CV-440 virtual air crash investigation

[Alain_F355]

Just finished a short flight from Detroit to Chicago by ditching my CV-440 in lake Michigan a couple of miles away from Meigs Field. I was flying at 230 kts and 11.000 ft when suddenly the plane violently increased altitude and airspeed. At the same time both engines failed. I do not understand what causes this. As far as I know I was operating within the limits. I think Map was 38 and rpm was 2300. What did I do wrong?

[Drzook]

Not entirely sure what went wrong; I had a similar situation happen last week with my Cessna Conquest; at 20500 feet the altimeter all of a sudden went to 8500 feet which caused the autopilot to compensate which made the Conquest pitch up then stall out. I was able to shut off the autopilot and get it back to normal before anything went bad however. Just wondering, did you have autopilot engaged?

[Dangerousdave26]

What weather settings were you flying under?

[srgalahad]

I have encountered and researched a bunch of those anomalies. it's possible in FS9 to have a "super" high/low pressure area (seems to be a misread of a pressure report from a station). I saw a differential near EGLL that went from 30.21" to 28.65" and back in a 20 nm distance. In those cases the auto will see a sudden pressure change that says "too low" and tries to initiate a climb and ( with auto-throttle at least) a blast of power. With the CV-440 it may be that there is a damage module that led to rapid structural failure or other events.
Bet you can't recreate it

[Alain_F355]

I was not using autopilot under real world weather. There were areas of dense clouds and turbulence. I will try the flight again, if it was weather related it should not happen this time. When it does happen again I will now it is something I did wrong.

[Wild Bill Kelso]

It won't help you with this current crash investigation, but if this is going to happen again, having Autosave installed, might be useful:

Pete Dowson's freeware Autosave module

Automatic flight saving for FS98, FS2000, FS2002 & FS2004. Place AutoSave.dll in your FS98 Modules folder. It will save flights ("STN" files) at regular intervals, so you can retry that crashed landing or quickly recover from other problems. File size 10K.

http://flyawaysimulation.com/downloa...4-details.html

[middle]

FYI, autosave is also available at flightsim.com...type in Pete Dowson's name in the search box.

[zmike]

This event sounds so unrealistic that it is probably due to a false METAR (QNH) value imported into FS9 causing both altimeter error and an air density so low that the engines 'flamed out' before the engine automixture control or crew could respond. One solution to this ever present problem is for the provider of every relevant payware or freeware real weather add on to provide gross error trapping, followed by deletion from the download of the grossly erroneous METAR. The superior solution is for those who host 'real weather' for this community to perform the gross error trapping automatically before upload for use by this community. 'We' cannot otherwise avoid frequent and annoying recurrence of this problem.

However I suppose the original poster wanted someone to address the aircraft operating issues involved even if this was rather obviously a data error event and not virtual crew error. In which case it is worth explaining that the crew in question were pushing their luck pretty hard, (if using my 'realistic' freeware Convair flight dynamics available from calclassic.com.

The fact that the incident report did not include outside air temperature (OAT) at the time of the event suggests the reporter may not be fully aware how dangerous 'max cruising', in very cold air, in a fragile classic era propliner may be. The (current) supplied on screen handling notes for *all* varieties of Convair Liner hosted at calclassic.com contain the following warning.

<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
****************************
Convair CV440 Handling Notes 4/2009
****************************
WARNING - NO MACHMETER - NO BARBER POLE - OBSERVE ALL IAS RESTRICTIONS WITH CARE
Note: Calclassic panels load with a barber pole to assist inexperienced users. Click on ASI to remove this cheat mode.
<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<

Tom Gibson and I discussed this at some length before deciding that many flight simulation enthusiasts would not cope with the supplied levels of realism unless a (cheat mode) barber pole was provided by default, which real aircrew can remove with a single click before flight.

Whether or not the weather boundary event encountered was 'false' the crew in question were probably operating the aeroplane in a way which risked loss of control, potentially followed by structural failure, as soon as they encountered a 'weather boundary' (wind shear) or turbulence of 'ordinary' proportions.

The supplied CV44 on screen handling notes provide further warnings, power setting, by power setting;

****************************
Design Cruise power:
(Original purchaser cruise)
COWLS = CLOSED
MAP = 37
RPM = 2300
Plan 1300 PPH
Yields 250 KTAS at FL200
****************************
Max Cruise power:
Use to battle significant headwinds
> FL190 DO NOT EXCEED 212 KIAS
> FL170 DO NOT EXCEED 224 KIAS
COWLS = CLOSED
MAP = 39 inches
RPM = 2300
Plan 1400 PPH
Yields 253 KTAS at FL180
****************************

As soon as we exceed design cruise MAP in level flight, in very cold air, in any Convair Liner, we risk transonic shock. As the on screen handling notes illustrate the colder the air the lower the IAS at which we will induce shock. In International Standard Atmosphere (ISA) conditions the highest (coldest) flight level at which 230 KIAS is safe is FL160. In deep winter over Lake Michigan the crew took an unreasonable risk by exceeding design cruise power, even at 11,000 QNH. Mach number depends on temperature which in real weather may be very much colder than ISA. We must take great care when using more than design cruise MAP in very cold places and may need to avoid that risk by very conservative application of manifold pressure during both cruise and descent.

