trisager Posted February 23, 2009 Report Posted February 23, 2009 I'm using ASA with FSUIPC wind smoothing enabled in FSX with very good results. I have the smoothing set to 4 seconds per knot change which completely eliminates wind shift problems. However, I've noticed that if open an addon (e.g. open FSUIPC or edit PMDG failure options) while in flight, when I return to FSX I always get a big wind shift, resulting in an altitude and/or speed excursion. This is not much of a problem, but I'm wondering why it occurs. Is FSUIPC somehow buffering up changes that get released instantly when coming back from the addon? Tom
Pete Dowson Posted February 23, 2009 Report Posted February 23, 2009 I'm using ASA with FSUIPC wind smoothing enabled in FSX with very good results. ASA's wind smoothing, or FSUIPC's? I'm not sure you should use both. I have the smoothing set to 4 seconds per knot change which completely eliminates wind shift problems. However, I've noticed that if open an addon (e.g. open FSUIPC or edit PMDG failure options) while in flight, when I return to FSX I always get a big wind shift, resulting in an altitude and/or speed excursion. This is not much of a problem, but I'm wondering why it occurs. Is FSUIPC somehow buffering up changes that get released instantly when coming back from the addon? Not FSUIPC no, it isn't really running then, but it doesn't smooth the wind setting it sees set when you exit menus, because you might have been in the menus to download new weather or manually change the weather. There's no way FSUIPC knows what you are doing in the menus. So, if you are using FSUIPC's wind smoothing, it doesn't smooth across Menu excursions. Effectively when you come back to flying, you get the weather you should have -- i.e. the target weather, not the weather en-route to the target at n knots per second. I expect that is what you are seeing? Regards Pete
trisager Posted February 23, 2009 Author Report Posted February 23, 2009 I'm using FSUIPC wind smoothing only, not ASA's. Here is a typical scenario (just tried it again): I'm in cruise with winds apparently stable (i.e. wind vectors on display not changing). I open FSUIPC and immediately exit to FSX. Slam, a huge wind shift causing a 50kt drop in indicated airspeed. I was thinking that somehow, maybe because of the slow rate of change that I allow, wind changes got buffered somewhere by FSUIPC and then instantly released to the sim when I exited the menu. Anyway, as I said, not much of a problem, I was just wondering. Tom
Pete Dowson Posted February 23, 2009 Report Posted February 23, 2009 Here is a typical scenario (just tried it again): I'm in cruise with winds apparently stable (i.e. wind vectors on display not changing). I open FSUIPC and immediately exit to FSX. Slam, a huge wind shift causing a 50kt drop in indicated airspeed. FSUIPC cannot differentiate the reason for going into the menu system, it is just told by SimConnect that simulation stopped. If FSUIPC insisted on smoothing continuation after returns from the menu it would make it very difficult for folks wanting to set weather explicitly or even loading a new flight with its own weather already saved. FSUIPC's wind smoothing (and all the other smoothing actions it did in FSUIPC3 too) have always done this, for 7 or 8 years now. It isn't different or new. Maybe the use of ASA now has somehow induced more distant target for FSUIPC to reach? Certainly if you have smoothing set to 4 seconds for every knot or degree of change i should think it will hardly ever actually catch up with any wind changes. Perhaps if you tried a more reasonable 1 or 2 knots/degrees per second you might not get such violent effects? After all the idea of smoothing is to smooth changes, but still allow them to happen reasonably timely. Regards Pete
Andydigital Posted March 8, 2009 Report Posted March 8, 2009 Hi Pete, further to what you have said above, would it at all be possible to add an option to have FSUIPC not release control of the weather smoothing features during menu transitions? It would actually help greatly with the aircraft that so much rely on your excellent weather taming facilities.
