AS332L2 Ditching off Shetland: 23rd August 2013
Join Date: Apr 2003
Location: in a figment of my imagination
Posts: 130
Likes: 0
Received 0 Likes
on
0 Posts
HC says
When I watched the 225 being introduced on to the NS, I could see that overnight pilots were being expected to move from 60's technology to 00's.
The level of automation of the 225 has been in airline use for a considerable number of years. Airlines have learned the hard way of the mistakes that can be made and have modified SOPs and cultures in light of that.
Should a fundamental review of SOPs for the L2 and 225 come about, I would highly recommend that either senior NS pilots go to an airline and see how they do things, or, better still, bring in airline people to offer advice.
As I mentioned earlier, I suspect that after this, there will be a tightening up of SOPs for use of automation, something which has mostly been at pilot's discretion up to now.
The level of automation of the 225 has been in airline use for a considerable number of years. Airlines have learned the hard way of the mistakes that can be made and have modified SOPs and cultures in light of that.
Should a fundamental review of SOPs for the L2 and 225 come about, I would highly recommend that either senior NS pilots go to an airline and see how they do things, or, better still, bring in airline people to offer advice.
Join Date: Mar 2001
Location: UK
Posts: 206
Likes: 0
Received 0 Likes
on
0 Posts
Lonewolf,
It seems, from the report, that the crew had not "selected" an airspeed, and were operating with just VS and Localiser coupled. In this instance the vertical mode is being controlled by cyclic, snd airspeed is being controlled manually by collective. If the collective is set in the right position airspeed is maintained but if not the airspeed will change as it is the only parameter available to the ac to maintain the chosen VS (in this case, the same happens in level flight if Alt is engaged).
If the collective is set too low, ie not producing enough 'power', then the aircraft will reduce airspeed to try and maintain the desired VS. If you are only slightly out then it will stabilise at lower speed. If you're a bit more than slightly too low the airspeed will keep reducing as the system continues to try to give you what you've asked for. However, once the speed passes Vy (which is usually around 68/70kts) the rate of speed drop can be very quick, until at an airspeed none of us can remember the autopilot gives up the ghost and the upper modes drop out.
In a 225, there are protections built in such that, should for some reason you elect to fly it 3-axis, at 60(?) kts the IAS upper mode automatically kicks in and the ac becomes 4 axis coupled.
Bit long winded but I hope that helps.
{Only intended as a description of the autopilot, not as any indication that this is what happened to G-WNSB!}
It seems, from the report, that the crew had not "selected" an airspeed, and were operating with just VS and Localiser coupled. In this instance the vertical mode is being controlled by cyclic, snd airspeed is being controlled manually by collective. If the collective is set in the right position airspeed is maintained but if not the airspeed will change as it is the only parameter available to the ac to maintain the chosen VS (in this case, the same happens in level flight if Alt is engaged).
If the collective is set too low, ie not producing enough 'power', then the aircraft will reduce airspeed to try and maintain the desired VS. If you are only slightly out then it will stabilise at lower speed. If you're a bit more than slightly too low the airspeed will keep reducing as the system continues to try to give you what you've asked for. However, once the speed passes Vy (which is usually around 68/70kts) the rate of speed drop can be very quick, until at an airspeed none of us can remember the autopilot gives up the ghost and the upper modes drop out.
In a 225, there are protections built in such that, should for some reason you elect to fly it 3-axis, at 60(?) kts the IAS upper mode automatically kicks in and the ac becomes 4 axis coupled.
Bit long winded but I hope that helps.
{Only intended as a description of the autopilot, not as any indication that this is what happened to G-WNSB!}
Last edited by obnoxio f*ckwit; 5th Sep 2013 at 17:48. Reason: To add last comment
OF, good explanation.
The 155B is the same, using 60 KIAS. The B1 refined it further to include rate of descent detection, so it may revert to 4-axis at higher speeds if required.
In a 225, there are protections built in such that, should for some reason you elect to fly it 3-axis, at 60(?) kts the IAS upper mode automatically kicks in and the ac becomes 4 axis coupled.
Join Date: May 2011
Location: Glasgow
Age: 40
Posts: 642
Likes: 0
Received 0 Likes
on
0 Posts
I believe the Airbus system is more simple, with the envelope protection system on roll / yaw / pitch only (and it has quite a different auto-thrust philosophy).
I would be curious to know if there are any 225 back to L2 issues.
225 back to 332L1 is pretty easy - It all goes manual. But L2 is a bit of hybrid, 'similar but different' scenario.