It is entirely possible that the crew in question pushed the aeroplane continuously beyond Mno (Mach - normal operating). When we decide to do that bad things can happen, even if we encounter entirely realistic weather boundaries. Unless the virtual crew understand how to calculate Mno from IAS and OAT via TAS in between they should never push the aeroplane beyond the last DO NOT EXCEED IAS cited in the Calclassic on screen handling notes. In this case flight simulation enthusiasts are unwise to push a CV44 beyond 224 KIAS in very cold weather at any altitude, unless they can calculate the DO NOT EXCEED value for the current weather using real world techniques.

If the crew did not need to battle a *significant* headwind, then pushing MAP above 34 inches during cruise was a waste of fuel anyway, and if they did need to battle a *significant* headwind they should have requested a lower cruising level. A lower level will usually diminish the headwind and it minimises risk of transonic shock by only ramping the aeroplane up to abusive IAS within warmer, lower altitude air.

Two general points.

Firstly most flight dynamics for large propeller driven aircraft available for download have no realistic dynamic limits of any kind encoded. This gives flight simulation enthusiasts a very false impression of how fragile classic era transport aircraft were (are) in real life, and flight sim enthusiasts become accustomed to ignoring Mach limits, or supposing classic era propliners, (and bombers or maritime patrol aircraft), don't have very restrictive Mach and profile drag limits. They do, and when 'realistic' FD are in use they can bite suddenly.

You almost certainly avoided the second problem, but it is probably worth explaining anyway. Many flight sim enthusiasts report in 'incident reports', or in 'bug reports' that they were flying an aeroplane of TYPE A based on the MDL and a repaint texture in use. When asked to supply the aircraft.cfg it often turns out that what they were actually flying was something else. What we are flying in MSFS is always and only the air file named at sim = filename.air in the relevant aircraft.cfg.

Unless that says sim=cv44 no one is flying a CV-440. Some flight sim enthusiasts reading this post, may for instance be flying a CV-440 MDL and textures, actually aliased from CV-340 flight dynamics (sim=cv34). This matters since as the relevant on screen handling notes disclose the CV34 has much lower transonic shock limits than the CV44. The CV34 on screen handling notes cite the following CV34 warnings, power setting, by power setting;

****************************
Design (Fast) Cruise power:
> FL200 DO NOT EXCEED 188 KIAS
> FL180 DO NOT EXCEED 200 KIAS
> FL160 DO NOT EXCEED 212 KIAS
COWLS = CLOSED
MAP = 37
RPM = 2300
Plan 1200 PPH
Yields 238 KTAS at FL190
****************************
Max Cruise power:
Use to battle significant headwinds
> FL200 DO NOT EXCEED 188 KIAS
> FL180 DO NOT EXCEED 200 KIAS
> FL160 DO NOT EXCEED 212 KIAS
COWLS = CLOSED
MAP = 39
RPM = 2300
Plan 1300 PPH
Yields 244 KTAS at FL180
****************************

This is all relevant because after a while, and after adding many repaints, MDLs and textures sometimes end up in the wrong folder using the wrong air files, often in combination with the wrong aircraft.cfg. That problem maximises when completely different aircraft share the same dimensions and use of the wrong FD does not cause false contact points, or other discrepancies which flight sim enthusiasts manage to notice. The need to match repaints to the appropriate 'base pack' is fully explained at calclassic.com, but not always as well explained in repaints uploaded elsewhere.

When flying CV34 Fight Dynamics (regardless of MDL or texture) transonic shock is a much bigger risk, or the same risk at much lower (warmer) altitudes. In very cold weather it is unwise to exceed even 212 KIAS in a CV34 at any altitude unless you can calculate the consequence.

The (cheat mode) barber pole supplied in every Calclassic base pack calculates the transonic shock consequence for flight sim enthusiasts, (unless they remove it). With the (cheat mode) barber pole present all we need to do is prevent the barber pole merging with the ASI needle and we will avoid loss of control due to transonic shock. The real world crew did (do) not have that safety feature and must calculate accordingly.

So we must ensure we have the latest CV44 files from Calclassic.com, including the latest VC, panels and gauges, and also ensure that we have any CV44 MDLs and textures aliased only from the correct 'base pack' else flight sim enthusiasts may be using CV44 on screen handling notes with the wrong Flight Dynamics, and then what the on screen handling notes assert is safe will actually be unsafe. The price of flight simulation realism is complexity of both installation and subsequent operation.

