Question about autopilot functionality on the pmdg

[Jonathan Johnson]

Flying last night into Boston was a mess for me. The whole flight up until the end was good. But it started off okay. ATC came online, I was guided to 22L, everything going smooth until FSX just froze.. I think something went wrong with the scenery I had (the airport then suddenly disappeared).. in my frozen state it caused the aircraft behind me to have to go around.. embar[Mod - Happy Thoughts]ing.. but once the simulator fixed itself my airspeed shot way up.. don't know at all what happened so I myself have to call a missed approach.

 

My problem (and subject) is not with that (though that was pretty bad).. but with an autopilot problem I had with the PMDG. After declaring missed approach, I was told to climb/maintain 4,000 feet. I switched from Vnav to altitude hold (already had Lnav off from when ATC gave me heading directions), but it wouldn't hold altitude. The altitude hold... did nothing? That's the best way to describe it.. Then later on, near the approach for the glide scope, the autopilot master (CMD A) shut off, and I couldn't turn it back on. I could engage/disengage any of the other autopilot switches (N1, Speed, HDG, APP, etc...), but could not engage the autopilot master.. neither A or B would engage, and the same goes for CWS... none of the master switches would engage. And I have no idea why this happened.. I didn't do anything insane.. It just seemed like after I switched from using the LNAV to an altitude hold my flight just went terrible.. I eventually had to disconnect from Vpilot to try and fix the issue, which ended with FSX crashing and me now having to start all over with the flight (and being too frustrated to want to try the flight over again..)

 

I should also note that this is the first time I've had to switch to an altitude hold in flight.. usually (and maybe I'm doing this wrong which would be nice to know) when I get near the airport and approach starts giving me directions/altitude holds, I can just set the altitude on the autopilot and keep the vnav on, and it will maintain the altitude they give me.. so if ATC says descend 4k, it won't descend lower then that...

 

hope I'm making sense.. don't want to repeat these issues the next time I fly

[Simon Kelsey]

Hi Jonathan,

 

Lots of questions! Caveat -- my 737-specific knowledge is thin on the ground. However, in general you would be well-advised to refer to the FCOM which I know PMDG supply with the NGX and have a read up on the procedure for carrying out a missed approach.

 

How comes you were pressing ALT HOLD? The normal procedure in the 747 for a missed approach would be first to press the TOGA switches, because this is the only way (without disengaging the autopilot and turning off the flight directors) to get out of LOC | GS if you are established on an ILS approach, and I am sure that the 737 is the same. Obviously you already set and checked the missed approach altitude in the MCP when you captured the glideslope right?

 

After pressing TOGA you should then get (in the 747 -- 737 should be vaguely similar but there may be slight differences in terminology or function) THR | TOGA | TOGA. Select the flaps to whatever the go-around setting is (20 in the 747, check the FCOM for the 737). Confirm that the thrust increases and is sufficient for the go-around (ask Emirates), rotate to the go-around pitch attitude (or confirm that the AP is doing so. Check that you have a postive rate of climb, then (and only then - ask Emirates) gear up.

 

Above 400 feet radio you can then select a roll mode (LNAV or HDG SEL depending on the situation/ATC instructions). Make sure the missed approach route is being tracked or the [Mod - Happy Thoughts]igned heading is being flown.

 

You then have a couple of options -- the old way was to select FLCH then SPD at acceleration altitude, set the speed to the manoeuvring speed for the planned flap setting and clean the aircraft up, the newer way is to leave it alone until it captures the missed approach altitude (check that it does), then select VNAV (or FLCH then SPD) and clean up as required. At this point you can then do whatever you need to -- HDG SEL for vectors back for another approach, LNAV to a waypoint, whatever.

 

There are lots of reasons why the 737 autopilot may not engage and I couldn't pretend to know them all, but I know that amongst other things it is possible in the PMDG NGX to break the shear pins by overriding the autopilot with the control column (resulting in no A/P engagement for the rest of the flight) and that the A/P will not engage if the aircraft is not in trim.