I remember converting onto the 225 and being shown how it will fly a coupled approach, even having lost almost everything. Very impressive, but my gut reaction was that I just wanted to take out the automatics and fly manually. At least that way, I knew who was controlling what, not that I have airspeed on the collective, the AP has Glideslope on the cyclic etc etc. Unless you're a TRE and forever in the Sim, I think you are liable to cock it up.
225 back to 332L1 is pretty easy - It all goes manual. But L2 is a bit of hybrid, 'similar but different' scenario.
I remember converting onto the 225 and being shown how it will fly a coupled approach, even having lost almost everything. Very impressive, but my gut reaction was that I just wanted to take out the automatics and fly manually. At least that way, I knew who was controlling what, not that I have airspeed on the collective, the AP has Glideslope on the cyclic etc etc. Unless you're a TRE and forever in the Sim, I think you are liable to cock it up.
TS, yes, and the BHL Nigeria 332L, the Bond 225, the KLM S76B, the BHL Australia 330J, G-TIGH 332L and the BIH S76A in the Fulmar field. Different means of getting to the same basic predicament - low power plus low airspeed equals high rate of descent!
Kaky C - I think one problem is that when the L2 was introduced (1992) all this talk of automation, automation dependancy etc etc hadn't been invented even by FW. Once you start down a bad path its harder to put it right. I recall doing an OPC in an L2, during the autorotation excercise I engaged IAS mode at 80 kts (Vy). The Training Captain / examiner took exception to this and said "you can't do that" but couldn't explain why not. Of course with IAS engaged, the whole excercise becomes much easier and you are guaranteed to maintain Vy until the flare, something which rarely seems to happen when people are flying it manually!
We were still banned from using the (admitedly limited) automation of the 332L during checks, being forced to fly the otherwise fully serviceable autopilot but without using the upper modes. Crazy! For the L2 I tried to introduce some use of automation during checks, but met a lot of resistance! So it was regarded as an extra set of sissy buttons that real men wouldn't use, and therefore the training given in its use was minimal, SOPs for its use didn't exist.
When we introduced the 225 (2005) we did try to look across at FW and hopefully took on board some of the lessons. For example, our SOPs require extensive calls and monitoring for use of the automation. One pilot calls presetting a parameter (eg ALT.A), the other pilot calls that he has crosschecked the setting. One pilot calls engaging a mode, the other calls what he now sees on the AFCS status display (ie what mode has actually been engaged). When there is an automatic transition between modes, eg an ILS localiser going from Armed to Engaged, that is called too.
We also show a video on automation dependancy - ie sometimes its best to drop down a level of automation. eg from coupled to ANAV on an overlay approach, drop down to HDG mode, or even drop down to manual flying (gasp!).
So I believe and hope that many of the bases are covered on the later fleets, its the earlier fleets that perhaps lag behind - although I have no idea what CHC's SOPs for automation use on the L2 are.
We were still banned from using the (admitedly limited) automation of the 332L during checks, being forced to fly the otherwise fully serviceable autopilot but without using the upper modes. Crazy! For the L2 I tried to introduce some use of automation during checks, but met a lot of resistance! So it was regarded as an extra set of sissy buttons that real men wouldn't use, and therefore the training given in its use was minimal, SOPs for its use didn't exist.
When we introduced the 225 (2005) we did try to look across at FW and hopefully took on board some of the lessons. For example, our SOPs require extensive calls and monitoring for use of the automation. One pilot calls presetting a parameter (eg ALT.A), the other pilot calls that he has crosschecked the setting. One pilot calls engaging a mode, the other calls what he now sees on the AFCS status display (ie what mode has actually been engaged). When there is an automatic transition between modes, eg an ILS localiser going from Armed to Engaged, that is called too.
We also show a video on automation dependancy - ie sometimes its best to drop down a level of automation. eg from coupled to ANAV on an overlay approach, drop down to HDG mode, or even drop down to manual flying (gasp!).
So I believe and hope that many of the bases are covered on the later fleets, its the earlier fleets that perhaps lag behind - although I have no idea what CHC's SOPs for automation use on the L2 are.
Join Date: May 2010
Location: UK
Age: 79
Posts: 1,086
Likes: 0
Received 0 Likes
on
0 Posts
Which takes us back to the subject of automation confusion which seems to be becoming a major issue nowadays in many different guises.
When the unexpected happens you dont always have time to figure out what it is doing, especially when low and slow.
When the unexpected happens you dont always have time to figure out what it is doing, especially when low and slow.
Avoid imitations
Join Date: Nov 2000
Location: Wandering the FIR and cyberspace often at highly unsociable times
Posts: 14,573
Received 422 Likes
on
222 Posts
Just read the latest AAIB report (link emailed to me).
