Possible wind smoothing issue

[trisager]

Once in a while I get a problem with the FSX winds, usually on final approach to a landing. What will happen is that the wind will change direction over an arc of about 120-180 degrees within a few seconds. The wind direction will then jump back to the starting position and go through the arc again, etc. This goes on for a while while I climb and go around, until it stops as suddenly as it began.

It happened again today, this time while I was still parked at the terminal (I had wind smoothing at the surface level enabled in FSUIPC). This time I tried disabling FSUIPC surface level wind smoothing, and the problem was gone. Of cause this is only a single data point, but in case this problem does have something to do with FSUIPC wind smoothing, what kind of data can I collect the next time that can help find the cause of this?

I have all sorts of stuff installed with FSX, but things I think are relevant are: FSX SP2, FSUIPC 4.26 (and 4.261 since yesterday, problem seen with both) configured to smooth wind and pressure, and Active Sky X (wind smoothing in ASX disabled) on a different computer. OS is Vista32. I've seen the problem both with and without surface wind smoothing enabled, and it has happened with two different airplanes from different vendors.

Tom

[Pete Dowson]

Once in a while I get a problem with the FSX winds, usually on final approach to a landing. What will happen is that the wind will change direction over an arc of about 120-180 degrees within a few seconds. The wind direction will then jump back to the starting position and go through the arc again, etc. This goes on for a while while I climb and go around, until it stops as suddenly as it began.

Sounds like wind variance. Check the METAR report -- variable direction and the range is indicated in there. Or use WeatherSet and see the Variance column.

Suppress it if you don't like it (in FSUIPC's Winds tab). I've not had much feedback on whether it is realistic or not, because it appears to be pretty infrequent except when the winds are very light (1-10 knots). What were yours?

It happened again today, this time while I was still parked at the terminal (I had wind smoothing at the surface level enabled in FSUIPC). This time I tried disabling FSUIPC surface level wind smoothing, and the problem was gone.

With the wind smoothing disabled, FSUIPC cannot simulate the three different wind effects, so you are left with FS's feeble efforts.

Regards

Pete

[trisager]

Hi Pete,

The METARs at the time did not indicate any wind variance. In the first case the wind speeds were at 20-25 kts, the one time it happened on the ground had winds around 5 kts.

I don't know how you simulate wind variance, but I would expect an element of randomness. In these cases there was no random shifts in direction or strength - the wind direction moved through an arc at the rate allowed by wind smoothing, then instantaneously jumped back to the starting direction and the pattern repeated, exactly the same on each iteration. No changes in the wind velocity.

As previously mentioned this is an infrequent occurence, and this far it has only happened during approach or on the ground.

Tom

[Andydigital]

I've also come across this problem on 2 or may be 3 occasions over the last month or so. I thought it was a problem local to me, hence the reason i hadn't mentioned it before. I did find that ticking the variance box stopped it within 3 or 4 seconds of returning to the flight.

[Pete Dowson]

The METARs at the time did not indicate any wind variance. In the first case the wind speeds were at 20-25 kts, the one time it happened on the ground had winds around 5 kts.

Which METARs did you check? You would need to know which three weather stations contribute to the interpolated weather at the aircraft, and in general that is not easy to determine. You would need to run something like WeatherSet to see the current interpolated weather being supplied from SimConnect, or engage FSUIPC's weather logging and check the reports in the Log.

I don't know how you simulate wind variance, but I would expect an element of randomness.

Yes. Target directions within the range specified are chosen at random, and then these are approached at an interval controlled by parameters until a new random target is computed at a random time later. Under test conditions it looks quite good.

In these cases there was no random shifts in direction or strength - the wind direction moved through an arc at the rate allowed by wind smoothing, then instantaneously jumped back to the starting direction and the pattern repeated, exactly the same on each iteration.

Ah, that does sound wrong. FSUIPC most certainly does go to enormous lengths to try to prevent any instantaneous jumps. It sounds like all my efforts have been in vain after all. :-( :cry: :cry: :cry:

I've really no more tricks up my sleeve, no other avenues I can even begin to investigate. Odd that every report until yours, over many weeks, said how well the smoothing worked, so this comes as a really bitter blow.

Do you think you can reproduce it to order? If so, could you supply a saved FLT+WX file set so I can see it here?

... this far it has only happened during approach or on the ground.

Well, approach is one of the worst times for it to happen, so that's a poor excuse. As for on the ground, it is generally best to have wind smoothing disabled on the ground so that new weather conditions can be instigated without enforced delays.

I'd love to be able to fix it, but I would first need to reproduce it. There must be some way FSX is bypassing the avenues for wind setting which I am already trapping. As I implied, I really don't have much hope of going further than I have, so it may be simply an exercise in verifying that it is happening then documenting it as a warning. The wind smoothing has been received so well by so many that I don't feel I can withdraw it now. :-(

Pete

[trisager]

Hi Pete,

Since this is a relatively rare occurence I for one will definitely not stop using the wind smoothing features in FSUIPC - I much prefer the overall smoothness that you have achieved, with the occasional issue on approach a drawback but not a showstopper. Keeps me on my toes, after all, so I don't forget the possibility of a missed approach :) Much worse before, with aircraft overspeeding and crashing due to overstress.

I don't think I can reproduce it to order, but the next time it occurs I'll grab the files that you requested.

The METAR data that I was looking at was what ASX showed me for location I was at, not any other stations. I'll take a look at ways of obtaining that information.

Tom

[GHD]

Odd that every report until yours, over many weeks, said how well the smoothing worked, so this comes as a really bitter blow.

I have had the same experience when following AI, there were instantaneous jumps to 102, 86, 2, 220, 94, 2 with it returning at one second intervals back to 294, the setting in World/Weather.

George

[Pete Dowson]

I have had the same experience when following AI, there were instantaneous jumps to 102, 86, 2, 220, 94, 2 with it returning at one second intervals back to 294, the setting in World/Weather.

Suddenly, after weeks of everyone saying it all works well, reports are streaming out of the woodwork about how bad it is after all? :cry: :cry: :cry:

Your results seem to be the exact opposite of Toms. So the whole implementation seems doomed.

Some days it is better to stay in bed. :-(

Please, give me files I can reproduce this mess with!

Pete

[Pete Dowson]

I have had the same experience when following AI, there were instantaneous jumps to 102, 86, 2, 220, 94, 2 with it returning at one second intervals back to 294, the setting in World/Weather.

Suddenly, after weeks of everyone saying it all works well, reports are streaming out of the woodwork about how bad it is after all? :cry: :cry: :cry:

And didn't suppressing the variance fix it for you as it did for Andy?

Also, your results seem to be the exact opposite of Tom's. Please, give me files I can reproduce this mess with!

Some days it is better to stay in bed. :-(

Pete