Two roads diverged in a yellow wood...

Alexb2000

En-Route
Joined
Nov 28, 2010
Messages
3,530
Location
Dallas, TX
Display Name

Display name:
Alexb2000
I thought I'd start by quoting Robert Frost accurately in hopes that this will be a productive, factually accurate, thread. Good luck right?:)

In my other recent thread about the SR22T (that has become a petty single vs. twin debate) I noticed something that I thought might be interesting to discuss the two different paths I see pilots taking. What I mean is some pilots have become monitors of the airplane, perhaps managers is a better word. While others tend to be manual operators only using technology as far as it helps them operate the machine directly.

I was trained as a stick and rudder airplane operator and I came face to face with the other perspective when I flew a Boeing 737NG sim in France. The French instructor was all about automation and we argued when my reaction to any adversity was to hand fly.

Flying the SR22T the other day with the salesman I noticed a similar difference, he managed the airplane I was always trying to fly it.

Before all you old school guys jump up and say, "children of the magenta line!" Let me give you a practical example.

We did a simulated engine out at night in the Cirrus. My first instinct in that scenario was... Hand fly, best glide, etc. I would have quickly pulled up the nearest and really focused on hitting best glide perfectly and holding a direct course to the nearest so as to lose the least amount of altitude possible. That's how I was trained. Once trimmed out and on heading and speed, I would have begun the checklist, etc. maybe a couple of minutes to accomplish. Once headed that way I would have been doing the mental math, 8:1 glide, altitude AGL, wind direction and speed, etc. if the calculation was close I would have kept flying it down hoping to make it. If I didn't I would have been looking for alternatives visually, etc.

The salesman never even considered hand flying. He pulled up the nearest, direct to, enter, enter, and set IAS mode to best glide. That took a few seconds then he began putting in the Vertical profile of the field to see if we would make it, another few seconds. Then as we slowed to best glide he could have done the checklist, maybe a minute. As soon as we hit best glide he looked at the Vpath indicator and knew we wouldn't make the field. He did no math, the airplane was perfectly on course and speed, he only had to do one thing at a time and he could concentrate.

Looking back, he did a far better job in this emergency scenario than I did because he leveraged technology.

So I talked this over with a freight dog instructor I know that hasn't had an autopilot to use flying turbines the last 5,000 hours. He said, that while he is obviously a big believer in hand flying skills, he deals with this issue a lot. He says EVERY pilot that comes to him for instruction or flies with him professionally thinks their hand flying skills are razor sharp, but 90% of the time they are not even close. He said that those flying for pleasure would most often be better off letting auto fly when it counts like on an approach.

I can see advantages to both sides and obviously the perfect answer is to be able to do it all. I'm just not sure that is possible for many pilots.

Fire away, but let's try and keep it civil. This board is sinking fast when every thread gets blown up on the first page.
 
""""Looking back, he did a far better job in this emergency scenario than I did because he leveraged technology.

He did better because he was a SIM instructor and had practiced that scenario hundreds of times.
 
Fire away, but let's try and keep it civil. This board is sinking fast when every thread gets blown up on the first page.

I was going to do a point by point and decided it wasn't necessary.

Here's the thing. All that automation stuff is there to help the pilot manage the flight. Yes I said it. Manage the flight. A good glass setup connected to a decent autopilot can fly the airplane better than a human pilot can. If you let the automation do what it does best, that frees up you the human pilot to manage the flight better. You have more brain power available to assess various situations. That is especially helpful in an emergency situation where stress levels tend to hinder ones ability to think clearly and in a timely manner.

In that regard, using the automation is a better deal.

HOWEVER, one has to use the automation enough to be proficient with it. If you are out there flying with it every day or on a regular basis, you become very familiar with it and pretty much instinctively know what its going to do and what you need to do in any given situation.

BUT, if you don't have a good grasp on what the automation is doing and can do for you, you quickly get mired in trying to make the automation do what you want it to do. When the automation becomes a hindrance, then it is time to take it out of the loop and do what needs to be done.

Bottom line? Automation is a GREAT tool for those that are proficient with its use. In my opinion, being a manager is not a bad thing. But if one cannot maintain proficiency with it, it soon becomes a detriment and one is better off hand flying.

