PSXAloft II : new upper wind and turbulence simulation for PSX

[Hardy Heinlin]

Carl, I suspect your VisualPSX is injecting slightly inconsistent elevation data from your external scenery. According to your description, the injection time intervals are approx. 950 ms, 50 ms, 950 ms, 50 ms, i.e. there seem to be two 1-second cycles, whereas one cycle is injecting data X (correct), and the other is injecting data Y (incorrect). The TDZ display picks up these changes instantly, but the aircraft has a great inertia and therefore doesn't change its altitude so quickly, hence the barometric altimeter doesn't change quickly either. If this theory is right, maybe VisualPSX sits in a P3D scenery between two different runway elevations, causing a decision problem.


|-|

[Britjet]

Thanks Carl and JP..

Peter

[cavaricooper]

Hardy-

I suspect you are correct (as usual)- this has not resurfaced elsewhere.  Sheesh- what are the odds of me picking that field for the initial PSXAloft II flight!  Perhaps I had better go purchase a Lotto ticket (with it currently above $1.3 billion it does seem like a remarkably good idea).

Ta- C

PS- about every third post from you has me desperately wishing I were smarter... I still stand in awe!

Carl, I suspect your VisualPSX is injecting slightly inconsistent elevation data from your external scenery. According to your description, the injection time intervals are approx. 950 ms, 50 ms, 950 ms, 50 ms, i.e. there seem to be two 1-second cycles, whereas one cycle is injecting data X (correct), and the other is injecting data Y (incorrect). The TDZ display picks up these changes instantly, but the aircraft has a great inertia and therefore doesn't change its altitude so quickly, hence the barometric altimeter doesn't change quickly either. If this theory is right, maybe VisualPSX sits in a P3D scenery between two different runway elevations, causing a decision problem.


|-|

[Hessel Oosten]

Jean-Philippe,

Completely nauseous here....

You did IT !
It works without any problem here -and-fluent (even the turbulence ...

Thanks a lot !

Hessel

[JP59]

Hello,

Many thanks Hessel for your kind words. I'm happy to see it works fine everywhere (fingers crossed) 

I wonder if some real 747 pilots use PSXAloft, may you feedback your feeling about turbulence generation for Light,
Moderate and Severe modes ? What about turbulence frequency, length, intensity, sound, sound duration, etc ? If you have any suggestion to improve this feature do not hesitate to share your experience.

Best regards,

[Britjet]

I hope to try it today...
Peter

[Hessel Oosten]

Well, it seems (IS !) better to await Peter's/Hardy's comments, I presume (I AM SURE)... .

But my feeling was, that yawing was less in relation to pitch movements

Hessel

[Britjet]

Very impressed, JP - although I have yet to work out how to export flight plans from PFPX etc ..

If I understand correctly..

Without using any of the "flight plan" options - PSXAloft will interrogate Activesky and produce winds etc for any flight, anywhere? Great.
I was able to do a descent page download without an ASky flight plan (didn't try en-route winds) and they loaded - is this the ActiveSky winds? They seemed to agree very well with my PSXAloft upper winds, but I'm not sure..

- Turbulence! He He! I nearly had a heart attack when the "severe turbulence" sounds started! Did you by any chance record them from an old Citroen CV with a broken exhaust?! Very clever, anyway! I think that the sound may be a bit overdone though - a little bit of seat rocking, the occasional seat belt tab clinking (always annoying!) and perhaps a bit of glassware "tinkle" would be my suggestion..

I tried all three levels of turbulence. I think they are very good. You certainly wouldn't want any more pitch movement - particularly in severe (I only tried severe in descent so far) - it would be quite sick-making with an immersive wrap screen such as I have..

I don't think there needs to be any increase in yaw movement. Although the rw back end swings about a bit in turbulence it isn't something that you feel or see at the front.

Question - Can you change the turbulence points again once in flight? eg after passing point AAA on route AAA BBB CCC with turbulence at point BBB - can you select AAA and start the turbulence?

I'm enjoying the realism - nice one!

Peter

[JP59]

Hello Peter,

Without using any of the "flight plan" options - PSXAloft will interrogate Activesky and produce winds etc for any flight, anywhere? Great.

