All,There has been a lot of discussion regarding Takeoff Torque settings since the release of FS2Crew. A few folks have taken the initiative, and are actively looking for ways to approximate the performance data. There are some good ideas floating around, and I just wanted to add my take on the issue.I first attempted to adapt the real data for use in the simulator, but unfortunately, the amount of data that I have is too small of a sample to be able to extrapolate with any kind of accuracy. Second, the real data didn't really match the simulator in most cases, which I chalk up to the fact that the FSX turboprop model is totally boned.So instead, I spent the evening flight testing the J-41 in FSX and recording the simulator data. I tabulated that data, and the result is a simple spreadsheet that covers the entire operational temperature range, and from Sea Level up to 6000ft (Pressure Altitude). I'd be willing to do the work to get the data valid out to 8000ft if there is demand for it, but we'll see how that goes.I tested the data at about half a dozen airports using current weather and had very good results. Please test it out and verify that the data works throughout the range of conditions, and provide any feedback so I can tweak as necessary...Note: I found that the most accurate method is to use the closest altitude and temperature from the chart. For example, if the field elevation is 5150, and the temp is 11C, it's best to use to use the value that corresponds to 5000' and 10C. In the real world, you would typically go to the next highest value in order to find the most conservative limit, e.g. 5500' and 13C. That method is not preferable using this data.Let me know if this is helpful. I'll repost this to the FS2Crew Forum as well.

Yes, the table is useful. It is more realistic for use in flying then my attempts at using a formula based on density altitude. I think the effect is about the same (seeing as temp factors very largely in to the density altitude equation) idea, but having a table to look up values is much more practical than calculating DA and then putting that in to a formula.

Thanks Eric. Unfortunately, I'm not a math whiz so I usually go the "brute force" route. In this case, I manually generated 217 data points (31 discrete temperatures at each of 7 altitudes (Sea Level to 6000ft)) and interpolated to generate the data in between. I found that after accessing the FSX Weather dialogue about 35 times (about once per test run), FSX would crash. Then I'd have to reload FSX and get the airplane all setup from a cold and dark state and pick up where I left off. Add to that the amount of time it takes to set power without invoking the TTL, and it ended up taking a lot more time than I had anticipated.I wish I were smart enough to figure out a formula!

Excel solver has a quadraic regression tool for fitting a second order polynomial to a data set, so you get some volunteers to 'beta test' for you and put their data in a spreadsheet and press a button. The is the PHD (push here dummy) method. I used it a couple years ago trying to fit fuel consumption data for the 747-400 but I wasn't clever enough to delegate the data collection. Gosh that took a lot of time.

quadratic poly who? :(

HS level Trigonometry, basically. I would have never though of using regression equations for torque data. I may just have to try this.

Excel solver has a quadraic regression tool for fitting a second order polynomial to a data set, so you get some volunteers to 'beta test' for you and put their data in a spreadsheet and press a button. The is the PHD (push here dummy) method. I used it a couple years ago trying to fit fuel consumption data for the 747-400 but I wasn't clever enough to delegate the data collection. Gosh that took a lot of time.

Well, quadratics may not be necessary. I already began working with data myself and using a regression model to model this. I elaborated on what I was doing more in this thread in the FS2Crew forums: http://forums1.avsim.net/index.php?showtopic=286655Quadratic equations still only analyze data with respect to one variable. If the data is linearly distrubted, there is no need or advantage to quadratic regression (Excel has several regressions built in including linear, quadratic and polynomial). I discussed some basic theories of heat transfer and support for a linear distribution in this thread here: http://forums1.avsim.net/index.php?showtopic=286732Along with the data I have from the first referenced post, this appears to be an accurate modeling for the static thermal transfer properties involved.