The caveat is that one should not rely on the automation so much that the motor skills atrophy to the point of being useless. The conscientious pilot knows how to stay proficient at both.
 
""""Looking back, he did a far better job in this emergency scenario than I did because he leveraged technology.

He did better because he was a SIM instructor and had practiced that scenario hundreds of times.

Well, he was actually talking about two different people. One was the 737 sim instructor, one was the Cirrus salesman. It goes to support my point that if one maintains proficiency with the automation, it is a great tool. If not, one is better off hand flying.
 
Well, he was actually talking about two different people. One was the 737 sim instructor, one was the Cirrus salesman. It goes to support my point that if one maintains proficiency with the automation, it is a great tool. If not, one is better off hand flying.

And if one knows when to abandon the automation and switch to handflying. The original Automation Dependency video had some good examples in it: change of plans where you are continuing on the same path but the AP is setup for something different, closing with other aircraft, or anytime things get intense in the cockpit.

In OP's case, I'd critique having the avionics setup and turn the plane instead of the PIC turning generally toward the emergency field and pitch roughly for best glide before setting up (or better yet, have SIC set up) the automation to fine-tune and manage everything.
 
And if one knows when to abandon the automation and switch to handflying. The original Automation Dependency video had some good examples in it: change of plans where you are continuing on the same path but the AP is setup for something different, closing with other aircraft, or anytime things get intense in the cockpit.

In OP's case, I'd critique having the avionics setup and turn the plane instead of the PIC turning generally toward the emergency field and pitch roughly for best glide before setting up (or better yet, have SIC set up) the automation to fine-tune and manage everything.

I think your point goes to my problem. I spent a lot of time flying aircraft that wouldn't hold a perfect airspeed descent, somehow this became my habit and mentally I just kept thinking hand flying is always the best way to go in most situations and AP's are good in cruise to keep down the fatigue. I've since flown a good number of hours in aircraft that can descend on a perfect speed, or pitch, or rate, or whatever and yet I never really thought about it enough to change my habits. This feels like a mental block, similar to having a iPhone for years and then some 8 year old picks it up and can do more with it than you can in 30 minutes.

So to answer your question, I'm not 100% sure anymore when the right time to turn off the autopilot is, but last week I would have been sure and argued about it.
 
So to answer your question, I'm not 100% sure anymore when the right time to turn off the autopilot is, but last week I would have been sure and argued about it.

Is the environment dynamic or static? Are you maneuvering or trying to maintain a speed/altitude? Are you reacting to a change or deliberately going through pre-set items?

On of these things, the Automation is excellent for. The other, not so much.
 
As opposed to quoting Robert Frost inaccurately?
 
Bottom line? Automation is a GREAT tool for those that are proficient with its use. In my opinion, being a manager is not a bad thing. But if one cannot maintain proficiency with it, it soon becomes a detriment and one is better off hand flying.

I agree 100% with Greg's statement. A couple years ago I decided to do my IPC in a C-182 with a G1000. I flew it about 4 or 5 hours, then decided I was crazy. The 182 was not mine, and there was no way I would stay proficient enough with it to fly it in IMC.
 
When I trained for my Citation type rating and during recurrent, they emphasized using whatever tools I had available, including the co-pilot, auto-pilot, controller, anything to help complete the flight safely! :D I hand fly a good bit, but a good auto-pilot is a good thing. :D
 
Well, he was actually talking about two different people. One was the 737 sim instructor, one was the Cirrus salesman. It goes to support my point that if one maintains proficiency with the automation, it is a great tool. If not, one is better off hand flying.

To play Devil's Advocate: If one is not proficient with the automation, what makes you think their hand-flying skills are any better?

Maybe the key is better/simpler/more-intuitive automation tools?
 
Someone posted an airline safety talk and the guy mentions handflying vs reprogramming the plane to resolve conflicts. It was an interesting bit maybe someone can find it.
 
Someone posted an airline safety talk and the guy mentions handflying vs reprogramming the plane to resolve conflicts. It was an interesting bit maybe someone can find it.

"Automation Dependency" aka "Children of the Magenta". An AA pilot seminar circa 1997. It's out there, but AA keeps trying to pull it down.
 
Last edited:
I would say you should be proficient in both approaches.

