YES There IS ICING in FS9

[johnl]

Firstly, apologies to TG for pinching his intended thread title.

Secondly, Charles "Dutch" Owen has been investigating FS9/FSX icing for some time, and has released a digital gauge, (icev10.zip on www.flightsim.com) which not only provides a 3-level threat warning of wing icing, but can also simulate the effect of freezing rain, thereby making the whole situation much more "interesting".

I haven't installed the gauge yet, but I definitely recommend a thorough read of the icing.txt file which accompanies it before leaving the ground in it's company.

[Tom Goodrick]

Okay, I suppose there is enough evidence of icing in FS9 to warrant at least a discussion. The gauge he describes (with code) is only for use with the DC-3. In the code, he uses changes in the weight to create estimates of the ice. He makes several local variables like (L:IceWeight, pounds).

You have to work very hard to get an icing effect that is significant. Rather than turning back to land at the same airport from which you took off, you hae to be sufficiently bull-headed to plow on through the soup until it eventually takes all your power and adds enough weight. Then you find you are not flying anymore.

That gauge has some code that, after detecting ice by weight growth, changes the setting of the elevator. So it will give you an exaggerated aircraft response.

That's one way to do it!

[Tom Goodrick]

Here is a report on my efforts. The efforts were flawed. But I tried the same test again with improved weather specs and still ran out of ice just above 6000 ft allowing me to climb and escape.

The most important thing you can learn from this is how to set up icing and how to tell when you are in it.

The custom weather in FS9 can be set to give icing conditions that will interfere with flight. The extent of the interference depends on the aircraft and what you do when flying in the icing area. When I flew a Cessna Skylane in this stuff, it almost knocked me out of the air before I could climb above the icing level set at 9000 ft. But I did manage to get up there to 9200 ft eventually and then cruise uneventually to Huntsville, AL, 436 nm away where I landed through the same stuff but with no indication of icing.

To set custom weather for icing, click on Weather, User-defined weather and click on the Customize Weather button. Click on the top line button which selects all stations for this new weather. Next click on the button at the bottom for Advanced Weather. The next page sets clouds. This is where both precip and icing can occur. i set cloud type to Stratus, Cloud coverage to Overcast, tops at 9000 ft and base at 2000. I also set turbulence to light and icing to severe.I set Precipitation to snow and the base altitude of the snow at 680 msl (elevation of Aurora, IL, KARR. (I chose this airport bevcause on the day I did this snow was forecast to cover the Chicago Area. Aurora is a handy airport for entry into the Chicago Area.)

Note that in Advanced weather, we have to also set Wind, Temp/pressure and Visibility on separate pages. Click on Wind next.I set three altitude regions for winds. You can, of course, do whatever you wish.

0 to 2500 ft 8 kts fr 71deg
6000 ft 35 kts fr 36 deg (moderate shear, light turbulence) gusts to 38
10000 ft 50 kts fr 315 deg, gusts to 51, shear gradual, light turbulence

I think we are setting only the exact wind at the specified altitude with the wind between lines varying according to the shear setting.

Next we set the temperatures making sure that ice can form where we want it. We set both the outside temp and the dew point at several altitudes.

I set 22/19 F at 750 ft,
15/15 at 5000 ft,
5/5 F at 9000 ft,
-14/-20 F at 21000 ft

I expected ice from the surface to 9000 ft. That's what I found on takeoff.

I set visibility at 1 mi from the surface to 6400 ft. Later when I moved the cloud tops up to 9000 ft I forgot to change this. You do have to try to make these different pages consistent.

The extent of the icing area and of the regions and altitude levels to which weather is assigned is not easily determined. I would say it does not work as prescribed. I originally set it up for ice and overcast to end at 7000 ft. Then I changed it to 9000 ft. But the cloud tops have remained at 7000 ft though the icing continued to 9000 ft coming out of Aurora. The same clouds, winds, temperature and visibilty were encountered during my descent at Huntsville. But I saw no eveidence of icing during the descent. I leveled off at 3000 ft about 30 nm out.

Here is how I knew I had ice. The paper about FS9 icing mentioned Ice Weightg as a parameter in FS9. I could not find it listed in parameters in the SDK. That is not surprising. The SDK has a few other omissions. But I have a digital weight gauge so I decided to watch that and jot down values.