If anyone trying to operate this complex family of propliners safely has not read and understood the 'Martin and CV mini tutorial' supplied in every relevant 'base pack' I strongly advise them to do so.

Below I have extracted only the two most relevant sections,

>>>>>>>>>>>>>>>>>>>>>>>
BEGIN
HURRICANES OF DRAG.
Even flight simulator enthusiasts who have grasped that IAS measures profile drag often fail to associate the values they see on the ASI with their real world understanding of hurricane force winds. Even a category 1 hurricane = 64 KIAS starts to rip moderately well built structures apart. By the time we reach 98 KIAS we are battering the airframe with F2 (Fujita scale) tornado drag forces. F2 tornadoes can do bad things to even well engineered structures. Yet many FS enthusiasts think that even 98 KIAS is not much force to apply to a structure.

Hurricane force drag overlaps tornado force drag, but hurricanes ‘only’ go to category 5. At some level we know that category 5 hurricanes can do really bad things to well engineered structures but fail to associate 135 KIAS on the ASI with the destructive force of a category 5 hurricane. We should not be amazed that extending flaps might require us to reduce the force on the motor, the flaps, their hinges and support brackets, below category 5 hurricane force before trying to extend them. A category 5 hurricane equates to an F3 tornado. Aeroplanes and their fragile moving parts do not have exemption from the laws of nature.

Real aircraft designers must understand the equivalence above and flight simulation enthusiasts who desire realism must understand it and act accordingly.

......

Convair had no clue about transonic flight. They knew how to make an aeroplane that was much stronger than the Martin in nice warm air, but they had no idea how to delay transonic shock, or minimise the consequence of transonic shock in cold air. Yet AAL wanted to climb into cold high air where the performance envelope would be limited by transonic shock propensity and consequence. In real life when an aeroplane suffers transonic shock it does not normally suffer structural failure directly. It departs controlled flight and then as a direct consequence of the invoked out of control energy state, or as a consequence of attempted recovery, it suffers structural failure. MSFS nevertheless imposes structural failure directly.

Boeing had no idea how to solve such problems either, but they knew how to help aircrew avoid them. They provided a dynamic 'Mach Bug' on the ASI of B-29s, C-97 Stratofreighters, and their civilian derivative the B377 Stratocruiser. The Mach Bug functioned just like a later 'Barber Pole' (see 2008 Propliner Tutorial). The CAB should have insisted that Convair, (and all other producers of pressurised propliners), did likewise, but failed to do so. Instead they published IAS limits which varied with altitude. MSFS users who lack relevant licences will be unable to comply with the complex legislation which surrounds that method of Mach avoidance and we do not have a Pilot Not Flying or Navigator to do the necessary calculations.

Within the Calclassic handling notes I substitute simple IAS limit tables.

Those tables are potentially unsafe because they assume that the weather encountered will be invariant. They relate to the International Standard Atmosphere (ISA) which is the weather at 45 degrees North averaged across 365 days. The real weather never matches that average anywhere on any date. However, even in real life, IAS limits were invoked as a clumsy way to limit Mach. Consequently the published limits had to be conservative, but many airline pilots decided they were not conservative enough and refused to cruise at high level and high IAS in any weather in aeroplanes without a Mach Bug or Barber Pole.

In a CV-240 it was safe to apply max cruise power at FL150 most of the time and safe at a maximum of FL160 under ISA conditions, but only until mid cruise weight. Below mid cruise weight MCP imposes unsafe Mach, even at FL160, even at 45 degrees North. The design cruise power AAL had demanded simply imposed the same safety issue at higher level in colder air.

****************************
Design (first user) Cruise:
WARNING - NO MACHMETER or BARBER POLE
WARNING - MAY CAUSE STRUCTURAL FAILURE
> FL200 DO NOT EXCEED 182 KIAS
> FL180 DO NOT EXCEED 192 KIAS
> FL160 DO NOT EXCEED 202 KIAS
COWLS = CLOSED
MAP = 37
RPM = 2300
Plan 1250 PPH
Yields 249 KTAS at FL180 <<<<<<<<
****************************

..(in a CV24).. It was barely safe to apply Design Cruise Power at FL180 at 45N in average weather. It was unsafe below mid cruise weight. Many airlines soon promulgated 'Normal Cruise' power settings below the original design cruise settings. However many Convair aircrew were disinclined to employ more than econ cruise power.