Pete Dowson Posted March 8, 2009 Report Posted March 8, 2009 Hi Pete, further to what you have said above, would it at all be possible to add an option to have FSUIPC not release control of the weather smoothing features during menu transitions? It would actually help greatly with the aircraft that so much rely on your excellent weather taming facilities. I'm not really sure why you would interrupt an otherwise immersive flight by going into the menu system in the first place? Can you explain why this has suddenly become an issue? The way FSUIPC works has been okay for everyone, at least judging by the lack of such a request, ever since the wind smoothing was introduced back in FS2002 times I believe. Maybe before. I'm not sure I really want to mess with it at present. It does get quite complicated because it is all based on averaging over elapsed time. Some justification for change is needed at this stage, I believe. I don't understand it. Regards Pete
Andydigital Posted March 9, 2009 Report Posted March 9, 2009 It does get quite complicated because it is all based on averaging over elapsed time. Can you not stop your time keeping when you see the user enter a menu. I sorry if this is a really stupid comment/question but I'm no programmer in any sense of the word, but I imagine from your response that it is a major thing I am asking for, and the last thing I want to do is create any more headaches for you Pete :lol: The reason its become an issue is because the new autopilot systems in the likes of the PMDG MD-11 and Level-D 767 get very messed up when the winds or wind speeds shift suddenly which most of the time result in a stall or overspeed condition. As you probably know far better than me winds can very often be as much as 180 degrees out to what they were 1 second ago in FSX and within a few seconds it can be back to where it was again, by which time the complex custom autopilot has tried to correct for the change only for it to be completely wrong again. The systems you have in place now work extremely well as long as you don't need to go into a menu and change any settings, which for some people is unavoidable I'm sure. For instance I tend to have to go and reduce my scenery density or autogen sliders at times when getting to a heavily built up area because my frame rates drop too much. Maybe I need to go into FSUIPC itself to tweak something or add a new button assignment or some such option. Needless to say there are many reasons to have to go into a menu unfortunately. I've also noticed that it can happen if you have to pause the flight for more than a few seconds, its a shame real life get in the way but us sim pilots are lucky we have the pause option when we are on short final and the door bell rings :roll: As I've mentioned previously if this is a major undertaking for you Pete than please don't think any more about it, you have done more than enough to warrant the asking price of FSUIPC many times over.
trisager Posted March 9, 2009 Author Report Posted March 9, 2009 I think this issue may be related specifically to FSX and PMDG aircraft. Maybe that is why this is coming up now - more and more users flying airline sims in FSX, particularly the MD-11? Re the rate of smoothing that I'm using: With the Level-D 767 and other compex addon aircraft smooting at the rate of a few knots per second works fine, but with the 747x and the MD-11 it has to be fractional knots per second to avoid altitude and speed excursions. The general advice in the PMDG forum seems to be to smooth at 8-10 seconds per knot change. (Currently I'm using 3 seconds per knot, but it doesn't make much difference from before when I was using a slower rate - the AP still goes bonkers when coming back from the menus). A typical reason for me to go into the menus in-flight would be to reset a failure, or to allow another failure to happen after the first one has triggered. I try to avoid it, obviously - partly for immersion, but also because it mostly means spending time recovering from an overspeed or low-speed condition. And like Andy I sometimes change display settings en route, e.g. to add some AI if flying online in an empty region. Cheers, Tom
Pete Dowson Posted March 9, 2009 Report Posted March 9, 2009 It does get quite complicated because it is all based on averaging over elapsed time. Can you not stop your time keeping when you see the user enter a menu. I sorry if this is a really stupid comment/question but I'm no programmer in any sense of the word, but I imagine from your response that it is a major thing I am asking for, and the last thing I want to do is create any more headaches for you Pete It may or may not be major, I've not looked at it yet. I am simply seriiously concerned about messing with it for little good reason. The earlier posting on this was actually mentioning a stupidly slow smoothing rate of some 1 degree /knot every 4 seconds. At such a rate the wind will almost never be what it