My eyes were also drawn straight to the paragraph about the autopilot modes in use at the time.
As a "3 axis only" rotary autopilot user, (not Puma) I'd be most interested to know what the company SOP is, regarding which AP modes should normally be engaged during an ILS approach.
"VS" mode isn't one I would choose to use at that stage of flight, although I do often use it for the initial descent, or for a non-precision approach.
My eyes were also drawn straight to the paragraph about the autopilot modes in use at the time.
As a "3 axis only" rotary autopilot user, (not Puma) I'd be most interested to know what the company SOP is, regarding which AP modes should normally be engaged during an ILS approach.
"VS" mode isn't one I would choose to use at that stage of flight, although I do often use it for the initial descent, or for a non-precision approach.
ST, it was an NPA - localiser/dme
TAG yes I think there is some truth in that, but surely the answer is to improve training rather than to get rid of the automation?
Training is a limited resource in practice. Why is it that we spend hours doing single engine stuff when (turbine) engines never seem to fail in reality, and much less time (and its not mandatory) dealing with the likes of partial automation? When you look at the causes of accidents, it seems to me that we don't really train to address them, we spend hours training in accordance with the legal requirements, on stuff that is never ever put into practice and never causes accidents. Something wrong there, too many legislators living in the past of unreliable piston engines, single pilot mandraulic ops etc!
Training is a limited resource in practice. Why is it that we spend hours doing single engine stuff when (turbine) engines never seem to fail in reality, and much less time (and its not mandatory) dealing with the likes of partial automation? When you look at the causes of accidents, it seems to me that we don't really train to address them, we spend hours training in accordance with the legal requirements, on stuff that is never ever put into practice and never causes accidents. Something wrong there, too many legislators living in the past of unreliable piston engines, single pilot mandraulic ops etc!
Last edited by HeliComparator; 5th Sep 2013 at 17:44.
Ob Fwit ...
First off, deeply grateful for your explaining to me some of the nuances of the auto systems.
You will I hope forgive me for offering the opinion that this is counterintuitive to helicopter flying. Maybe it is due to me having flown and instructed in both fixed and rotary wing, but I have found that it is best to pick an airspeed as an approach speed, and adjust rate of descent with power (and then of course modest counter corrections with stick/cyclic) when trying to stay on a glide path. I do realize that once one has the landing spot in sight, the transition through translational lift and to touchdown requires adjustments.
I got into a discussion on the Asiana thread about pitch and power equalling performance over at R & N -- I had airline pilots inform me that they were pointing their noses and relying on power via automation to get their airspeed right. That is counterintuitive to how the coupling between pitch and power gats you performance, in a general sense. It took a while for me to understand that they were discussing flying in a mixed mode of hand and auto. While I think I better understand the issue now, I do not find the practice to make sense ... and the Asiana crash is a data point firmly in my favor.
Having taught a bit of instrument flying in rotary wing, a while back, I hold that the same basics apply to flying a helicopter on instruments in actual IFR as they do in fixed wing. Power in this case is collective, as you note.
I do understand the logic behind choosing a RoD in an effort to get a nice constant descent angle. Not a bad thing by itself. Your explanation of the mixed mode traps is an eye opener for me.
That's all well and good until one gets below max conserve airspeed, and/or into translation lift since one will begin feel the effects of being on the wrong side of the power curve.
While the details of "the power curve" are not identical between fixed wing and rotary wing, the result of "the bottom drops out from under you" is very similar. I've experienced it in both kinds, without bending metal, and feel comfortable making such comparisons from experience.
"You've got it, I'm outta here" says HAL, eh? 
It appears to me that maybe those "protections" are overrated.
To sum up:
Please don't beat me up too much for being a dinosaur. I am a bit disturbed with the traps that mixed modes provide when combined with a likely demand from management to use the auto features to the greatest extent possible. No doubt each company will have differences in SOP.
I guess even helicopter pilots must now ask themselves more frequently:
Are you flying the aircraft, or is the aircraft flying you?
Awaiting the more detailed findings with great interest.
First off, deeply grateful for your explaining to me some of the nuances of the auto systems.
It seems, from the report, that the crew had not "selected" an airspeed, and were operating with just VS and Localiser coupled. In this instance the vertical mode is being controlled by cyclic, snd airspeed is being controlled manually by collective.