It is vital that we understand that we may have more than enough power to cruise at high level, but that doing so will cause transonic shock. The updated 2009 CalClassic Notepad only calculates Operational Ceiling by testing whether sufficient power is available at our current weight in the current weather. It does *not* test whether that altitude is safe (warm enough). We should never climb above the maximum level cited in the handling notes for any given cruise power; and in cold places or on cold days in warm places even that level may be unsafe. That's how it is in real life too.

END
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>

Most flight sim enthusiasts fail to relate the IAS on the ASI to the wind force with which they are deliberately abusing the aeroplane's structure. The value of 230 KIAS in the submitted report is at the high end of the destructive forces in an F4 tornado. It is not a 'modest' force to apply to any public transport vehicle. While we try to rip every part of the aeroplane structure apart with almost F5 tornado forces we must be careful not to add transient transonic shock at different and asymmetric locations on the structure as different local curvatures and structure conjunctions cause different velocities and Mach numbers locally. If we fail to avoid localised transonic shock then loss of control will ensue, and may soon (suddenly) be followed by structural failure as IAS rises to F5 tornado forces of structural abuse during the loss of control event.

Provided the (cheat mode) barber pole is retained and we do not apply such excessive power that the ASI needle is almost superimposed on the supplied barber pole then a small margin for pilot error will be present at the next (unsmoothed) FS9 weather boundary, but we need to be especially careful when heading into colder air as the (cheat mode) barber pole will then creep ever lower down the ASI scale. In real life Pilot Not Flying must calculate the current IAS limit for the current OAT, else the real crew must avoid Mno by a significant margin, and in very cold air that is never consistent with more than economy cruise power in very powerful classic era propliners. The supplied on screen handling notes explain how to constrain the R-2800-CB17 engines of the CV44 to economy power.

****************************
Econ Cruise power :
(Second hand user cruise)
COWLS = CLOSED
MAP = 34
RPM = 2300
Plan 1200 PPH
Yields 239 KTAS at FL220
****************************

If the crew in question obtained the current files and documentation before flight, and had self briefed as required, they were unwise to disregard the supplied documentation. Their decision to exceed 34 inches MAP, and thus to exceed 224 KIAS, in what was probably air far below ISA temperature, high over Lake Michigan this January, was a contributory factor to the loss of control event, but the cause of the accident was probably injection of grossly false QNH data into the FS9 weather model, which the crew could not prevent.

One of the purposes of a flight simulator, when using sufficiently realistic flight dynamics, and matching handling notes, is to promote understanding of the complex relationships which limit the performance of the aeroplane in question in real life, as weather varies. We can understand why and how much it is constrained at 'high' cold latitudes, and why it is more constrained in winter. The limitations can be severe and in aeroplanes designed before anybody understood the effect of Mach number there may be no way to measure Mach number, even though measuring Mach number is vital, because excess Mach threatens loss of control, or even structural failure, at high IAS, in very cold air, in that aeroplane. This is not theoretical. It has caused the loss of many real aeroplanes.

In real life most public transport aeroplanes have easily more than enough power to induce both transonic shock and structural failure in level cruising flight as weather varies. The (real or sim) crew must restrict power applied accordingly. Most FS9 flight dynamics, (for large aeroplanes with propellers), just pretend otherwise to create 'uber-planes' which greatly outperform the real thing, to the delight of most flight sim users. However the Convair and Martin flight dynamics from calclassic.com replicate the real world structural limits (except Va) and require the virtual crew to avoid continuous cruising > Mno in real weather, by one means or another, else sooner or later they will induce a transonic shock event at a weather boundary that terminates their flight with a loss of control event.

It is only by compliance with the restrictions in the on screen handling notes, that we can experience a realistic 4D profile for the flight, arising from appropriate reaction to the weather encountered in real time. However on screen handling notes can only ever be abbreviations of concepts explained more fully in the lengthy, necessary, and supplied 'tutorials' which explain what the abbreviated handling notes mean, and how to comply, in more detail. Comprehensive self briefing for the flight is part of the realism process. If we do not fly a complex aircraft frequently we may need to undertake extensive self briefing before each infrequent flight.

It is always worth checking regularly to see whether updates are available. If anyone has Convair Liner handling notes which do not match the above then they have not installed the Spring 2009 Flight Dynamic and handling note updates. Other parts of these complex freeware simulation products have been updated more recently at calclassic.com where the update process is always ongoing.

FSAviator.

[adhockey]

Very interesting reading! I had no idea FSAviator's flight dynamics where THAT thought out.

[tgibson]

FSAviator,

Thanks again for your usual thorough treatment of an aspect of propliner flying that only a few understand and watch out for. I appreciate that!

The latest Convair Base Packs can be downloaded from http://www.calclassic.com/convair.htm

I also second the reading of the FSAviator Propliner Tutorial, available from my Tutorials page - it gives you much of what you need to understand this when flying all the propliners:

http://www.calclassic.com/propliner_tutorial.htm

Thanks,