When an aircraft is put on autopilot, does it follow a constant heading such that it would spiral toward/away from the earth’s poles, or does it follow a great circle route? For short trips (e.g. DTW-ORD) I expect there wouldn’t be much difference, but if you’re trying to get from Chicago to Tokyo, tne difference matters.
Or are both modes available, and the pilot able to select between the two?
Mine can follow a GPS route with multiple legs.
Autopilots vary - they typically are appropriate to the aircraft in which they’re installed, and the kind of flying it’s expected to do.
IIRC, the problem with flight KAL 007 - shot down by the Soviets for intruding over Kamchatka airspace - was that the pilots entered “waypoint coordinates” into the autopilot or navigation computer and transposed some digits on one waypoint, which directed them on the wrong course. Was that the autopilot, or the autopilot was tied to a navigation computer? That was well before GPS.
There was a collision IIRC a decade or more ago in northern Ontario where the two planes collided head-on (clipped wings or something) as both were following the same GPS track - which shows how accurate that was. GPS navigation I assume is point to point so great circle. Since then the instruction for pilots is to always fly to the right(?) of a GPS track by 100 yards.
(Wasn’t there something similar between a commercial jet and an army jet or business jet in Brazil?)
On a related question: Most of the scenarios involving MH370 shows that it most likely was flying on autopilot west of Australia curving away from Australia. The arc displayed on USA television seems to have a long radius (over 1000 km) and centered somewhere near India. Since an autopilot is a computer, anything is possible, but why do they believe that it was flying in such a wide arc.
Autopilots can usually be set to follow a heading (e.g. 210°), or a “direct-to” (e.g. fly from here to Tokyo, which would be a great circle route).
As a sidenote, the inventory of the autopilot, Lawrence Sperry, is also generally considered to be one of the first two members of the Mile High Club. We only know this because he was riding with a married woman (not married to him) in one of his planes when the autopilot failed and the plane ditched in the water.
Both passengers were safely recovered, and claimed that their clothing had been lost in he crash.
Same guy helped to develop the Curtiss-Sperry Aerial Torpedo, which could have been the first cruise missile back in World War One, but they never quite got the autopilot and all of the other parts to work right all at the same time. The plane would have launched, flown on a set path, and after a certain distance (determined by rotations of the prop, IIRC), the wings would fall off, depositing the body of the plane (and its explosive payload) somewhere hopefully near the target.
Let’s just say there’s a good reason the Germans got credit for developing the first cruise missiles in WWII a few decades later.
It depends on the autopilot mode.
Heading mode (HDG) has the autopilot fly the heading you set with the HDG bug. It is typically used during departures and arrivals when you are getting micromanaged by the departure / approach radar controller. The headings are typically for short periods of time until you are either in a position to intercept your flight planned track outbound or intercept the final course of your approach to the runway. If left at a constant heading for a significant time the heading will stay constant and will therefore be a rhumb line that spirals towards the magnetic poles (the heading is typically referenced to magnetic north at most latitudes.)
There is also lateral navigation mode (LNAV). This will fly whatever the flight management computer has been setup with. The FMC will have your flight route made up of waypoints joined by great circle tracks. You can be in LNAV and be flying a great circle track between waypoints, or you could be flying a great circle track direct to a waypoint (the “from” waypoint simply being the position you were at when you selected “direct to”), or the FMC can also be set to steer a magnetic heading in which case it is similar to being in the HDG mode.
A further NAV mode will allow the autopilot to fly a constant radial from or to a ground based navigation beacon (VOR). A radial is a bearing from or to the beacon measured at the beacon rather than at the aircraft. Flying on NAV as opposed to LNAV is uncommon now days as the FMC is a much more capable navigation system than using raw ground based aids. I think some of the more modern airliners can’t even fly with direct reference to ground based aids, instead having to fly the FMC track which uses ground based aids as part of its computed position.