Yes. In dynamic mode, PSXAloft retrieve wind and OAT data directly from your scenery generator through SimConnect. Data can come from any weather add-on (ActiveSky, REX, Opus,...) and is available anywhere in the world. Usefull in case of diversion 

I was able to do a descent page download without an ASky flight plan (didn't try en-route winds) and they loaded - is this the ActiveSky winds? They seemed to agree very well with my PSXAloft upper winds, but I'm not sure..

If you request datalink without an ASN flight plan file loaded within PSXAloft user interface, FMC will be fed with "default" PSX weather data. It also works like this if you load a flight plan different than your PSX FMC route (say, you load an ASN LFPG-RJAA flight plan, and your FMC is flying a KJFK-EGLL route). If you are lucky data will match, especially for short legs, like descent, but I recommand to always load the ASN flight plan within PSXAloft interface before to request datalink.

- Load you flight plan within ASN interface (exported from PFPX in FSX format)
- An "activeflightplanwx.txt" file will be generated by ASN
- Load this file within PSXAloft via user interface (button "BROWSE"). The file is generally located at C:\Users\UserName\AppData\Roaming\ASNP3D\Weather
- You can go with datalink

- Turbulence! He He! I nearly had a heart attack when the "severe turbulence" sounds started! Did you by any chance record them from an old Citroen CV with a broken exhaust?! Very clever, anyway! I think that the sound may be a bit overdone though - a little bit of seat rocking, the occasional seat belt tab clinking (always annoying!) and perhaps a bit of glassware "tinkle" would be my suggestion..

Sorry if I frightened you. I'll add a caution message on the turbulence button  . The severe sound is different from the light and moderate ones. I'll work on it.

Question - Can you change the turbulence points again once in flight? eg after passing point AAA on route AAA BBB CCC with turbulence at point BBB - can you select AAA and start the turbulence?

Yes. Just press the turbulence configuration button, reset turbulence zones up to you, and press OK. The turbulence generation algorithm will calculate your new turbulence zone.

I'm enjoying the realism - nice one!

Thank you very much Peter for your feedback.

Kind regards,

[Britjet]

Thanks "JP"!

Peter

[Britjet]

Hi Jean-Philippe,

I am "bumping around" with your severe turbulence in the cruise!
It looks to me as though what is happening is that you are injecting a pitch change into PSX which causes a height fluctuation. To me that is the wrong way round, but I realise that this may be very difficult to simulate.

What happens, as I see it, for the height change, (I think!) is that you get a sudden vertical movement of air changing the angle of attack, and the aircraft starts to change height. It then pitches on autopilot to compensate - so you would get:-
1) Quite rapid (but short time) height change tendency
2) Pitch change to compensate

What I see (?) happening in PSXAloft (and please correct me if I am wrong) is:

1) Pitch change (injected?)
2) Resultant height change
3) Autopilot then compensates back to target

I suppose that the desired effect might best be achieved by injecting a sudden loss or gain of main wing lift - but of course I don't know if you can do this..

As for speed - I think if anything the instantaneous speed "excursions" (ie speed trend arrow) could be maybe 25% bigger. The trend arrow can be quite alarming!

Pitch change - too large at the moment. Quite vomitable actually! I suppose this is a function of the way you may be causing the height change? I would certainly never expect to see it get up to 6 degs, which is what is happening at the moment (in lee waves this would certainly be possible, but not otherwise).

I hope this helps!

Peter

[Hardy Heinlin]

Hi,

it is possible to inject certain turbulence impulses, but there are no pitch impulses. Injectable impulses are for vertical speed, bank, yaw, and headwind/tailwind. Pitch reactions are a consequence of injected vertical speed impulses or headwind/tailwind impulses.


Cheers,

|-|ardy

[JP59]

Hi Peter,

I am "bumping around" with your severe turbulence in the cruise!
It looks to me as though what is happening is that you are injecting a pitch change into PSX which causes a height fluctuation. To me that is the wrong way round, but I realise that this may be very difficult to simulate.

As Hardy said, I can just inject some altitude, bank, yaw and wind variations. The pitch changes you describe are the reaction of the autopilot to altitude changes I inject. I cannot inject pitch variations directly.

For your understanding of the background, PSXAloft turbulence generation algorithm is a big randomizer thread. It generates :

- Random altitude, bank, yaw and wind variations. Sometimes you can get all of these 4, sometimes 3, 2, or 1
- Random intensity of each of these variations
- Random number of data injections in PSX during the same loop (makes the effet stronger or lighter)
- Random time between injections
- Random time between each loop
- Random sound duration

For each level of turbulence (Light, Moderate or Severe) randomized lower and upper limits are different. The only thing I can do is to decrease altitude changes in severe turbulence. Thus it will indirectly limit the pitch variations the autopilot commands.

