Studying Lateral Stability in FS9

[Tom Goodrick]

I happened to find an article by Robert Blodget in a 1972 issue of FLYING that raised some interesting aspects of lateral stbility and its effects on flight safety. I can quote some portions of it if anyone is intersted. It deals with the general difficulty of poorly-trained pilots to keep an aircraft intact when descending through cloud layers or when flying at night when the artificial horizon acts up. This area of flight has alays interested me and my interest peaked back when JFK Jr flew into the ocean.

I never quite bought the idea that a "graveyard spiral" can kill you because it happens so gradually that any decent pilot will spot it before it gets out of hand. But I did some tests in FS9 that shook me up a bit. I have always found plenty of lateral stability in FS9, maybe a bit too much. But i ran a test in a plain old Piper Saratoga that showed a dangerous spiral could develop within a few seconds and could quickly become very difficult to recover from without losing the wings or tail surfaces.

There are two ways of diddling with lateral stability in FS9. The most obvious and direct is to edit the aircraft.cfg file and set roll stability to zero in the Flight Tuning section. You need to first get a plane in steady level flight before making this change. It is a little complicated because, after making and saving the edit, you have to switch to a different plane (while paused) and then switch back. Then you turn off PAUSE and watch what happens. It will quickly go into a vertical spin. It is not clear whether the aircraft will break up before or after it hits the ground.

So next you can let back in some lateral stability by increasing the amount of lateral stability to .05, .10, .15, .20, etc in the Flight Tuning section and checking the controllability and dynamic stability.

The most complicated way to study lateral stability is by changing the roll stability derivatives: Cl_beta, Cl_p, Cn_beta and Cn_p. Note that Cn is the yaw moment but any yaw disturbance quickly becomes a roll disturbance and vice versa. Moments of inertia must be set correctly because they take part in the interference between the axes.

But another thing that can cause problems in roll and yaw is the fuel balance. With fuel carried in the wings in almost all aircraft (except the Piper Cub), if you forget to switch tanks and let a large imbalance build up, a quick entry to a divergent spiral can happen when you turn off the autopilot. You might think all is fine because the autopilot is holding the plane nicely; then you turn off the autopilot to make an approach and all heck breacks loose.

In a test of this, the aircraft entered a spiral just as fast as it had with zero stability. But I was able to recover after losing 1300 ft.

Try this stuff with your favorite planes.

[Tom Goodrick]

It would be nice to duplicate the fuel tank switches in real aircraft. But FS has a feature that makes this difficult even if you can design your own gauges. The command in FS "Select_Right_Fuel_Tank" actually selects any tank on the right starting at least with the right aux tank and perhaps with the right tip tank (if there is one). This is the command you would use in any gauge to select the right fuel tank. The FS method when you select a main tank is to use the aux tank first and then the main tank in that automatic sequence.

If you do have a selector that can point to aux and main separately, you should select aux first and use it up so that when you selct the main, only the main tank can be selected. But this does not let you use a little from each tank as you might like.

In real aircraft, you selct each tank with a rotating switch. Care is generally taken not to switch through "Off" or through an empty tank which can kill the engine at least temporarily. For example on the Piper PA-32 series (Cherokee 6 or Saratoga), there are switch positions, left to right, of OFF, Left Aux, Left Mani, Right Main, Right Aux. The preferred method is to start with the fullest tank on the left and switch through to the right using each tank for a chosen time like 1/2 hour. Then switch back to the left. The goal is to use each tank a little at a time to maintain balance.

I am now working with the DC-4 that adds a tip tank on each side with two tanks in the fuselage. default useage starts with the tip tanks and the Center2 tank, then the aux tanks and the Center1 tank, finally the main tanks, using both sides at the same time for the four engines.

The GA single engine aircraft have either of two selectors: the Cessna Selector or the Mooney Selector. The Cessna Selector has left, both and right positions which is realistic. But it lets you be lazy and select both and forget about it, at least in FS. I can say that, in the real case, your payload balance will determine how off-balance the fuel load will become if you just selct "both" and sit there. I suspect the same applies to FS though I have not flown that many long trips in singles. The Mooney Selector has only left and right positions so this is best for most aircraft and keeps you on your toes to switch tanks regularly to keep the fuel load balanced.

