Simulating complex flying scenarios: stuck throttle
- Credit: Archant
There are several strategies to deal with a jammed throttle, but if the cloudbase is low and the power stuck at max then your options are limited, as Eugenio found out in this month's sim session...
Flight Takeoff R24 from Denham, departing to the North
Surface wind 270/07KT
Clouds Broken at 3,600ft amsl
Scenario Stuck throttle
I’ll be honest. After two pretty challenging simulator scenarios (an elevator failure, and a full electrical failure in IMC) I thought this third emergency scenario−a jammed throttle−was going to be a bit of a breather. After all, I had already practised this malfunction in the real aeroplane a few times, and each time I handled it well and without much drama. But, as they say, the devil is in the detail, and little did I know that this flight, too, was going to be quite a rollercoaster.
So here I am, taking off from Denham Runway 24 on a standard northbound departure. The local procedure is to turn to the right after takeoff and leave the Denham ATZ at a maximum of 1,000ft, in order not to infringe Heathrow’s airspace. But, as it turns out, today I will be able to follow the standard routeing for only a minute or so.
When I power back to level off at 1,000ft, the throttle has no effect, the engine noise does not change, and the aircraft accelerates well beyond cruise speed. Not only does the engine noise not decrease, it actually gets louder as the speed builds up! The aeroplane wants to accelerate and climb, and for a few moments I break through the 1,000ft ceiling into Heathrow’s airspace. I don’t have carb heat on this Cessna 172SP, and switching one magneto off does not reduce power by much. All the while, the rev counter has gone beyond the red line, the engine is really roaring and the speed is in the yellow arc. I’m in trouble.
Suddenly I realise that, in my previous exercises in real aircraft, I had simulated a throttle sticking at cruise power, and that this is a whole different story. In this case, with the engine running at full power, I am being cornered (NB this is the failure mode for Rotax 912 spring-biased throttle linkages−Ed). I have to either red-line the engine, or overspeed the flaps, or bust an upper ceiling (airspace in my case, but it could also be cloud). In the heat of the moment the latter seems to me the best option (although I will end up questioning this after the flight) and so I declare an emergency to D & D on 121.5, telling them that I’m climbing in order to slow down and lower full flaps.
I pull up steeply and it takes 700ft of altitude gain to slow down to the white arc, at which point I extend full flap. A useful piece of learning here: trying to do this while gaining the minimum height means that the manoeuvre is quite abrupt, and I end up slower and more nose-high than I would like. If I had to do this again, I’d rather trade off one hundred additional feet of altitude for a smoother (and less upset-prone) manoeuvre−more on this later.
So, here I am with full flaps, and the engine happily below the red line. This, when I previously simulated the malfunction with a throttle jammed at cruise power, was usually where the main drama ended. But today things are different, and it will still take quite a lot of work to get out of this situation. The aircraft still wants to climb−at about 200fpm when flying at 80kt, to be precise−I can’t lower the nose, since the white arc ends at 85kt, and a lower speed only increases my ROC.
Can’t fly straight and level
The next card up my sleeve, then, has to be a sideslip. A note of caution here: some manufacturers explicitly prohibit sideslipping with the flaps down, for a variety of reasons. But today I do not have much else to try, I can’t keep climbing, and it turns out that the results are not bad: a full sideslip while maintaining a straight track at 80kt allows me to descend at about 250fpm. If I increase the bank (that is, I’m no longer going straight and I start orbiting), the ROD increases to about 450fpm.
- 1 Broadband from stratospheric airship
- 2 Flight test: Comco Ikarus C42C
- 3 Last Airbus 380 delivered to Emirates
I don’t feel comfortable red-lining the engine again, or shutting it down, so I decide that the ‘drag option’ will be the strategy I will follow. I will go back to Denham (to make the scenario a bit more challenging I’m pretending I’m somewhat short of fuel and that a diversion to another airfield is not available) where I will position myself at the right altitude by sideslipping, and then I’ll shut down the engine when I’m certain to make the runway.
The plan is simple, but I find the execution more difficult than I thought. The numerous tight orbits that I’m flying to descend are quite disorientating, and I suspect it would be even more so in the real aircraft, given the acceleration. To help a bit with this, I find that picking a prominent feature (in my case, the lake to my east) provides a good reference to maintain a sense of direction.