[Jonathan Johnson]

This was the first time I've ever done a missed approach (never had to do a go around either..) so I'm pretty sure my nerves got to me as well as I totally forgot to press the toga switch. I did just what you said before (disengaged autopilot and F/D) to get out of the glidescope .. I was already nervous since FSX had decided to lock up on me (and cause a go around), so yeah I missed that . What I did:

 

1. Disengaged the flight director/autopilot.

2. Turned back on f/d and autopilot master (CMD A). Set 4,000 as altitude, turned Alt. hold on.

3. Flew runway heading, contacted approach, eventually got turned around.

4. Realized alt. hold wasn't on when approach contacted me. Tried to fix.

5. Couldn't fix altitude issues, soon noticed autopilot was off. Eventually disconnected from vatsim to try and correct issues (and not disturb other traffic in area), game crashed.

 

Missed approach was in the MCP, but I definitely failed at the rest of it. Later tonight I'm going to do an offline flight and practice a missed approach. Now that I understand what to do. I do also believe that it was not in trim, and thus wasn't going into autopilot.

[Simon Kelsey]

No worries! If it's any consolation, remember that the missed approach/go-around (same thing) is one of the most regularly botched manoeuvres in real life flying as well: it's not regularly practiced (at least not with all engines operating!), depending on the reasons for the go-around it can be unexpected, and there's a lot to do and check in a short space of time, so in terms of actually handling it you did a pretty good job to remember how to get out of APP mode and get the aircraft flying in the right direction!

 

In real life, the go-around actions would often be discussed as part of the approach briefing before top of descent -- it doesn't do any harm to do the same in FS as well, and if you mentally rehe[Mod - Happy Thoughts]/talk yourself through it (I sometimes 'touch drill' the actions as well -- think about which buttons you need to press and in which sequence) every flight, when it comes to actually having to do it it'll be second nature! As I say, find the Go Around/Missed Approach section of the Amplified Procedures in the FCOM and have a read through -- bullet point the actions if it helps and keep them handy for your rehearsal/approach briefing.

 

In general terms, ALT HOLD is not a very common mode to use (it will level you out and capture whatever altitude you are currently at) -- do whatever the FCOM says you should do, but as a general rule I would go with FLCH or VNAV as these modes will sort out your autothrottle mode as well and capture the MCP altitude.

 

It's a great idea to have a practice of the missed approach, and I would suggest practicing from a variety of different positions -- as a starting list of things to try (or at least think about):

 

- From minima (i.e. 200ft ish above the runway)

- On final approach with the autopilot engaged

- On final approach without the autopilot engaged

- After touchdown (what is the latest point at which we can initiate a G/A? Are there any 'gotchas' in your aircraft type (ask Emirates re: 777)? How will the procedure be different compared to a go-around in the air?)

- From an early stage in the approach (let's say you've been vectored on to the ILS and cleared for the approach and you're descending on the glidepath at, say, 3000ft. ATC instruct you to go-around. What do you do? What if the missed approach altitude is below your current altitude?

 

Have fun!

[Trent Hopkinson]

Interesting fact: If you apply enough deflection on your yoke (in the PMDG 737, in the real world it's "Force" on the yoke) the linkages that allow the servo motors on the autopilot to control the aircraft break. After that, no more autopilot control is possible. Usually it's only going to break the servo linkages for the autopilot that is engaged (ie if Autopilot A is in command, Autopilot A will break, but Autopilot B will still be available).

 

If you engage both A and B for approach... well..

[Jonathan Johnson]

I only had A on for approach, and the simulator froze pretty close the airport. I believe I was uhh.. 500 ft above? Can't remember.. but do remember that once it sorted itself (really thought I was going to have to force close FSX) the speed jumped way up and I overshot the runway.. which once I knew I was going to overshoot I called missed approach.. And then the issues continued lol..

 

Those practice suggestions are definitely nice! You never know what can happen (as I found out) so it'll be nice to practice.

[Magnus Meese]

Also, if you only toggled the F/D on your side, you just simply made the other Flight Control Computer (belonging to the guy in the other seat) the Master FCC (the MA indicator just next to the F/D switch), meaning that the flight modes (LOC and GS) does not cancel, and you're stuck in the situation as described above, with the system rejecting any engagement of CMD A/B. To reset the system, both F/D switches must be toggled, with both being in the off state simultaneously at some point (or you just play pong with the Master status, making no real change).