I personally use all the available automation when doing a hard IMC flight under IFR. For example if you're flying with a G1000 as in the case of that Cirrus. I would say single pilot IFR with the G1000 is almost trivial. But seeing I want to maximize my probability of living (along with my passengers :D ) yes I'm going to use every available resource I have. So if I have that at my disposal, you bet I'm going to use it. Including all those great tricks you have VNAV navigation, and so on.

But it is boring because you're basically just managing the flight. You're not flying. The computer is flying and you're just telling it what to do.

So when I fly for fun using VFR I personally use no automation. If I can get away with not talking, even better.
 
I thought I'd start by quoting Robert Frost accurately in hopes that this will be a productive, factually accurate, thread.

If you're suggesting that the usual quote ("Two roads diverged in a wood") is inaccurate, that's not the case. Both wordings appear in the poem. And if you include the less-traveled-by part (as the excerpts typically do), then only the usual formulation is accurate.
 
If you're suggesting that the usual quote ("Two roads diverged in a wood") is inaccurate, that's not the case. Both wordings appear in the poem. And if you include the less-traveled-by part (as the excerpts typically do), then only the usual formulation is accurate.

I was just having a little fun. I was watching something on TV and no one could quote the first line correctly, even though some were very confident about their knowledge. I thought it would be a good parallel to many aviation discussions.
 
To play Devil's Advocate: If one is not proficient with the automation, what makes you think their hand-flying skills are any better?

All automation is different, so it depends. As an example, back when I was flying 7 different types regularly, 4 of them had some derivative of a Century III autopilot, another had a KFC200, another had a KFC300, and I forget what the last one had. I wasn't trained on proper use of the KFCs, so when I had a night ILS to mins (first approach in the plane), I hand flew it. there was a runway at the end, so I must not have screwed up too badly.

That said, had I been trained and familiar with the KFC300 in the plane, I would've used it instead.
 
To play Devil's Advocate: If one is not proficient with the automation, what makes you think their hand-flying skills are any better?

Maybe the key is better/simpler/more-intuitive automation tools?

Hand flying skills once engrained come from a different part of the brain, hand eye is much simpler to retain and regain when needed than is analytical logic of deciphering a complex system.
 
I have a simple panel with a Brittain wing leveler and an Apollo 50 GPS. I do not fly enough to be really good at hand flying or at digging deep into what the GPS can do. I feel more comfortable using my tablet to find information since a) it is up to date and b) I can practice with it at home.
Yesterday on my 1.4 hour flight taking a visitor around the area, I pulled on the wing leveler everytime I got pointed towards a new site. I hand flew all the turns and pointed out the sights.
On the other hand, on my flights to and from OSH I got to altitude, trimmed and set the WL and folded my arms and monitored the progress.
I don't know if I would have done differently in either case if I had a fancy panel because I probably wouldn't be as knowledgable as a salesman of the product.
 
I believe it is important to keep in mind that being able to turn on the autopilot or even use it to fly an approach is not exactly what I am talking about.

IMO to really see the differences, you need to fly a bunch of different procedures especially emergencies with a pilot that has really good skills in areas you don't and compare results. As I mentioned this caught me completely by surprise.

For example, I was thinking about classic hand flown maneuvers like the impossible turn. I wonder if in practice it could be flown as well or better with an autopilot? The turn rate wouldn't be as steep, but it would be precise and allow you to focus on other things. I don't know the answer, but this is the kind of thing I would think would be interesting to experiment with.
 
Last edited:
For example, I was thinking about classic hand flown maneuvers like the impossible turn. I wonder if in practice it could be flown as well or better with an autopilot?

No. Because the computer wouldn't be programmed for that eventuality and by the time you got a turn initiated and a descent programmed, it is much easier and probably more precise to hand fly the plane.

Having said that, that is a limitation of the current systems, at least as far as the Boeing stuff I am familiar with is concerned. I suppose the computers could be programmed to look at the current conditions, wind, altitude, climb rate, etcetera, and compute whether or not the turn could be made and execute it to the ragged edge of the aircraft's performance capabilities. But, as far as I know, that hasn't been done.
 
