obviously how the placement was chosen - away from the diesel, near the centerline.

[belizesailor]

Originally Posted by RedHerring

> Wind is not a likely culprit as the software will tell you if you are turning to fast.

I'm theorizing that the wind may be making the boat turn faster going from upwind to downwind than the other way around. Basically, wishful thinking that I can solve it without having to dismantle anything

...

Wind will effect turn rate of course, but as long as you are within acceptable turn rate its OK.

So, yeah, wishful thinking.

[Wotname]

Do you have DC wiring anywhere near the sensor?

[RedHerring]

> DC wiring anywhere near the sensor?

Windlass power wires live in the bilge about a foot under under the sensor. No current in them, of course.

[Wotname]

Originally Posted by RedHerring

> DC wiring anywhere near the sensor?

Windlass power wires live in the bilge about a foot under under the sensor. No current in them, of course.

No current = no compass interference from them

[donradcliffe]

Is the deviation the same on all headings??

[Cap Erict3]

You autopilot might be showing COG.


------------------------------
Looking for another pretty place to work on the boat.

[David M]

In a nutshell, three things cause deviation, non-magnetized iron, magnetized iron and magnetic fields caused by DC currents. Find those things, move them or move your compass. You may need a repeater compass to read from if you find it necessary to move your auto pilot compass further away from things that have a magnetic influence. I had to put my autopilot compass near the bow of the boat where it was furthest away from two large Diesel engines.

[Wotname]

To elaborate on David M's excellent advice, most steels are magnetic except for Austenitic stainless steels like 300 series.

So using 316 hardware around your sensor is acceptable.

[svlamorocha]

Originally Posted by RedHerring

I have a problem with my Raymarine autopilot compass unit. It has too much deviation - up to 17 degrees. Two attempts at doing the full circle routine resulted in roughly the same
(Snip)
What would you do to solve this riddle?

Can you please provide a table showing deviation (between Raymarine and handheld) as a function of (handheld) compass course? Ideally the table would be made with 16 data points spaced 22.5 degrees or so.

My starting hypothesis is that you have calculated deviation against a compass that has much deviation. Best way to avoid that is to use as baseline not the steering compass but a handheld compass held in the centerplane of the boat but far away from any metal or wires.

C

Sent from my GT-I9192 using Cruisers Sailing Forum mobile app

[jackdale]

Originally Posted by Cap Erict3

You autopilot might be showing COG.


------------------------------
Looking for another pretty place to work on the boat.

Or the fluxgate compass is set for magnetic.

Is the deviation east or west?

[RedHerring]

I don't know what my main compass deviation looks like. I should certainly figure that out, but haven't seen anything drastic so far.

Raymarine's automatic deviation adjustment doesn't involve using any other compass, and consists of two steps.

1. A steady slow 360 degree turn. The program assumes a linear change of heading, and detects relative deviations from that assumption.

2. You then point the bow along some known transit and set the unit to display this known heading. Thus turning those relative deviations into absolute ones.

My autopilot complains about deviation being up to 17 degrees right after step 1. I have it set to display M headings. It matches the main compass on a some headings, and mismatches on some others.

My current plan:

* live with it for now - not a safety issue, really,
* make a deviation table for the main compass,
* do another autopilot adjustment circulation, this time making sure that the actual heading (taken from the main compass sans its own deviation) is actually changing linearly,
* if it fails to produce decent results again, then move the autopilot sensor to some other place and see if that changes anything

And then I'll start moving things around.