[Nick Warren]

Hand fly? Stick and rudder? I don't mean to be curt or simplistic in my response, but it seems to be so much of a lost art and in this case with the automation challenges is even more beneficial. I just posted a link to a great article on this very topic in this forum https://forums.vatsim.net/viewtopic.php?f=8&t=72621 but it really didn't generate any discussion.

[Robert Shearman Jr]

Hand fly? Stick and rudder? I don't mean to be curt or simplistic in my response, but it seems to be so much of a lost art and in this case with the automation challenges is even more beneficial. I just posted a link to a great article on this very topic in this forum https://forums.vatsim.net/viewtopic.php?f=8&t=72621 but it really didn't generate any discussion.

... only silent agreement.

[Simon Kelsey]

Hand fly? Stick and rudder? I don't mean to be curt or simplistic in my response, but it seems to be so much of a lost art and in this case with the automation challenges is even more beneficial. I just posted a link to a great article on this very topic in this forum https://forums.vatsim.net/viewtopic.php?f=8&t=72621 but it really didn't generate any discussion.

 

I have read the article in question and I confess I don't really get what the author's issue is.

 

I shall expand.

 

Hand-flying is great. I am not, at all, suggesting that we should not be proficient in hand-flying our aircraft and being able to point it in the right direction.

 

But modern transport-category jets are not just big C172s, and this is not the 1960s when traffic density was much lower and airspace and environmental concerns did not rely on the accuracy of modern nav systems, RNP, etc etc. The author of the article misses the point somewhat, which is that if you do not understand the automation and use it appropriately it will kill you!. Hand-flying whilst leaving the FD engaged and ignoring its commands is, from a technical point of view, what lead to the Asiana crash at SFO, and there are a catalogue of other incidents and accidents that have arisen as a result of doing exactly the sort of thing described in that article. That, I would suggest, is why the author's instructor was so exercised about it. It was not the act of hand-flying that was the problem -- I'm sure that had the FDs gone off, there would have been no issue. There is no point in throwing away all the automatics to make a turn within a nanosecond of receiving the clearance if you then end up in a smoking hole in the ground because you got the system (and/or yourself) confused.

 

The automation is an essential tool to help reduce flight deck workload and keep everybody safe. Yes, of course pilots need to understand its limits and be proficient in hand-flying. But just taking the "real men hand-fly" attitude is not a solution to poor systems knowledge and in a high-workload environment there are many, many examples where it has ended in tears.

 

I'll take your article and raise it with this one: https://mmsba.wordpress.com/2013/09/03/the-automation-paradox-making-training-relevant/

[Nick Warren]

Simon,

 

I appreciate you taking the time to provide some discussion on the matter. Full disclosure for myself is that while I am a pilot, I am not an ATP in real life, nor have any ratings in jet aircraft. In that, I appreciate the response and the article you submitted.

 

In reading it though, it struck me that the constant theme throughout was the inherent struggles with automation. Problems caused with automation are often due to circomestance of training and or misuse at least from what I gauged by the article. I think this is where the two articles meet. Where your article stresses the importance of automation with comprehensive training and competency, the one I posted lends focus to when perhaps automation fails, or just may not be a good idea in the first place.

 

I certainly understand the complexities with current and future navigation and airspace. Couple that, with as you said, trying to stuff more and more aircraft in smaller boxes. So, I certainly can appreciate automation as a means to ensure proper adherence to navigation and certainly to reduce overall workload and safety. Again, however, there has to be a time to turn the switches off, and to possess the knowledge and skill to hand fly the aircraft accordingly.

 

In the OP's example, having done a missed approach, this may have been in my opinion one of those times. He wasn't following a published MA procedure, and was given a straight instruction to climb. As automation, in the OP's case, failed for whatever reason, would this not be a perfect time to have just hand flown the aircraft to the [Mod - Happy Thoughts]igned altitude and vectors as opposed to succomebing to significant frustration in not being able to establish the automation?