Last edited:
I fly for sport. I like steam guages, a good GPS, a throttle on the left and a stick in the middle. I want to be a pilot, not a computer programmer. I want to steer the plane and make it do everything it can do. Its hard to land because its a taildragger each landing is an adventure. Not everyone can fly my squirrly bouncy taildragger, only brothers (and some sisters too). Its a challenge and Im pretty good at it.

I recognize there are other ways to fly planes, other equipment. That's cool. The less automation the better, I say.
 
No. Because the computer wouldn't be programmed for that eventuality and by the time you got a turn initiated and a descent programmed, it is much easier and probably more precise to hand fly the plane.

Having said that, that is a limitation of the current systems, at least as far as the Boeing stuff I am familiar with is concerned. I suppose the computers could be programmed to look at the current conditions, wind, altitude, climb rate, etcetera, and compute whether or not the turn could be made and execute it to the ragged edge of the aircraft's performance capabilities. But, as far as I know, that hasn't been done.

On a G1000 I could put it in heading mode and IAS descent in literally a couple of seconds. It would take me much longer to recognize the emergency. I expect that I could hand fly it better, BUT, I would be 100% tunnel vision focused on doing the maneuver. On AP I might be able to look around for alternatives, or do an emergency checklist and restore power, or… Single pilot also changes things. I don't know for sure, but my mind is now open to the possibilities.
 
For example, I was thinking about classic hand flown maneuvers like the impossible turn. I wonder if in practice it could be flown as well or better with an autopilot? The turn rate wouldn't be as steep, but it would be precise and allow you to focus on other things. I don't know the answer, but this is the kind of thing I would think would be interesting to experiment with.

You want to fly the "impossible turn" at standard rate? You're not going to make it. And descents in FLCH mode are problematic, especially close to the ground. In VS mode, you'll get behind the power curve in a significant turn at wings-level best glide and lose a lot of altitude. VS mode behind the power curve is unstable. If you descend too fast, it will pull back on the stick and make you descend faster.

I have yet to see a coupled autopilot that will turn faster than standard rate.

Don't overestimate its capabilities. Think of it as an über-precise student pilot. It will do exactly what you tell it to do, even to the point of crashing you. Like, follow altitude profile with insufficient engine thrust until you stall (Asiana).
 
I expect that I could hand fly it better, BUT, I would be 100% tunnel vision focused on doing the maneuver.

Yeah, that is the problem. You are much better off finding the best option more or less in front of the airplane and taking your chances with that.

On AP I might be able to look around for alternatives, or do an emergency checklist and restore power, or…

ONLY ONLY ONLY if the autopilot was following a computer generated profile. You are talking about the impossible turn here. By definition you don't have TIME to do any of the stuff you are talking about.

I don't know for sure, but my mind is now open to the possibilities.

The possibilities, given current technology, is to FLY THE DAMNED airplane and don't lose control. The impossible turn is more risky than taking your chances straight ahead.
 
You want to fly the "impossible turn" at standard rate? You're not going to make it. And descents in FLCH mode are problematic, especially close to the ground. In VS mode, you'll get behind the power curve in a significant turn at wings-level best glide and lose a lot of altitude. VS mode behind the power curve is unstable. If you descend too fast, it will pull back on the stick and make you descend faster.

I have yet to see a coupled autopilot that will turn faster than standard rate.

Don't overestimate its capabilities. Think of it as an über-precise student pilot. It will do exactly what you tell it to do, even to the point of crashing you. Like, follow altitude profile with insufficient engine thrust until you stall (Asiana).

The GFC700 can do 22 degrees.

What if the problem was a lost fuel pump and doing the checklist restored power?

What if you wouldn't make it regardless of technique and you flew right over a golf course because of tunnel vision?

Are all of us really proficient in a 45 degree power off banked turn starting from a climb attitude performed at 1.03 VSO close to the ground?

I just don't know the answers.
 
The GFC700 can do 22 degrees.

Not steep enough.

What if the problem was a lost fuel pump and doing the checklist restored power?

Then it does.

What if you wouldn't make it regardless of technique and you flew right over a golf course because of tunnel vision?

If that is the case, you are going to miss other things too. I am not sure what your point is.

Are all of us really proficient in a 45 degree power off banked turn starting from a climb attitude performed at 1.03 VSO close to the ground?

What's your point?

I just don't know the answers.