I started with a gross weight of 2751 lbs (solo 200 lb pilot with 40 lb bag) parked at 11:33am. At 11:37 when I made the takeoff, the weight was 2748 lb. (I have a weight gauge on the panel though that is unrealistic in this case. It can't weigh the ice but that is the only way to account for weight gain. Fuel burn is decreasing the weight.) At 1139 while climbing, the weight was back to 2752. Here's the log of data:
Time__Weight__Altitude__KIAS
1141__2757____3700______126
1144__2764____4940______109
1147__2774____6340______90__Levelling at 6000 to get speed
1154__2786____6000______95__Adjusted mixture for more power to climb
1157__2775____6680______100__above overcast
1201__2752____7000______124__Power reduced to 79%.
Here I stopped to check the weather settings. I saw the page for icing to 9000 ft and clicked OK. I had thought I'd checked it OK before but maybe not.
1205__2753__7000 ft__112.5__More ice coming now.
1214__2778__8100 ft__72__Trying for 300 fpm getting -60fpm

Reached 8200 ft tried diving to get speed, stalled but recovered and climbed back. Lost some ice.
1219__2760__8260 ft__93.6__+323 fpm
1220__2752__8760 ft__96.41__+319 fpm
1223__2736__9200 ft__115.5__level
1225__2724__9200 ft__119.2__level with a strong tailwind.

About two hours later I landed at Huntsville. There was no weight gain during the approach even though I held level at 3000 ft for several minutes.

The idea of using weight to indicate icing is a little strange becausethere is no way to determine this weight in flight. I use a "computer output" on my panels as some airliners have today. On the ground these actually measure and report weight and CG position. But in flight they estimate weight based on measured fuel burn. There is no way for them to measure or to estimate the ice weight.

This was not a nice clean test. I messed up a couple of times. It is not even clear if the Skylane could have made it to 9200 ft without help from my incomplete specification. But it does show there is an icing effect in FS9. Now we need to investigate it more.

[hanspetter]

The focus seems to be on increased weight only. Another obvious parameter would be the changes in airfoil shape as the ice builds up. In other words, not just a heavier aircraft but also one with less efficient airfoils.

[Tom Goodrick]

What kind of parameter is this: "obvious parameter would be the changes in airfoil shape..." A parameter is a number. There is no measureable number in FS to reflect this.

No! The emphasis is NOT on weight as a flight problem but we use it simply as an indicator of the extent of icing. The amount of weight gain would have no effect on flight at all. The Cessna 182 was well below its max gross. There was only one pilot and a 40 lbs of clothes and camera gear on board. I only list the weight to show that it is, in fact, a gain! This gain does not happen for any other reason.

The flight problems are a loss of power, added drag, and mushy handling if you have to try to recover from a stall. It is hard to identify particular aspects of handling that change with ice in FS but there are some. Here are two more data sets. in each case I am taking off into "known icing" but think I can climb above it (ie- above 9,000 ft) to continue safely with my trip.

Cessna 182S Skylane, one 200 lb person and 40 lb baggage:
Heading south to Huntsville, AL

Time__CG%___GWT___ALT____KIAS___FPM
1149_17.15__2751___720____0______0
1153_17.13__2749___2680___109____822
1155_17.14__2754___3900___110____522__Leaning
1157_17.15__2760___4860___103____522
1158_17.16__2763___5440___94_____530
1200_17.17__2768___6300___67_____183 (Etrim at limit)
1204_17.19__2783___6100___96_____0
1209_17.16__2774___6280___96_____259 Out of ice!

(Wx setup adjusted): Bonanza A36, same payload:
Heading west to Rochester, MN
Time__GWT___ALT___KIAS
1141__3679__?_____134
1144__3689__5400__115
1146__3690__6880__118
1148__3694__7600__113
1150__3704__8700___96
1151__3704__9080___87
1153__3704__9740___78__42 kts GS
1158__3681__10000__134__103 kts GS__Wind 315 at 49

In each case I am climbing using the autopilot. I have set the climb rate a bit lowr than usual, anticipating problems. When problems occur, I adjust the climb rate lower. In the first case with the 182 (previous thread) I found I could sustain a safe climb with ice at 300 fpm. That was better than a descent! By using the autopilot, I avoid problems due to ham-handed upsets. But the danger is that the autopilot can drive the trim to the upper limit trying to hold a climb rate beyond its capability. This is what you must watch.

I have found a set of binary flags identified as "Ambient Precip State" that I can read with a gauge. This would be more realistic than reading weight. I'll experiment with this value, trying to mask out the bits to see exactly what it says in various precip conditions. This will take a while.

[hanspetter]

Yes, a parameter is a number. Would Oswald efficiency be a candidate? I understand that FS tries to mimick the effect of icing by reducing power and adding drag. Then there's mushy flight controls -- how is that accomplished in terms of FS flight dynamics? My guess is that some curves in the air file are shifted.