I got into a discussion on the Asiana thread about pitch and power equalling performance over at R & N -- I had airline pilots inform me that they were pointing their noses and relying on power via automation to get their airspeed right. That is counterintuitive to how the coupling between pitch and power gats you performance, in a general sense. It took a while for me to understand that they were discussing flying in a mixed mode of hand and auto. While I think I better understand the issue now, I do not find the practice to make sense ... and the Asiana crash is a data point firmly in my favor.
Having taught a bit of instrument flying in rotary wing, a while back, I hold that the same basics apply to flying a helicopter on instruments in actual IFR as they do in fixed wing. Power in this case is collective, as you note.
I do understand the logic behind choosing a RoD in an effort to get a nice constant descent angle. Not a bad thing by itself. Your explanation of the mixed mode traps is an eye opener for me.
If the collective is set too low, ie not producing enough 'power', then the aircraft will reduce airspeed to try and maintain the desired VS.
While the details of "the power curve" are not identical between fixed wing and rotary wing, the result of "the bottom drops out from under you" is very similar. I've experienced it in both kinds, without bending metal, and feel comfortable making such comparisons from experience.
However, once the speed passes Vy (which is usually around 68/70kts) the rate of speed drop can be very quick, until at an airspeed none of us can remember the autopilot gives up the ghost and the upper modes drop out.
In a 225, there are protections built in such that, should for some reason you elect to fly it 3-axis, at 60(?) kts the IAS upper mode automatically kicks in and the ac becomes 4 axis coupled.
To sum up:
Please don't beat me up too much for being a dinosaur. I am a bit disturbed with the traps that mixed modes provide when combined with a likely demand from management to use the auto features to the greatest extent possible. No doubt each company will have differences in SOP.
I guess even helicopter pilots must now ask themselves more frequently:
Are you flying the aircraft, or is the aircraft flying you?
Awaiting the more detailed findings with great interest.
Avoid imitations
Join Date: Nov 2000
Location: Wandering the FIR and cyberspace often at highly unsociable times
Posts: 14,573
Received 422 Likes
on
222 Posts
212, thanks. My first look at this and I promise to try to keep up in future.....
Join Date: Mar 2001
Location: UK
Posts: 206
Likes: 0
Received 0 Likes
on
0 Posts
Lonewolf, you're right it is counterintuitive, it is however the way the L2 is set up. First upper mode selected goes on the cyclic, whether Alt or IAS. If the other one is selected, making it 4-axis, then Alt is on collective and IAS is on cyclic.
225 simplifies it, Alt is always collective, IAS is always cyclic
Exactly!!
225 simplifies it, Alt is always collective, IAS is always cyclic
That's all well and good until one gets below max conserve airspeed, and/or into translation lift since one will begin feel the effects of being on the wrong side of the power curve.
Join Date: Oct 2012
Location: Canada
Age: 53
Posts: 215
Likes: 0
Received 0 Likes
on
0 Posts
Pitts
The industry has the system forced upon it by the same people who conduct witch-hunts based on scant evidence, chase scapegoats as early in the process as possible, and then sue if a mistake is made. Add in the fact that aviation has a much higher profile than, for instance, road accidents or boats.
HC
italics added
Defies logic? I doubt it. But finding the root cause, separating the wheat from the chaff, will take more time than just stating the facts.
I can think of several very plausible and different sequences of events that may or may not have led to this result. Most are AP related but not all. We recently had an intermittent 26VAC bus Inverter issue on a 76C+. The result was seemingly random decoupling of the upper modes with minimal indication to the crew.
FFF
It certainly sounds like a good place to start.
These are the things the AAIB must now determine. They have a significant amount of data to sift through, including the CVR, to determine what the crew intended, what really happened, and how the difference between the 2 came about.
SAS
Sorry SAS, I have to disagree here too. Given the anti-Super Puma/225 witch-hunt that was going on, it was inevitable and, SADLY, necessary to tell the world that there was not an obvious design issue.
HC
Finally, something to discuss. OUR 76C+ also lacks some of the power limiting functionality I hear about in the 225. As result, in the highly compressed sim sessions we are now being given (bean counters pay attention with shame!) one can expect nearly every approach to end OEI. The standard I have seen is a consequent discouragement to use 4-axis to prevent troubles related to power limiting. We tend to fight the way we train, which is to then avoid 4-axis.
I personally dislike rigid automation SOP's. I prefer to 'mix it up' from flight to flight in order to explore the capabilities and limitations of both the system, and my knowledge of it. I therefore am comfortable choosing the modes that will best help me when things get busy or difficult, but not surprised when the system is degraded.
HC
As mentioned above a couple times, easy to miss on many types.