Another issue, though, is the effort entailed with the very prolonged sideslip, combined with the difficulty in controlling airspeed while banking steeply with speed near the top of the white arc. These are not impossible tasks, but they are more difficult than I anticipated.
Despite these difficulties the strategy ends up working quite well. First, I descend to about 2,000ft amsl in going back to Denham from the north, updating ATC of my intentions, since I’m still in Heathrow’s airspace. Then, when I’m certain that I can make the runway, I cut the mixture, the engine runs down, and I gently glide in for a rather uneventful landing on Runway 24.
My legs are somewhat tired from operating the rudder and I didn’t enjoy the endless sideslipping orbits, but I’m back on the ground in one piece after a short flight of less than twenty minutes.
Alternative ‘PFL’ and full speed strategies
So, the ‘drag option’ worked, although it did not feel particularly elegant nor easy. However, as Charlie Huke mentioned in his ‘QFI Logbook’ (April 2020 issue) there are other strategies that you can follow, and I also tried these.
One that actually works quite well is to just position yourself at a known point, as you would do when you fly a Practice Forced Landing (PFL), then turn off the engine (by cutting the mixture) and follow a PFL profile to a landing. In one case I simulated this strategy with CAVOK conditions, and it worked really well. I shut down the engine at about 3,500ft agl and found myself a bit high at my High Key point, which I usually place at 2,500ft agl. However, I had plenty of time and space to adjust my path, and managed to be on profile at my Low Key (1,500ft agl) and then all the way to the ground.
The strategy still worked, but proved more difficult to use, when I placed a broken cloudbase at 2,000ft agl in my simulated scenario. In this case I cut my engine at 1,800ft to avoid entering cloud and ended up substantially below profile at my Low Key point. I adjusted my path but didn’t manage to make up all the ‘missing height’, and my profile looked like I was going to barely make the airfield. So, I set mixture fully rich, the engine re-started and I re-positioned myself for a second attempt, which this time worked. So, the ‘PFL option’ worked perfectly with fair weather, but proved more difficult when I was forced down by clouds.
Another option that, frankly, I wouldn’t use in the real aircraft if the throttle was stuck at full power and beyond the red line, is what I call the ‘full speed’ option. The plan here is to approach the runway at whatever speed the engine forces you to fly at. Then, about a mile out, pull up steeply to the white arc for full flap. At that point, if the profile looks good for a landing, you can turn the engine off and land, otherwise you can lower the nose and go around for a second approach.
I tried this option with the throttle stuck at a cruise setting and it worked well, so that is a viable option in that case. With the power jammed full forward, however, I would consider using it in the real aircraft only if the RPM remained under the red line. That is not unrealistic: depending on a variety of factors, you might well find that your particular aircraft will not over-rev in straight flight at full power. In the simulator, instead, the rev counter was at least 120rpm above the red line anytime the aircraft was flying level for a few seconds, and so I did not even venture there (I saw the same happen in real flight, on a few different types of aircraft). But, if the red line is not a limitation, the ‘full speed’ strategy is certainly another option.
The importance of gliding
Although in the sim I initially decided to use simply the ‘drag option’, after flying a few sorties I ended up with the impression that the best way to deal with this scenario is probably a combination of the drag and PFL strategies. That is, if I had to do this in the real aircraft, I would probably sideslip (and even use flap) to prevent height gain, then I would position myself at the High Key point, and carry out a PFL from then on. But, of course, there will likely be quite a few variables that affect the scenario (primarily the presence of an upper ceiling, the red line limit, and the power setting at which the throttle is stuck), so that may mean that other strategies−or combinations thereof−may work better.
While the main strategy to deal with this malfunction may vary, there are a few firm points that I learned from the exercise. Firstly, I found speed control during an extreme sideslip near the upper end of the white arc more challenging than I anticipated, and at the same time an interesting training exercise that I had never tried. So, if you feel so inclined, try it yourself (in the sim though, I would not try it in the real aircraft as you risk damaging the flaps).
