How does the autopilot work?

[john_morris_uk]

Maybe for you, but many of our cruises involved long legs without a cross tide. From Essex to the West Country, or around the Southern Baltic were our hunting-grounds for many years. There is no law against following a track across the Channel. It’s just a lousy way of doing it.

I agree. Just pointing out what’s not intuitive to many people.

 

[Ric]

OK so I assumed I knew how the autopilot worked, as I'm sure do we all. Then I set one up and now I'm reading about great circles. Now I'm in one of those loops we all get into occasionally...

So I select a destination, for instance the Carribean, and I'm in the UK, and the plotter knows the great circle. How does it do the steering? I thought I knew that the plotter told the autopilot where the waypoint was and the AP steered to it. Then I realised probably not, and maybe the plotter just sends a course to steer, but then I realised probably not since I have to acknowledge the waypoint on the AP, not the course and with a GC that would need to be updated. So....does it just send out a heading constantly, or does it send an initial course and the boat would end up on a rhumb line?

All theoretical of course, given all the other considerations on a yacht like weather I can't imagine I'd ever use a great circle for long enough for it to make a difference, and certainly not with an electronic AP!

I would be very surprised if most APs used anything other than great circle. Aircraft autopilots definitely use GC so it would be odd if the designers of marine autopilots were to do anything different. The software then just steers a course to minimise cross track error along the GC.

 

[dunedin]

I would be very surprised if most APs used anything other than great circle. Aircraft autopilots definitely use GC so it would be odd if the designers of marine autopilots were to do anything different. The software then just steers a course to minimise cross track error along the GC.

Big difference is that for aircraft the journeys are regularly hundreds or thousands of miles, often with the ability to fly there directly. And aircraft have engines and a relatively narrow optimum cruise speed.
The vast majority of boat autopilot routes are for much shorter distances between harbours and / or navigation obstructions.

Also, as noted, on a saily boat, even if travelling thousands of miles you generally want to route for optimum weather / wind angles and or tide / current. A fixed route could be days longer and potentially more dangerous than a weather routed course.

 

[john_morris_uk]

Big difference is that for aircraft the journeys are regularly hundreds or thousands of miles, often with the ability to fly there directly. And aircraft have engines and a relatively narrow optimum cruise speed.
The vast majority of boat autopilot routes are for much shorter distances between harbours and / or navigation obstructions.

Also, as noted, on a saily boat, even if travelling thousands of miles you generally want to route for optimum weather / wind angles and or tide / current. A fixed route could be days longer and potentially more dangerous than a weather routed course.

I’m sure about ‘hundreds of thousands’ in one go. Tens possibly, but all the long haul flights I’ve been on recently where you can see the routing on your seat display show that the routing is rarely in a straight line. They’re either choosing which airspace to fly in or routing for winds of to avoid weather or turbulence.

 

[AntarcticPilot]

It sure about hundreds of thousands in one go. Tens possibly, but all the long haul flights I’ve been on recently where you can see the routing on your seat display show that the routing is rarely in a straight line. They’re either choosing which airspace to fly in or routing for winds of to avoid weather or turbulence.

Yes; I prefer to watch the flight track than most of the other "entertainment" offered! And I fly to Hong Kong about once a year on average. It's very noticeable that although the flight path approximates the Great Circle route, it deviates from it quite substantially in places, especially over mainland China where it very definitely DOESN'T follow a Great Circle! But even where it could, they often modify it to use the jet-streams. Over that distance, it makes a substantial difference if they can pick up a tail-wind; the flight out is usually 2 hours longer than the flight home for that reason (I think that's the right way round; given the time changes etc. I can never remember which way it is!). Weather routing is important even for airliners.

 

[James_Calvert]

Our Autohelm 2000 is called George.

Just googled some old manuals.

Neco named theirs Pilots but autopilots in the text. Cetrek and Wagner similar. The proper name appears to be automatic pilots.

A firm I worked for in the late 70s used to sell and install these sort of things in fishing boats, motor yachts and the like.

Very expensive compared to the yachty stuff just arriving. Seafarer did a tiller item called a Seacourse. Space Age Electronics had one that could be integrated with wheel steering, there was something called a Tiller mate, and Nautech introduced their Autohelm - which was definitely a trade name.

 

[ip485]

Yes; I prefer to watch the flight track than most of the other "entertainment" offered! And I fly to Hong Kong about once a year on average. It's very noticeable that although the flight path approximates the Great Circle route, it deviates from it quite substantially in places, especially over mainland China where it very definitely DOESN'T follow a Great Circle! But even where it could, they often modify it to use the jet-streams. Over that distance, it makes a substantial difference if they can pick up a tail-wind; the flight out is usually 2 hours longer than the flight home for that reason (I think that's the right way round; given the time changes etc. I can never remember which way it is!). Weather routing is important even for airliners.

Absolutely correct, direct are almost never flown.