Most aeroplanes, including light piston engined singles and twins, with an autopilot and a GPS as a minimum will have some form of these basic lateral modes.
The arc shown in the predicted paths is based not on an autopilot setting, but rather calculated from a satellite ‘ping,’ which gave the angle from the satellite to the aircraft. They estimated it’s height, and that gives an arc along which the airplane could have been. If they had a bearing along with the angle and calculated height, they would have had a fairly good estimate of location at the time of the ping instead of the arc.
Anybody want to know about the old Lear L2 autopilot or how the aerial mappers made it able to make flat turns?
.
.
.
.
Thought not. :dubious: ::::: grump, getting old sucks. :::::
Does it just point the nose in a given direction, or will it compensate for the wind and actually fly a course in the given direction. Seems like the first case wouldn’t be very usedul (although the pilot could figure out the crosswind component, dial in a little crab, and fly the course he wanted), and the second case would require some sort of position fix like GPS (in order for the autopilot to know what course it’s on) in which case just holding a course seems less useful than flying to a waypoint.
What about autopilots before GPS, how did they handle this situation?
Most planes I’ve flown work like this:
Heading mode simply flies a heading. It doesn’t account for wind - that’s up to the pilot. In that mode the autopilot is not referencing any navigation information.
Nav mode means the autopilot is following some sort of guidance. It could be a VOR radial, a localizer, or GPS. It will also turn the plane in order to keep the course if it experiences wind.
GPS navigation is different from VOR in that it always (as far as I know) plots courses by great circle. The effect is negligible for short trips, but important for longer legs. I believe the Garmin 430 manual advises not flying direct longer than 200 miles for this reason. Rather, you should insert some intermediate waypoints. Plotting it out flat and direct will not give you the same course the GPS does, and this could cause an airspace incursion.
The airliner I now fly uses a combination of mechanical and computer inputs to the autopilot. Other planes do it purely mechanically. Interestingly, some older planes with mechanical autopilots have a hard time with GPS. They were designed for for VOR tolerances, which are much larger than typical GPS error values, which with WAAS enhancement can be as low as around three meters. So the older autopilots tend to snake around the course if they are taking GPS input.
I want to hear! What’s a flat turn?
Prior to GPS, laser gyros had begun to be used for inertial navigation. LORAN units were becoming available for GA. Both of these had about a decade run before GPS got feasible for non-military use. I never heard of LORAN driving an autopilot, but INS normally would.
Fiber-optic gyros were kind of still-born as GPS came online. They have even worse phase locking problems than the gas laser gyros.
A post upstream mis-stated ELMER Sperry’s first name.
My first job out of college was for Sperry. I worked on military flight controls, which interfaced to the autopilots.
Sperry had a very risk averse culture when I worked there. Honeywell was eating our lunch on laser gyros because their’s used only three mirrors instead of four, which reduced mirror losses, sealing, and alignment issues. I always thought the square optical path was emblematic of inside-the-box thinking that management rewarded.
Flat turn is a turn using rudder with opposite aileron to make the wings remain level with the ground as the rudder forces the turn.
Very uncoordinated turn but necessary for making certain kinds of products with a normal camera and also very useful when using a ‘strip’ camera.
It is done by adding a cheater knob to the rudder circuit and tweaking the aileron circuit very tight, all the way up to ‘twitchy’ with a cut out switch for going back to normal operation.
You could do the same thing using the Directional Gyro Bug ( heading selector ) but then it would go to max effort with no way to control the amount of rudder input. Had to be able to control all three axis with separate knobs. Sometime even had to use the pitch adjustment at the same time and tweak the trim circuit too. You never had enough hands. Had to keep the RADAR altimeter as close to 1950 feet as possible go keep the scale on the negative within specs.
Sometimes it was just easier to do it all by hand or if the autopilot lost all the smoke out of the wires, you just had to do it.