Second, the exercise made me think about how the engine should be shut down (again, you can find helpful hints on this in the April 2020 ‘QFI Logbook’ column). At some point, unless the throttle is jammed at a low setting, you will need to shut the engine down, but if you do it by cutting the mixture then you are keeping your options open. In case your glide turns out not the way you wanted it, simply going mixture rich should restart the engine, and you can use the starter if it fails to do so or if the prop has stopped turning.
A third learning point from this simulation is that I came to appreciate even more the importance of gliding. I used to think that for a ‘power pilot’ gliding is essentially only associated with the unlikely event of an engine failure or a fire. It turns out, that is not true: a jammed throttle almost inevitably ends up turning into a gliding exercise. And that reminded me of the importance to learn about the real gliding characteristics of your aircraft, including the difference between gliding with the engine idle (as we practise) and the engine off (as in a real emergency). The Cessna 172SP does not show a great difference between the two (at least in X-Plane) but this is something to keep in mind in case of a ‘real’ glide.
There are two additional options to deal with this malfunction that I did not mention among the main options, since they seem a bit extreme (however they could be considered and practised, at least in a simulator). The first ‘extra’ option if you find yourself with the throttle stuck at full power and, say, you are about to enter clouds, is obviously to shut down the engine (again with the mixture) to lose altitude, hoping that it will fire up when you need it. This would be an additional way (besides carb heat, one mag, flap and sideslipping) to lose altitude. However, I personally would not feel comfortable using it in the real aircraft, and that is why I didn’t use it in the sim.
A second ‘extra’ option (again, a bit extreme) is to make the aircraft descend in a semi-stalled condition while flying on the back of the drag curve. I practiced this in a clean configuration, and the end result is that the aircraft would still not descend. It probably would descend if the flap was down, but then I would rather use a sideslip, so I did not practise this latter scenario. In any case, the flying characteristics on the back of the drag curve may vary considerably across different aircraft, and so this may be an option in other types−although I wouldn’t feel comfortable doing this in the real aircraft during an emergency.
A final piece of learning for me was to realise how fast the aircraft climbs when starting from 130kt with the throttle jammed at full power. As I described earlier my initial reaction when the RPM started to overspeed was to climb into controlled airspace, so that I could decelerate and put full flap. I had good visibility, I saw no traffic above me, and I thought of my aircraft as a slow and rather innocuous traffic. But I was, at least partially, wrong: when I pulled up the VSI went off-scale (2,000fpm or more), turning me into a rather unpredictable dart in a busy airspace. I had warned ATC and taken a good lookout, but still the manoeuvre turned out to be more dangerous than I had anticipated. If I found myself in the same situation again I would give more thought to other options (including shutting down the engine), I would give ATC a good thirty seconds warning before starting my climb, and I would then climb at a lower rate to give myself and other traffic more time to avoid a conflict.
Another interesting session Eugenio. As you say I discussed the problem of stuck throttles in one of my columns (April 2020 issue), and you’ve put most of what I said then into practice here. However as always there are a few additional learning points.
This scenario is useful to discuss how you can alert ATC of your problems (and your impending airspace infringement) if you are short of time. The best option is to set the transponder to emergency (7700 – usually a simple action on modern avionics). That will immediately alert the various radar controllers to someone with a problem climbing into controlled airspace. True, they wouldn’t know quite why, but they could immediately start avoiding action for conflicting traffic. I appreciate your very high workload at this point, and whilst you absolutely must fly the aircraft first, setting the transponder to 7700 would have been my choice, and would have been high on my list of priorities.
A second option to communicate is the one you followed: alerting ATC on 121.5. This is probably slower (you still need to tell them what’s happening), and a bit less useful for the controllers as your aircraft won’t be highlighted on their screens. You would still have been on the Denham frequency when you realised the throttle was stuck, and a third option would have been to declare your emergency to them. This would have been the least useful course of action. Unfortunately, Denham have no control over traffic in the London CTR, or the TMA above. Consequently they would have to phone Swanwick and it takes a while to take to get to the right controller(s). So, in terms of communicating, setting the transponder to 7700 is probably the best action.