[Tom Goodrick]

In my GA twin aircraft, I have used the same Baron 58 fuel tank selector. This is neat because it allows you to feed either engine from the tanks on the opposite side instead of the same side as the engine. I often find after taking off and climbing out that the tank on one side has been used more. Then I must select the opposite side tank for a brief time to get the fuel closer to balance. This is a realistic feature that keeps you working well into the cruise.

[Bill Von Sennet]

When I switch off the autopilot on the approach, I often have to correct for an uneven passenger load. If I have an extra 200 lbs on the left side of the plane, should I start out with more fuel on the right side to compensate? Since the fuel tanks are further from the centerline, maybe I only need 100 lbs more on the right. Is there a formula for this, or just trial and error?

[Tom Goodrick]

Such problems occur frequently but are random in nature so there is no sense trying for a standardized solution. If the plane has rudder trim or aileron trim (rudder trim is preferred), take the plane off autopilot early enough so you have enough time to trim out the imbalance. This requires trial and error because you inevitably over shoot the correction and have to make several little adjustments. If you cannot trim out the problem, you deal with it manually. It is something to practice.

Imagine the fun for those of us learning to fly in little Cessna 150's where our own weight was always offcenter. We had no rudder or aileron trim. The imbalance caused by our off center bodies caused more imbalance in fuel use. We always flew with strange attitudes. You just learn to handle it.

The danger in all this is in the misconception that our airplanes are all nice and stable and will recover naturally from any turn to fly straight and level. The fact is you have to pay attention at all times and be ready to take corrective action. In a Cessna 150 I hgad enough time to turn around and get my lunch bag from the back seat before the aircraft went out of control or took any particularly unusual attitude.

A frequent problem I see in FS when flying turboprops is that I will shut down and then restart while parked. I always use the Ctrl-E method. Then I forget to check that the generator is switched on and make a take off. 2 or 3 minutes later the fancy digital panel goes dark. I say a bad word at myself and turn on the generators. Then several minutes later I notice the plane is no longer on course. The loss of power means the autopilot shut off and I forgot to turn it on after getting the generator back on line. This is when a lateral stability problem could become an issue. So far it hasn't. The plane just wanders a bit off course and off altitude. Of course, in some cases those can be serious problems.

[n7rg]

HI Tom
I thought you might find this a good read

calclassic.proboards.com/index.cgi?board=fsaviator&action=display&thread=1987

look forward to your thought

Rich

[Tom Goodrick]

That is such a huge article, I don't know to what part of it you want me to respond. Most of it is true and correct. I was bothered a little by this statement:

"What we are flying in MSFS is always and only the air file named at sim = filename.air in the relevant aircraft.cfg."

which I'll attribute to FSAVIATOR.

More precisely, the FD of the aircraft are defined mainly by the aircraft.cfg file which supercedes the items in the .air file. The two files act in combination with anything not covered in the aircraft.cfg file influenced directly by the .air file. Both the .air file and the aircraft.cfg file aerodynamic specifications must be as set by the FD developer unless one has better information available. But the goal is always to give accurate performance for the aircraft.

As for the references to Real Weather problems, yes, it is common to find problems but i do not agree that all those problems can be found by scanning the METAR data because the problems often come from interpolation of the METAR data in regions between airports.

Also, winds aloft data are not included in METAR data as far as I know, but rather are separately measured and reported with significant interpolation by MSFS. Thus you can see high wind values rotating quickly around the compass or with abrupt changes at the invisible boundaries between data regions.

For this reason I always fly cross countries with the structural failure effect turned off. this prevents "bumps" that the autopilot cannot handle.

As far as the discussion on Mach number is concerned, it is very relevant to flight of pressurized aircraft, even with old radial piston airliners. Mach number depends on true airspeed and temperature. When the temperature is lower than usual (It is always cold at 20,000 ft but sometimes it is colder.), you can experience Mach effects that can ruin your day if you are cruising at high indicated airspeeds. I leave a Mach meter on my panels for pressurized aircraft. Mach effects, if properly set by the developer in the .air file, can degrade stability and control. Few .air files I have seen have this level of detail set correctly. To set it correctly, a developer would have to have access to real aircraft development and certification testing. I can only guess at parameter values for the appropriate equations. I seldom bother with this.

