FSUIPC Weather Interface

[Bernd Podhradsky]

Hi Pete!

Got a question about the weather station system FSX uses. I'm currently sending Extended METAR Strings (Custom Mode = Station Mode) to FSX via SimConnect (or FSUIPC which - as far as I know - uses SimConnect itself) for all weather stations 200 nautical miles around the aircraft position. That works fine, the weather's there!

However, even though the Dynamic Change Rate is set to 0, after a few seconds (!) stations lose their cloud layers. I thought that it was some internal FSX interpolation thing going on, so I extended my radius to 300 nautical miles. In theory, this should prevent nearby stations to change their weather. But that wasn't the case, there are still stations losing cloud layers that are 60 nautical miles away.

Have you heard of this issue before? Any ideas what it could be?

Thanks,

Bernd

[Pete Dowson]

However, even though the Dynamic Change Rate is set to 0, after a few seconds (!) stations lose their cloud layers.

The weather is dynamic in any case, irrespective of the change rate setting. After all, if there is any wind at all, things have to change, to move. That's what wind is -- an equalisation of pressures. And clouds move with the wind.

However, disappearing cloud layers is something I've seen reported but not really noticed myself. I can understand them moving away in favour of an area with less layers from the other direction, and maybe they can "dissolve" if the temperature/dew point setting for the layer they are in dictates a dry air situation. But simply getting lost for no good reason doesn't make sense to me, and isn't something I've seen.

If this is in an area with quite a few weather stations close to each other, like around Chicago and some parts of California, I would tend to think it due to conflicts btween those nearby stations, being resolved.

I thought that it was some internal FSX interpolation thing going on, so I extended my radius to 300 nautical miles. In theory, this should prevent nearby stations to change their weather. But that wasn't the case, there are still stations losing cloud layers that are 60 nautical miles away.

I doubt that the range is relevant to this issue.

There are a lot of bugs in the FSX weather engine, as there was in FS9's for that matter. This is why the more successful weather programs now choose to use Global mode -- making the weather the same everywhere -- and adjusting it dynamically by program. This solves a lot of these sorts of problems and certainly eliminates interpolation errors. Of course for ATIS reports on destination airports it means things are wrng -- the weatrher reports really then need to either be obtained from the same source as the weather setting program, or from that program itself. This is what I do with Active Sky -- you can either read its weather file or use the direct interface it has supported since ASE.

Regards

Pete

[Bernd Podhradsky]

Hi Pete!

Thanks for your answer.

You're absolutely right, the symptoms I have described are the same no matter what range I use to set the stations. The strange thing is, that weather stations only lose cloud layers, they're not getting additional ones.

Global Mode has is advantages, but it also has a lot of disadvantages: it doesn't work with AI Traffic, ATC and you don't have smooth weather changes. That's where station mode is much better.

Another issue that might be related to my problem, maybe you've heard about this: as I said, I set the weather for all stations within 200 nautical miles around the aircraft position. If I place the aircraft @ Runway 9L in EGLL and start setting the weather, EGRB (a weather station located in London Centre) gets is weather set and it displays correctly. However, as my setting process progresses, as soon as another station (about 30 miles away vom EGRB) is set, EGRB loses all of its cloud layers and suddenly has some very, very strange wind layers in FSXs weather menu. I send 12 wind layers (5000 feet, 10000 feet etc) but now EGRB has wind layers in 4990 feet, 5000 feet, 5010 feet etc. That's why I thought it's some kind of interpolation that's going on. Have you ever heard of this issue?

Thanks in advance,

Bernd

[Pete Dowson]

Another issue that might be related to my problem, maybe you've heard about this: as I said, I set the weather for all stations within 200 nautical miles around the aircraft position. If I place the aircraft @ Runway 9L in EGLL and start setting the weather, EGRB (a weather station located in London Centre) gets is weather set and it displays correctly. However, as my setting process progresses, as soon as another station (about 30 miles away vom EGRB) is set, EGRB loses all of its cloud layers and suddenly has some very, very strange wind layers in FSXs weather menu. I send 12 wind layers (5000 feet, 10000 feet etc) but now EGRB has wind layers in 4990 feet, 5000 feet, 5010 feet etc. That's why I thought it's some kind of interpolation that's going on. Have you ever heard of this issue?

Ah, the multiplication of wind layers at silly small height differences is the problem I found when FSUIPC was set to update weather via its filters. This is why that is defaulted off and warnings are provided about it. The filtering was intended to operate by reading the weather, making changes, and writing it back. Alas, no way this works well with this problem occurring.

I do have algorithms in place which tries to combat this by deleting the spurious layers. One of the problems is that the total size of the METAR string eventually goes beyond the SimConnect limit of 2048 bytes, after which crashes occur in the SimConnect interface. FSUIPC records this size problem.

Have you tried setting weather the other way around, distant stations first, local ones last?

Pete

[Bernd Podhradsky]

Hi Pete!

Yes, I've tried both ways. When starting off with nearest stations first, then EGRB is the station that gets these strange wind layers and loses all clouds. If doing it the other way around, EGRB is set correctly with another station (don't know the identifier) getting these wrong values. However, setting all stations twice eliminates the problem with the faulty wind layers (but doesn't help with the problem of weather stations losing cloud layers).

I'll check, if any of my extended METAR strings is more then 2048 bytes. Am I correct to assume that we're talking about 1-byte-per-character encodings (ie 1 character of the METAR string = 1 byte)?

Thanks a lot for your help!

Bernd

[Pete Dowson]

Am I correct to assume that we're talking about 1-byte-per-character encodings (ie 1 character of the METAR string = 1 byte)?

Yes. FS uses normal ASCII encoding with is an 8 bit = 1 byte code per character.

Pete