Route planning consists of a surpising plethora of factors, including weather (both taking advantage of jet streams and avoiding the nasty stuff), NOTAMS (areas which for various reasons are restricted or unavailable), and airways (yes, most of the time in most countries aircraft are still constrained to set routes, which is hugely wasteful, but still the way the system is designed). I cant speak for intercontinental stuff (well at least not from direct experience) but across Europe routes are entirely based on whatever you can get Eurocontrol to accept as a series of predefined wps that happen to basically get you from point A to point B. If you get lucky you may find air traffic will agree to you routing direct between certain wps on the route, missing out some others, to give you a far more direct route, but your route will have no resemblance to a point ot point between A and B. In these days of GPS we still live in strange times!

 

[pyrojames]

Some of the DIY vane steering gears were known to be very quick upwind compared to human helms.
I wonder if 'production' gears sacrificed performance for robustness?
Or whether the DIYers in the 60s tuned their gears to the boats, maybe unknowingly, experimenting a lot.

I have an AYRS book on self steering. From the 60s.There is basically no maths in it, which surprises me.
It's a feedback control system, lots of people in the 60s worked with the maths of all that. Proper maths, calculus and funny symbols, none of your 'beat it to death with a spreadsheet' crap.

I've read that, I did my undergraduate thesis on windvane steering gears. In addition to the wind and water testing, there was a lot more control theory than I had wanted...

 

[pyrojames]

I know that my Garmin calculates the bearing and distance to a waypoint via the great circle route. I put in the Scillies after I cleared the Grand Banks off Nova Scotia and the course was definitely not the rhumb line. At that stage the AP was not steering to a waypoint though and while our course was nominally a great circle one, we chopped and changed to keep the wind aft.

 

[Ric]

Big difference is that for aircraft the journeys are regularly hundreds or thousands of miles, often with the ability to fly there directly. And aircraft have engines and a relatively narrow optimum cruise speed.
The vast majority of boat autopilot routes are for much shorter distances between harbours and / or navigation obstructions.

Also, as noted, on a saily boat, even if travelling thousands of miles you generally want to route for optimum weather / wind angles and or tide / current. A fixed route could be days longer and potentially more dangerous than a weather routed course.

The actual route sailed (or flown) is different from the great circle route. But put any two points into either a marine or aviation gps then the route calculated between them is great circle route based on WGS84. It would actually be more difficult for computers to calculate a rhumb line.

Obviously in both aviation and sailing you deviate from the initially planned route for many different reasons.

 

[TernVI]

... It would actually be more difficult for computers to calculate a rhumb line.
..

I doubt that.
Isn't it just arcTan (lattitude difference/( (longitude difference).(cos lattitude) ) ) plus some stuff to deal with the quadrant you're in?
To the nearest degree assuming you're not going hundreds of miles.

 

[mjcoon]

I doubt that.
Isn't it just arcTan (lattitude difference/( (longitude difference).(cos lattitude) ) ) plus some stuff to deal with the quadrant you're in?
To the nearest degree assuming you're not going hundreds of miles.

Any such formula is only going to yield a single value, not a "line"...

 

[Buck Turgidson]

Absolutely correct, direct are almost never flown.

Route planning consists of a surpising plethora of factors, including weather (both taking advantage of jet streams and avoiding the nasty stuff), NOTAMS (areas which for various reasons are restricted or unavailable), and airways (yes, most of the time in most countries aircraft are still constrained to set routes, which is hugely wasteful, but still the way the system is designed). I cant speak for intercontinental stuff (well at least not from direct experience) but across Europe routes are entirely based on whatever you can get Eurocontrol to accept as a series of predefined wps that happen to basically get you from point A to point B. If you get lucky you may find air traffic will agree to you routing direct between certain wps on the route, missing out some others, to give you a far more direct route, but your route will have no resemblance to a point ot point between A and B. In these days of GPS we still live in strange times!

Long range can be a little more free depending on where you are but it's usually GC modified by wind then fitted to available airways or tracks. The only time I got to fly completely unrestricted routes was up and down the South Atlantic from ASI to MPN. Just like the Vendee skippers we had to play the winds on that route

 

[TernVI]

Any such formula is only going to yield a single value, not a "line"...

That's what a rhumb line is. A constant heading relative to grid North.
That is kind of the point of it.

Obviously it will get silly when you pass through many degrees of lattitude or close to the poles.
Most rhumb lines i.e. if it's not 0, 90, 180, 270, are spirals from pole to pole, if extended.

But for those of us only going the odd 500 miles here or there and using a flat chart instead of a globe, it generally suffices.

 

[[163233]]

A great circle would be drawn as a curved line on a mercator projection, so if you're not seeing a curved line and it's a mercator projected chart then it's not a great circle.

 

[lustyd]

Well that depends how the route is drawn. For power consumption it’s probably better to render a straight line between two points. That’s not related to the routing necessarily

 

[mjcoon]

A great circle would be drawn as a curved line on a mercator projection, so if you're not seeing a curved line and it's a mercator projected chart then it's not a great circle.

With the obvious (I hope!) exception of N-S lines!

(P.S. I have a vague recollection that I may be thinking of transverse Mercator, which I think that is the one usually used. And only the central meridian is the one that is accurately aligned.)