It is a good practice to observe a VNE KIAS that is safe at all altitudes within the operating envelope. For an estimate of this, multiply the clean stall speed by 2.5 or 2.0 if in turbulent conditions.

Is there a specific element of the article on which you want comments?

[louross]

The winds aloft and other WX probs can be very well rectified, but it costs. You need to purchase both Active Sky and have a registered version of FSUIPC.

lr.

[Tom Goodrick]

I learned a long time ago not to trust those addons. I take what I can get out of the standard "Real Weather" available . It has its flaws but can be relied upon in most cases. I was just pointing out that wind aloft does not have the same level of accuracy as local weather at an airport.

[louross]

That's a good point. You need to be real careful when buying add-ons. Some have no real use at all. Some are just bad quality. I would do very little, if any simming if I hadn't purchaased Active Sky, Flight Deck Companion, Radar Contact, and Ultimate Traffic many years ago. Haven't upgraded to the newest ones due to lack of money. Fortunateley I could buy them back then, like 1995.

lr.

[n7rg]

Ok
I had no idea that weather played that big of a roll FS.
thanks for your input
rich

[louross]

OOOPPS!! Not 1995, shud read 2005

In case you're not aware, you can get the aviation weather page at:

aviationweather.gov/

and, again, if anyone doesn't know, it is coded, so you need to translate it. METAR is the actual airport wx at the time it was taken- Winds Aloft is not the actual wind, they are forecasted figures. It is currently March 06, 1355 Z; you can get a forecasted wind that is valid after 1200 Z, March 07. Obviously it won't be exactly at those figures when the time comes.

As mentioned, the MS wx engine has it's problems. Some of these can be overcome if you have a registered copy of FSUIPC- there is a small cost involved. You can get the free copy, but most of the functions are unavailable. The wx engine by Active Sky is very very good. There is a definite difference between the 2 programs (FS and AS), but, there is also a cost involved.
lr.

[Tom Goodrick]

I think it is interesting to use FS to see how different aircraft perform the job of getting us and a few friends between two airports at least 300 nm apart. Any distance less than this can be easily driven in a car. But it would make a long day if you drove that distance, did something, and then returned. An airplane makes a day trip over this distance very easy. It doesn't even have to be a fancy airplane.

But weather is a big deal in real flying so I like to include it in trips I make in FS. We have this RW feature that we can easily use in our flights. But it is important to know what it is. RW starts with the periodic (every 15 minutes at some stations) weather reports at most airports in the world. These reports give sky cover, precipitation, temperature and surface winds. I live a few miles from KHSV, under the initial approach route to runway 18L used mostly by small aircraft and non-scheduled flights (although 747's and the Concord have landed there). I can see the weather outside and experience the wind and the temperature. Then I load RW and look at KHSV conditions. What I see in FS is often very close to what I see outside the house within a 15 minute interval.

Wind aloft data is separately measuered and reported using a smaller number of stations and using special techniques such as balloon tracking and LIDAR which involves reading motion of airbourne ions with lasers.

The folks at MS have created a program that merges the various weather reports together to create a continuous approximation of the weather. For the most part. their methods produce a reasonable weather system. I have tested the distributuion of temperature and pressure within the FS atmosphere and it agrees with several actual measurements adjusted to "standard" conditions.

But we don't fly in the weather at the airports unless we just practice circuits at one airport. When we fly between airports at reasonable altitudes (to 51,000 ft), we see a variety of weather. We can accept that the data at each airport is reasonably accurate for the reportiing time (within 15 minutes of when you download it). But how accurate is the data we see when flying at altitude between the airports?