should be. It's a wonder the AI traffic at airports aren't conflicting all the time with your own flights. With more sensible smoothing rates like 1 for 1, the Wind should never get that far behind what it should be that it is going to matter so much. The reason its become an issue is because the new autopilot systems in the likes of the PMDG MD-11 and Level-D 767 get very messed up when the winds or wind speeds shift suddenly which most of the time result in a stall or overspeed condition. But this was always a factor with the PMDG 747X as well, and it seemed to be all sorted, a year or two ago. Has all this new fuss come with the use of ASA in place of ASX? If so, isn't it in that direction you should be looking? As you probably know far better than me winds can very often be as much as 180 degrees out to what they were 1 second ago in FSX and within a few seconds it can be back to where it was again That is actually pretty rare except when using a weather control program, and it is due to exactly the same bug in FSX as was in FS9. This has been going on for years and years. The smoothing is not the only need, the weather setting programs like ASX, and I presume ASA, were also changed to try to deal with it. The systems you have in place now work extremely well as long as you don't need to go into a menu and change any settings, which for some people is unavoidable I'm sure. For instance I tend to have to go and reduce my scenery density or autogen sliders at times when getting to a heavily built up area because my frame rates drop too much. Maybe I need to go into FSUIPC itself to tweak something or add a new button assignment or some such option. Needless to say there are many reasons to have to go into a menu unfortunately. During a flight? I never do. It definitely destroys the whole point of the realism I've strivern for. I've also noticed that it can happen if you have to pause the flight for more than a few seconds Because, presumably, the smoothing is continuing based on real time. I'll have a look when i get a moment. Maybe it can be changed to work on FS time, though that could be messy as all timing in FSUIPC, throughout, is based on the Windows "tick" time. Changing a whole section of code to work with different timers could be very error-prone. It is really this risk of introducing bugs in complex code which worries me more than the work involved. Regards Pete
Pete Dowson Posted March 9, 2009 Report Posted March 9, 2009 I think this issue may be related specifically to FSX and PMDG aircraft. Maybe that is why this is coming up now - more and more users flying airline sims in FSX, particularly the MD-11? Pretty much all the testing on the smoothing in FSUIPC4 was done with the PMDG 747X, and it has been okay now for over a year. Suddenly it is a problem? Are you sure this isn'yt something in ASA which is different from ASX? That seems to be more significant in this change of needs all of a sudden, doesn't it? The general advice in the PMDG forum seems to be to smooth at 8-10 seconds per knot change. (Currently I'm using 3 seconds per knot, but it doesn't make much difference from before when I was using a slower rate - the AP still goes bonkers when coming back from the menus). Blimey! Really the facility wasn't designed to work that slowly, though of course the arithmetic will still work. At those sorts of rates you are never likely to get the predicted or stated winds, anywhere. Doesn't that cause problems with runway assignments and so on? I've always use a 1 for 1 rate with no significant problems. Pete
Andydigital Posted March 9, 2009 Report Posted March 9, 2009 Sorry Pete I failed to mention that I'm on the Beta test team for HIFI and so should have said that I use ASA as well. Believe me when I say and I'm sure you are probably aware that Damian and the guys are working really hard on ASA with PMDG assistance too to try and get the problems sorted out regarding wind shifts etc. Recently their understanding of what is causing the hunting/swaying problems in PMDG aircraft has shifted somewhat (excuse the pun). It seems to be caused by the way FSX depicts turbulence more than anything else. While HIFI may be able to tame the effects to a greater degree with ASA, there are other people who don't have ASA or ASX and they are still having problems with the default weather engine which requires your system to be used. I found that 2 seconds per degree or knot with turbulence suppressed/ticked in FSUIPC works pretty much perfectly for me in the MD-11 and LD767. I certainly don't need anywhere near 8 seconds Tom, which as far a I understand it was recommended by someone on PMDG beta team certainly not officially by PMDG themselves. I do really wish I didn't have to go into the menu's or have to pause the flight, but needs must at times like I mentioned already. Thanks for give this some consideration at some point in the future Peter it is very much appreciated by all I'm sure.