When I mentioned "changes in airfoil shape" I did so without considering whether FS could simulate the effect in any direct way. However, I understand that this thread focuses on what FS actually can do rather than how it might have been programmed to simulate icing

[flaminghotsauce]

Gee, ya'd think Microsoft with all it's billions of programmers crammed into that little closet of theirs could have figured that there would be realism enthusiasts on the other side of their cellar door that would want real icing. I wonder if they know there are grass and trees outside?

A simple "Icing" weather under the "Hard" setting somewhere? Here we are forced to figure this stuff out....

[Tom Goodrick]

Their sim certainly is not as real as it should be.

I made a gauge that reads the Ambient Precip State flag set. Because only one flag is set at one time, it is rather easy to just look at the value of the flag set and use that value as a code.

2 is for clear.
4 is for rain
8 is for snow.

I have seen no other values. When it shows snow, there may or may not be icing. on my setup icing at Aurora, IL, it shows 8 while icing adds weight. But in the preset weather "Heavy Snows" it shows 8 but there is no weight gain indicating icing.

I tried setting up the gauge so it actually displays "Clear" "Rain" "Snow" but got bogged down in XML. The number is easy to use as a code since these are the only values.

Tonight, I took off from ARR with a loaded C182 expecting to see some problems. I stayed at 5,000 ft heading south. At about 10 nm, just after leveling at 5,000 ft, the precip gauge switched to 2 indicating no snow. The weight gain stopped too. No idea why.

This gauge needs a little more work.

[Tom Goodrick]

Here is the XML code for the Precip gauge I am using. To make your own gauge, simply copy the code and paste it into a Notepad page. Name the file Precip.txt. Store it in your Gauges folder. Then change the file name to Precip.xml and leave it in the same directory. (You will need the file "Box.bmp" in the same directory. If you already use some of my .xml files in a Digital subdirectory, put the two files in that directory where Box.bmp probably already resides.)

<Gauge Name="AMBIENT PRECIP STATE" Version="1.0">
<Image Name="Box.bmp" ImageSizes="64,48"/>
<Element>
<Position X="14" Y="11" />
<Visible>(A:Circuit general panel on, bool)</Visible>
<Visible>(G:Var2) 0 == (A:Avionics master switch, bool) &&</Visible>
<Element>
<Text X="40" Y="14" Bright="Yes" Length="5" Font="Arial" Color="#00FFFF" Adjust="Center" VerticalAdjust="Center" Multiline="No" Fixed="No">
<String>%((A:AMBIENT PRECIP STATE, number))%!5.0f!</String>
</Text>
</Element>
</Element>
<Element>
<Position X="11" Y="32" />
<Element>
<Text X="50" Y="12" Bright="Yes" Length="8" Font="Arial Bold" Color="#114444" Adjust="Left" VerticalAdjust="center" Multiline="No">
<String>Precip</String>
</Text>
</Element>
</Element>
<Mouse>
<Tooltip>%Ambient Precip State</Tooltip>
</Mouse>
</Gauge>

This creates a gauge with a number in a slot and the name "Precip" below the slot on a white rectangular background.

The numbers are code numbers for the various precipitation states in FS. So far I have found 2 is for Clear, 4 is for Rain and 8 is for snow. I have not seen any combinations (for example 12 could be for rain and snow, mixed). Hail and fog might be other posibilities.

The gauge does not show if icing is occurring.

[Allen Peterson]

The code is correct for the .xml version, but the .txt version will need a correction. The && in the line:

<Visible>(G:Var2) 0 == (A:Avionics master switch, bool) &&</Visible>

won't work in the .txt version, I get a compile error when I open the .xml file.

Tom, compare your post version with your original text version, I think you'll see what I mean. I can't figure out how to show the && correction on the board.

[Tom Goodrick]

My text version has the line:

"<Visible>(G:Var2) 0 == (A:Avionics master switch, bool) &&</Visible>"

This is the exact same code that has been used in all my xml gauges. It seems the culprit is this FORUM's input processing. It has removed the "amp" from each of the "&" signs.

[Tom Goodrick]

The doggone Forum did it again!!!!! And again!!!!!

Each of those "&" characters should actually read, in your version of the text before changing the extension to .xml as follows: "&"

(It removes the amp.) The word amp should follow each & sign in the text version. When compiled, you won't see the amp.

[Allen Peterson]

OK, one more time time to try to fool the Forum.
For anyone having trouble following the above, replace the && in the Precip.txt file with & a m p ; & a m p ; and then remove the spaces.

[Tom Goodrick]

Right.

[ben]

I think I remember :-) a tragedy in Australian Alps ( approx 8000ft ) in a C172 (?) where four adults were killed.
I suppose if you get caught in icing conditions with close to max weight, it wouldn't take too long for icing weight to be a problem (compared to 1 or 2 passengers)

ben