The process that gives us this data is called "interpolation." I first learned about it when doing surveys of the wakes behind aircraft where parachutes had to inflate and exert a certain amount of drag to extract large cargo loads from the aircraft. I did wind tunnel tests measuring the total pressure at a set of points behaind the aircraft. I knew that the data at each point was accurate. But to get an estimate of the drag of a parachute operating in that region, I had to find a continuous distribution of the pressure throughout the region. I had to "interpolate" between the points where the pressure was measured. I checked with the resident "staff mathematician" for the Army laboratory where I worked and he found a few interpolation methods but they all were inappropriate to my problem. So I set out to develop such a function. It took a few months but I ended up with a "bicubic spline function" that could be applied to the measured data. It was simple enough so we could spot check calculations although the data would be processed on a large computer. (This was in the 1960's.) The function could be integrated to get an average pressure over an area such as the projection of a parachute. Some of the problems I had to overcome were the tendencies of many such functions to "blow up" at certain points between the data values. My method worked well and has been used in several applications.

While I was at NASA Marshall inthe 1990's, I dug up the old bicubic spline function to merge two sets of data also measured in the 1960's but unreconciled until I applied my method. They had force measurements on rocket models climbing through the atmosphere from one type of test and they had pressure mesurements from a set of holes in the sides of rocket models from other tests. I was able to show the tests actually agreed. Of course it was only of academic interest at the time but may now be used on new rockets.

So I understand about interpolation functions. Think of an interpolation function as a blanket you lay over a set of spikes (the data) to form a smooth surface between the spikes. Now when flying jets above 30,000 ft over the central US in FS, I frequently see the jetstream with speeds in excess of 150 knots. It is common near the edges of these jetstreams to see "rips or ripples in the fabric of the interpolation." Sometimes these ripples will show up as rapidly changing direction of the 150 knot wind. If you have structural damage turned on, these can ruin your day. Without allowing that damage to crash the airplane, the plane will usually just wobble and move around a bit but the flight will resume.

But you will see some exercises in lateral stability - the subject of this thread.

[n7rg]

for those that like to know the Winds Aloft, that is on the sim at any given time.
here A gauge that will give that info. wawind.zip at flightsim.com

[Tom Goodrick]

Do you mean a gauge that tells the wind speed and direction at your current altitude and position? I made one of those and have it on all my panels. That is why I know what the wind is doing at all times in flight. It is a normal indication on all Garmin 430 and 530 GPS nav gauges but it was left off of the Garmin 500 given to us in FS9. You cannot get that data in real life without having a GPS nav system. It compares airspeed and heading with measured groundspeed and track to get wind speed and direction.

Returning now to the topic of Lateral Stability, I did some more tests in a Saratoga with tanks run dry during cruise. First i did the tests with the reduced value, 0.6, of the Lateral Stability Scalar in the Flight Tuning section. I had only one occupant, the 220 lb pilot in the left front seat.

I set up steady cruise on autopilot at 5500 ft with full tanks and then zeroed the right aux tank. The autopilot held it steady though there was a small sideslip angle. Then I turned off the autopilot. It took 42 seconds to get into serious trouble with a 60 degree bank, an airspeed near VNE of 181 KIAS and a descent rate of -2798 fpm. It would be difficult to recover from that without bending something.

Next from steady cruise I zeroed the right main tank as well. Again the autopilot held a steady cruise with the selector on the left tanks. But when i turned off the autopilot, it only took 16 seconds to get into very serious trouble: 70 degree bank, 172 KIAS and -4999 fmp.

I repeated these tests with the Lateral Stability Scalar reset to 1.0 where the aircraft handles very well and realistically according ot all reports. When the autopilot was turned off with the empty right aux tank, after 42 seconds the bank was only 40 degrees, the airspeed had only increased to 166 KIAS and the descent rate was 961 fpm. Recovery would not be too difficult. With the right main tank also empty, the effect was also about half as bad as before. At 16 seconds the bank was 44 degrees, the airspeed was 157 KIAS and the descent rate was 2104 fpm.

The bottom line is that running a tank dry is not a good idea. But, with good normal lateral stability it is survivable if you do the right things: level the wings and then gradually level the nose. From 5500 ft you would be safe by 2500 ft. The imbalance of an empty tank can be managed in manual control if you know what has happened and counter the effect with rudder trim and with careful roll inputs. The FS autopilot can handle the situation.