Pete Dowson Posted March 9, 2009 Report Posted March 9, 2009 I've had a look at my code. It turned out that it didn't look too hard to change all the smoothing to operate based on the "ELAPSED TIME" values which i nowadays read as a Gauge variable -- the one provided at offset 04A8. I don't think this is as accurate as the Windows timer I normally use for everything. It is probably only at most as accurate as the 55mSec (1/18th second) ticker in FS, or possibly the frame rate. I don't think it is updated more often than the frame rate in any case. This may make it a little more uneven in how it smooths the wind. The only advantage this timer does have is that it stops whilst FS in in menus, or paused, and, Ithink, runs faster or slower according to the FS Simulation Rate. So, maybe it will do the job you want of it. Try it and let me know. I'm afraid I won't have the opportunity to do any proper testing of this till next week. It is provided as a session mode, selected via an INI file [General] parameter: SmoothBySimTime=Yes (defaults to No at present -- I may change the default if the results look really good). The wind smoothing is still reset when you load a new flight, or move the aircraft location via menu, or change the weather mode (theme <-> user <-> real, etc). download http://fsuipc.simflight.com/beta/FSUIPC4504.zip now (please make sure you've installed 4.50 first, though). Let me know, please. Regards Pete
Andydigital Posted March 9, 2009 Report Posted March 9, 2009 OK Pete thanks a lot for this I will certainly test it as much as I can over the next week, today is going to probably be out of the question though as I have a few beta's for various different FS developers to work through that were released in the last 24 hours. Thanks again.
Andydigital Posted March 9, 2009 Report Posted March 9, 2009 Just a quick note to say that initial testing of a 90 minute flight in the Md-11 was very encouraging, but I wont be foolish enough to say you have sussed it out just yet. I've learnt my lesson from past experiences of saying something works perfectly only for it to turn around and slap me in the face 5 minutes after I've made a post on a forum. It does look very promising so far though, I even paused for 7 minutes while taking a phone call and all was well on my return to flight.
Andydigital Posted March 15, 2009 Report Posted March 15, 2009 Morning Pete, I've been testing this during the last week (not as much as I'd like) and I can report that there are no adverse effects as far as I can see. I wouldn't be so bold as to say you should make this the default mode though, having it as a option would be more prudent I think especially until a such time as its been tested by a wider audience. The only thing I found a little odd was that variance seems to be different now, particularly wind heading variance. I only really had one flight where it seemed odd, but the variance seemed slow and too regular, although the METAR did specify variable winds so it was correct but just a little too predictable. Other than that I have nothing else to report, other than it does what I need well. Would I be right in thinking that this facility is still in the latest interim i.e. 4.505, I've not downloaded it yet because I was unsure if it was still in there. Thanks again for the changes.
Pete Dowson Posted March 15, 2009 Report Posted March 15, 2009 The only thing I found a little odd was that variance seems to be different now, particularly wind heading variance. I only really had one flight where it seemed odd, but the variance seemed slow and too regular, although the METAR did specify variable winds so it was correct but just a little too predictable. Hmm. I'll have a look -- but i don't think the changes i made can have influenced that. The randomisation of variance and turbulence would still be, er, randomised. Maybe it's related to the accuracy (or rather inaccuracy) of the FS timer value I'm using now (probably only updated once per Frame), compared to the Windows tick time. I doubt it but I'll look anyway. Would I be right in thinking that this facility is still in the latest interim i.e. 4.505, I've not downloaded it yet because I was unsure if it was still in there. Yes. I don't take facilities out. Well, not often! . I just haven't documented it yet because I was waiting for your feedback. I'll add the documentation for it to the next update. Regards Pete
Recommended Posts
Create an account or sign in to comment
You need to be a member in order to leave a comment
Create an account
Sign up for a new account in our community. It's easy!
Register a new accountSign in
Already have an account? Sign in here.
Sign In Now