Regarding the actual malfunction, you eventually conclude that you are likely to end up using drag to contain the aircraft performance, and then a glide to land. And that’s absolutely what I would envisage. I often give a stuck throttle as a way of getting people to demonstrate a PFL to me. I would only really ever continue to use drag to position to the runway as you did initially when the cloud base would preclude a forced landing pattern. So, it’s useful to ponder the options you have to create this drag. You correctly reduced engine power as much as possible with one mag (and would have used carb heat if you had it) and created drag with flap. However it wasn’t enough and you resorted to sideslipping. Fine, but its tiring to fly for a prolonged period. Rather I would have tried reducing speed. You did consider this, but dismissed it as ‘extreme’. I personally don’t think it is. At 80kt, you are probably close to the bottom of the drag curve, in other words around the minimum drag speed. Obviously drag will increase markedly if you go faster than this speed, but it will also increase if you slow as the induced drag rises. I guess a Cessna 172 with full flap will stall in the mid 40’s, so you could safely have come back to the low 50’s. I’m certain that this would be vastly easier to fly, less disorienting and no less safe. I also think it could well have made all the difference – give it a try!
Having gone back to the airfield overhead, about to commence the engine off landing, you correctly decided to close the mixture so that you could start the engine again if you messed it up. However, you had already switched one magneto off. Although the engine should re-start on advancing the mixture, it is more likely to pick up cleanly with both mags selected on. Indeed many installations only have one impulse magneto, and if you have selected this one off it may be reluctant to go from a slow wind milling. Consequently, having retarded the mixture I would have selected the mags back to ‘both’.
Toward the end you talked about being uncomfortable with the idea of shutting the motor down to prevent entry into cloud. I absolutely would. In this case you’re in an uncontained climb, in an odd configuration. It would be hard enough to sort this out VMC as you have shown – let alone in cloud. There is no reason for it not to start again once clear below, and I think the risks of entering cloud with this problem far outweigh the dangers of a possible forced landing. - Charlie Huke
Five top lessons
1. Setting 7700 on the transponder is the fastest way to alert ATC of your troubles, if you need to infringe airspace and are short of time
2. One mag, carb heat, flaps, sideslipping and flying on the back of the drag curve are all options to descend and slow down
3. If you turn the engine off, do so by cutting the mixture (and turning the magnetos back to both). If you need to re-start the engine, just go mixture rich, and use the starter if needed
4. If you need to orbit to descend, it might be helpful to pick a prominent feature as a reference so as to reduce the sense of disorientation
5. Be careful with the use of flaps near the top end of the white arc (or beyond, if you get tempted): you may lose one flap and end up with less drag, a steady roll tendency, and a remaining flap that could detach and hit the tail
The worst case scenario
For some things X-Plane is unforgiving. One of the ‘unforgivable sins’ in X-Plane is to overspeed the flap, since they will be damaged or ripped away as soon as you hit the upper limit of the white arc. Needless to say, this showed me – during one of the flights I practiced for this article - how a jammed throttle could turn into a much worse emergency really quickly.
In that case I was trying to descend in an orbiting sideslip but, as mentioned, I found it difficult to control the speed accurately. As soon as the ASI hit 85kt (the top limit of the white arc), the left flap got ripped – first problem. With no flap on my left wing, the aircraft suddenly rolled and accelerated well beyond the white arc, damaging the right flap – second problem. The flap on the right wing remained attached but seemingly down to only a few degrees. In addition, now I was concerned about its integrity as well: I did not want to move it, because it might have detached and hit the tail. So, there I was, with a throttle jammed at full power, only partial drag from one flap, a strong roll tendency, and the threat of a flap about to separate and hit the elevator.
Long story short I decided to pick a nice big field and land there, but the flight was a very good reminder of how it’s better to increase your own safety buffers (eg the white arc) and not push your limits when you are already dealing with a malfunction. Because that is exactly what caused a challenging scenario to turn into a much more serious one.
The flight was flown using X-Plane 10, a software that is almost identical to the one approved by the FAA for formal flight training. X-Plane uses ‘blade element theory’, a process which allows for a realistic simulation of scenarios and effects at the edge of the flight envelope, including turbulence, stalls, the P-factor, and so forth. We chose the Cessna 172 (the main default option on X-Plane 10) as a popular light aircraft whose flying characteristics are accurately reproduced by the simulator. Hardware used for this simulation included control column, pedals, and throttle.