You keep adding questions. The original question was cut and dried. The more crap you add the scenario, the more obvious it is that you need to fly the ****ing airplane and land straight ahead. Close to the ground you don't have time for much of anything else.
 
Not steep enough.



Then it does.



If that is the case, you are going to miss other things too. I am not sure what your point is.



What's your point?



You keep adding questions. The original question was cut and dried. The more crap you add the scenario, the more obvious it is that you need to fly the ****ing airplane and land straight ahead. Close to the ground you don't have time for much of anything else.

Just a little thinking out loud.

As you know, at some airports, straight ahead is not practical. My home field for example.

Relax, it's just a conversation.
 
The GFC700 can do 22 degrees.

The GFC700 will fly standard rate turns, up to 22 degrees. It will be a lot shallower at slow speeds. We've tried to use it to fly search patterns with 1 mile spacing, and it would just barely do it at search speed (90 KIAS), and then only if the wind was mild. At best glide, it's gonna take 4000 feet downrange to do a 180, and a full minute. Now, you've missed the airport entirely and lost a minimum of 500 feet.

And the GFC700 will not coordinate your turns. It slips like mad if you don't touch the rudder pedals. So, your 500 foot loss is even larger. It's not a set and forget device.
 
The GFC700 will fly standard rate turns, up to 22 degrees. It will be a lot shallower at slow speeds. We've tried to use it to fly search patterns with 1 mile spacing, and it would just barely do it at search speed (90 KIAS), and then only if the wind was mild. At best glide, it's gonna take 4000 feet downrange to do a 180, and a full minute. Now, you've missed the airport entirely and lost a minimum of 500 feet.

And the GFC700 will not coordinate your turns. It slips like mad if you don't touch the rudder pedals. So, your 500 foot loss is even larger. It's not a set and forget device.

I've never tried it really slow. I was thinking just use ROL mode at 22 degrees. Good points, thanks.
 
This is a specific case of the larger and perennial tools vs. skills debate. (For a legendary exposition of this question read the legend of John Henry the Steel Drivin' Man.) Skilled workers can do better than tools in the hands of unskilled folks until the tools get so good they can't. Good tools, in the hands of those who know how to use them are very valuable. No tool, no matter how wonderful, is a substitute for understanding the problem.

That said, we build better and better tools until they're at least good enough for most all the time. And then better.

A couple of examples from automobiles:

Cars in the early 1900's had manual spark advance. The driver (or riding mechanic in racing cars) manually controlled the spark advance. This allowed better performance and drivability than any fixed timing system in the hands of a sufficiently skilled operator. Then along came mechanical spark advance systems (or vacuum) based on engine RPM. These worked well enough for most conditions and better than semi-skilled or unskilled operators. Soon folks who had no idea how spark advance should work could drive cars and get reasonably good results. (And I'm pretty sure some skilled drivers argued that if you couldn't learn how to control the spark advance you had no business driving a car.) Occasionally weird conditions would crop up or a maintenance issue and someone who still understood how spark advance was supposed to work would have to get involved. (Points, condenser, coil, timing light, etc.) Now we have computer controlled spark timing which can vary the timing based on many inputs and get astoundingly good results in all conditions we encounter. And self-diagnose (mostly) if they get messed up. Nobody except the engineers who design and test the systems have to understand spark advance any more. They just expect the car to do it. At least until they break down, nobody cares.

Consider automatic transmissions as well. I still prefer driving a standard transmission in my sports car, but if I was racing (road racing, you understand, not 1/4 mile or roundy-round stuff) you can bet I'd have a paddle shifter and the transmission would shift that thing much faster than I can. The machine can do it better because somebody who really understands the problem has designed a machine to do it better.

Knowing what the automation can do, how to make it do it and, most importantly what it won't do well is part of being a TAA pilot. Keeping your hand-flying skills up for when the machine can't or won't do the best thing is also part of your responsibility. Knowing which to choose when is the most important part.

None of which changes the fact that I'd rather (on a nice VFR day) fly it myself-just like I prefer shifting myself. It's my own enjoyment of the skill in those cases. In IFR and/or when the stuff hit's the ventilator, give me the tools and hopefully the skills to use and, more importantly, understand them.

John
 
Back
Top