 

When we cite examples such as Asiana, the transition from automation in itself was rather foreign by company training. Company policy and training encouraged near full automation in their operation. One of the items highlighted in the analysis was a need for manual flight training. So again, when automation fails for whatever reason, what are we left with? To me that is the importance, and the reason I posted my original response.

 

Of course we can argue this is sim, or even a game to some. I truly believe the OP has a desire to learn the systems and learn them well to be better at his craft, and that is a good thing. What I see often though, is a culture who doesn't want to learn the basics even. They want push button flying from takeoff to landing. For me it's kin to insisting on an ILS approach on a crystal blue day with 30 miles of visibility.

 

Anyway, my two and a half cents. Thank you again for engaging some good discussion.

[Rick Rump]

John,

 

Kyle Rodgers has some great videos for the NGX I would highly recommend checking out.

https://www.youtube.com/user/scandinavian13/videos

[Daniel Churchman 1050104]

Pardon me coming late to the party...

 

Both the first article, linked to by Nick Warren, and the second provided by Simon Kelsey were great reads; experienced pilots presenting their considered opinions.

 

Both articles acknowledge the need for hand-flying skills. Both articles seem to acknowledge a need for more/better automation training.

 

Where they differ; the first article presents a pilot "falling back" on his manual skills because, as he admits, he wasn't able to recall the location of several automation controls. He found it easier to take manual control. Still, he somewhat defends this with the argument that manual skills are vital; nevertheless he went manual only because of his inability to promptly apply the necessary changes to his flight path with the automation.

 

The second article really makes the point that automation = safety. They chose to use it even though the weather was great because they judged it was better to free themselves up to keep their eyes outside the aircraft as much as practical to monitor traffic in the busy airspace.

 

Which one is right?

 

I think both have their points, but I come down in favor of the second article favoring the automation. Here's why.

 

The second article pilot clearly acknowledges the importance of manual skills. They in fact considered flying in manually, but professional judgement told them to use the automation for exactly what it is good for; relieving pilot load to free them up for things like maintaining situational awareness outside the aircraft as well as in. I think it is safe to infer from the article that both pilots keep their manual skills in practice when they can, so they should be prepared and able should automation fail.

 

The first pilot has made a decision because of lack of knowledge about his aircraft. That's a red flag right there. I understand that he was there to learn this stuff; his instructor made an issue of it, the pilot admitted this, yet he also still defends his action. He's right, but also wrong. Yes, given the situation he made the right choice. "Aviate" comes first; always fly the plane. But it is dangerous to defend this decision on the basis of "manual skills first". The reason he did this was a lack of familiarity with his aircraft, NOT to sharpen his manual skills. I don't think any argument can be made against the [Mod - Happy Thoughts]ertion that he should have used the automation to make the heading change; he just had a brain fart and forgot how to do it in the heat of the moment. I don't criticize him for that (he was learning), nor his decision to go manual rather than unduly delay the turn.

 

I only criticize his defense of the situation. If you are in IMC and get an unexpected directive, you really need to know how to use the automation to achieve the command. Sure, you might be able to fly a heading, change altitude or speed, without automation. But what about an unexpected "Fly direct to" instruction?

 

The message I hope should be, "Both manual AND automation skills are vital. Automation is a great safety tool for many, many reasons, but pilots also need to fully understand what the automation is making the aircraft do, where the aircraft is, and when and how to take manual control and the skill to do so competently and confidently."

 

I think that's true for virtual flying too.

[Daniel Churchman 1050104]

It has been pointed out to me that a Direct To can be flown by hand. Of course, instrument flying can be hand flown. The boon of automation is to relieve pilot workload.

 

My core point remains, however. The first article pilot makes a defense for his decision based a need (and lengthy argument) for hand flying skills.

 

As I said in my previous post; I feel the decision s he made in the cockpit were correct under the circomestances. However, the reason the decision was correct has nothing to do with the importance of maintaining hand flying skills. He forgot where a switch or two were. That's it. I think it's dangerous to not acknowledge simply and clearly what happened; doing otherwise sets off all kinds of warning in my mind. Incorrect [Mod - Happy Thoughts]umptions or attitudes come from deception or misinformation. I think pilots cannot afford to do this.