Don’t forget we had to keep the center line of the aircraft within 75 feet of the pipeline right of way or the line on the map depending on what we were doing. Before we had a view finder for the pilot there was a 3 inch hole in front of the pilots seat and a corresponding one on the belly, Inspection access that was not perfectly in like and a thick piece of glass that allowed me to see straight down sorta. I had to keep it in place with my heels, it would slip out of place in turbulence or any flat turn.
I still managed to smoke my Kool Filter Kings. Bawahahaha 
“Strip” camera = one that the film is pulled across the lens behind a narrow slit in synchronization with the ground image. No shutter as such. It just burned the film in a narrow strip when the film was stopped, it did not spread into the surrounding film because it was very very close to the slit.
We had at first one made with 70 MM film and later built another one that used 5" film. So we would end up with an exposure 5" wide and 250 ’ long if we did a run with no breaks. Hardly ever happened but we cut out a lot of breaks if the turn was not too sharp and I flew it just right. With the lenses we would get 1000. feet to the inch on the negative. Different lenses for the 700mm & 5 inch cameras. We need to stay at 1950 AGL ( Above Ground Level ) because the RADAR altimeter was not accurate enough higher than that & the accuracy was needed. Also easier for the pilot to stay online with the needed accuracy.
The cameras were gyro stabilized with an auto erection system and a moving counter weight to compensate for the film weight transfer during the pulling of film.
Those auto pilots had electrically powered graphite clutches and they were very flaky if not cared for properly.
The turns were very hard on the vertical stabilizer and we had to keep a close eye on it if we did not want to lose one. This was mostly done with C-180 mapping planes and this was from the mid 60’s to the mid 70’s when the better optics made the Fairchild T12 Cameras of WWII unable to maintain the specs that were starting to be called for in many of the jobs. The older cameras were still in use for some things into the late 90’s that I am personally aware of.
Not sure which post you referred to, the only Sperry I saw in the post was Lawrence, Elmer’s son. Elmer invented the Gyrocompass, Lawrence the Autopilot.
Wow, Gus’n’Spot, thank you. As a cartographer who worked some in the 1990s, I was indebted to the tremendous skills and knowledge of you and your colleagues. I regret never having learned more about exactly how you guys pulled it off, but your post vividly goes some way toward redressing this deficiency of mine.
In the mid 70’s were we doing good enough to get an Zeiss RMK much like this one. In the 6" version, then on to a better one with moving platten, gyro mount, video view finder with a feed for the pilot with next shot moving lines, moved up to a 1975 C-310Q turbo A sports car compared to the Piper Navaho which came next.
Remember, I started with Jack Biendorf, Gleason Romans and the other old timers who all knew each other. There was the ‘Wild’ camera guys & the Zeiss camera guys and they all started with old Fairchild surplus T-12’s K 22’s 6" 12" & 24" focal length versions. The 3½" versions had not come out until later in the Zeiss & Wild cameras.
It was mostly B/W film as no one thought color was worth the cost for the little gain in accuracy some could get with it.
I processed all our B/W film in hand wind developing cans. 9½" film up to 500 feet long.
Enlarger was home made with red dot lenses, eventually got an HE- 12
All enlargements were hand processed, contact prints were exposed with an Log-E unit, also exposed the glass plates for plotting.
Started with 2 Kelsh 2 projector plotters, add an HE 3 projector, then went to the Zeiss PG - 2 then the PG 4 and now I am not sure what they are using as I am no longer in the business.
this all happened from 1948 to 1969 when I got on board and on to 1998 when I left.
I have a few pictures of what was going on, a million stories, we were a small outfit compared to some but we put out some of the best work.
I am not up to writing a book but I am sure there are some.
Was in on the early civilian use of the Bendix nitrogen charged thermal cameras.
Was pilot for one of the first test flights of the next generation of thermal cameras with germanium lens. Did that with a C-206
OK, I’ll quit now, Bawahahaha 
Thanks again, Gus.