I won't quote you again, HC, on automation philosophy and how it has changed. As I have changed employers, continents, types, and years have passed I agree things changed from all hand flown rides, to all coupled, to something else. Inevitably there is an incident that highlights another chink in the armour and the world rushes to fix it. All this to say you are correct as to the progression.
-----------------
Again, my own personal defense is to, when allowed within the SOP's, use the various modes, in different combinations, and see how it reacts. This way I avoid, as much as possible, having to ask on a dark and stormy night: "what is it doing now?"
the industry has the system it deserves.
HC
They have said what happened in detail, but it takes time to work out why (since at first glance I suspect it defies logic).
Defies logic? I doubt it. But finding the root cause, separating the wheat from the chaff, will take more time than just stating the facts.
I can think of several very plausible and different sequences of events that may or may not have led to this result. Most are AP related but not all. We recently had an intermittent 26VAC bus Inverter issue on a 76C+. The result was seemingly random decoupling of the upper modes with minimal indication to the crew.
FFF
My posting a couple of pages back got trampled in a handbagging incident between Pitts and HC (Ladies, please!). Any thought on this as a theory?
http://flightsafety.org/hs/hs_may_june97.pdf
http://flightsafety.org/hs/hs_may_june97.pdf
These are the things the AAIB must now determine. They have a significant amount of data to sift through, including the CVR, to determine what the crew intended, what really happened, and how the difference between the 2 came about.
SAS
I shall take a bit of a Contrarian view on this Special AAIB Report.
My view is this report should not have been released.
The AAIB should have waited until they had digested the CVR data and decided what it meant....then released a Special Report that would go much further in describing the sequence of events and the time line of the CVR.
The Report as it reads now...tells very little and only generates more conjecture than it would otherwise had they waited to put out the CVR information.
Far too little information to justify a Report at this time.
My view is this report should not have been released.
The AAIB should have waited until they had digested the CVR data and decided what it meant....then released a Special Report that would go much further in describing the sequence of events and the time line of the CVR.
The Report as it reads now...tells very little and only generates more conjecture than it would otherwise had they waited to put out the CVR information.
Far too little information to justify a Report at this time.
HC
Because the L2 doesn't have the same power limiting functionality, its not normal to fly the L2 in 4-axis in the cruise or climb, but no reason not to for an approach. However much of the time the L2 will be being flown in 3 axis and perhaps that makes the fleet culture a bit different.
We are presuming of course that the collective axis was functional at the time, but that the crew chose not to use it.
As I mentioned earlier, I suspect that after this, there will be a tightening up of SOPs for use of automation, something which has mostly been at pilot's discretion up to now.
We are presuming of course that the collective axis was functional at the time, but that the crew chose not to use it.
As I mentioned earlier, I suspect that after this, there will be a tightening up of SOPs for use of automation, something which has mostly been at pilot's discretion up to now.
I personally dislike rigid automation SOP's. I prefer to 'mix it up' from flight to flight in order to explore the capabilities and limitations of both the system, and my knowledge of it. I therefore am comfortable choosing the modes that will best help me when things get busy or difficult, but not surprised when the system is degraded.
HC
there should be an aural warning anytime autopilot functionality was lost to the point the pilot had to take manual control, but so far that hasn't been implemented.
I won't quote you again, HC, on automation philosophy and how it has changed. As I have changed employers, continents, types, and years have passed I agree things changed from all hand flown rides, to all coupled, to something else. Inevitably there is an incident that highlights another chink in the armour and the world rushes to fix it. All this to say you are correct as to the progression.
-----------------
Again, my own personal defense is to, when allowed within the SOP's, use the various modes, in different combinations, and see how it reacts. This way I avoid, as much as possible, having to ask on a dark and stormy night: "what is it doing now?"
For example (on the 225), if you engage VS alone but have too little collective, the airspeed reduces until about 65 kts, then IAS automatically engages and stabilizes the speed at 65 kts (actual speed varies slightly according to dv/dt). That's fairly easy, but what if the collective channel is not working for some reason? ( its an MEL allowable defect, although I've never known it to be inop). Well then you have VS engaged on the cyclic with insufficient power to meet the needs. This time the IAS cannot engage automatically because the collective channel is inop. So what happens?
Well, what would you want to happen? You wouldn't want the IAS to go much below Vy, and that's exactly what happens. The VS mode starts to soft-limit the IAS to around 65 kts. So even though VS is still engaged according to the AFCS status zone, nothing engaged on the (inoperative) collective, the protections prevent the IAS getting too low.
So if, after that brief explanation of some of the protections, you still think they are over-rated, then perhaps you would tell us what sort of protections you think should be incorporated?
Last edited by HeliComparator; 5th Sep 2013 at 19:43.