Please try PSXAloft v.2.0.1 (link and change log on the first post of this thread) and tell me what's your feeling. I also increased airspeed variations a little bit.

Thanks again for your contribution
Cheers,

[Britjet]

Hi Jean-philippe,

Thanks (and Hardy) for the explanation. I have just flown 2.1
Just talking about severe..

1) I like the speed trend vector now :-)
2) Small changes in bank look good :-)
3) The reduced pitch effect is better :-)
4) I am not sure how you can program this, but the up and down pitching is too frequent - it feels like a small boat in a storm. Up, down, up, down (etc).
What would be better, I think, is to have the injection less frequent but longer duration
eg...
1) vertical draft - slowly increases over (say) 5 seconds to 700fpm - Autopilot compensates
2) draft decreases at similar rate - autopilot compensates.
3) A couple of seconds with no draft to allow pitch to stabilise
4) The same or reverse effect.

I was wondering if the (sometimes) large pitch changes are due to the autopilot making a correction and then being "caught out" by a vertical draft in the opposite direction?
I think the main thing here (for me ) is that the vertical changes seem too frequent and too short, and reverse too quickly?

I hope that helps..

Peter

[JP59]

Hi Peter,

4) I am not sure how you can program this, but the up and down pitching is too frequent - it feels like a small boat in a storm. Up, down, up, down (etc).

This is due to AFDS which over-compensate after a vertical speed injection. The only thing I can do to reduce this effect is to reduce intensity of V/S injections. This is what I did.

What would be better, I think, is to have the injection less frequent but longer duration
eg...
1) vertical draft - slowly increases over (say) 5 seconds to 700fpm - Autopilot compensates
2) draft decreases at similar rate - autopilot compensates.
3) A couple of seconds with no draft to allow pitch to stabilise
4) The same or reverse effect.

I worked on my code. Please download again PSXAloft v.2.0.1 and try it. I only changed severe turbulence for now. Waiting for your feedback before to work on other modes. I also worked on synchronization between sound and "aerodynamic" turbulence effect. The sound sucks (I'm searching for a new one on the internet) but tell me if you prefer this new synchro (or not).

Regards,

Edit : A new version is online since 17:05 UTC. Please try this one.

[Britjet]

Hi Jean-Philippe,

Thanks for the quick update! - I tried it (showing still as 2.01 but I can see it is different).

I think that is better. If anything I think you could increase the effect very very slightly on the draft - maybe stronger for longer ?

One for Hardy - I think the PSX autothrottle is too responsive here to the speed changes? - it immediately makes a large thrust setting change at the first sign of a speed change if the trend vector is large.

Now I'm not sure (as usual!) what the autothrottle uses but I don't think it looks at the speed trend vector - just the speed, so a change of speed of (say) 5-10 kts would not cause a huge thrust change even if the trend vector was very large. Eventually it would "wake up" and put the required thrust on if the speed continued to change, but not at first..I would be interested to know what parameters PSX uses..
In turbulence we would often have to "help" the speed control by overriding the thrust levers to maintain a sensible speed. You would have the autothrottle motors fighting you quite often as it hadn't yet changed the thrust.

Any other 747 guys who might like to chip in on this, please?

I think it is getting there, Jean Philippe, but of course that is only my opinion and I would welcome other 747 input.
I well remember losing a 90kt headwind over the Alps in 10 seconds one evening, it was rough as hell with speed getting to the "red"  from .82 Mach (even with thrust levers rapidly closed manually) but the altitude and pitch hardly did a thing. Maybe we need a "buttkicker"? :-)

Peter

[JP59]

Thank you so much Peter for your kind help. I "compiled" all your suggestions :

- 3 turbulence modes reworked
- Better sound synchronization for 3 modes
- Severe turbulence sound less loud

For you and all PSXAloft users, version 2.0.2 is up. You can download the new package with the link provided on the first post of this thread.

Best regards,

[Hardy Heinlin]

Hi Peter,

the A/T in PSX and in the real aircraft does not only look at the speed; the result would be as terrible as in FS5.

Like in all control loops, there are also trend and acceleration parameters involved to avoid overreactions on one hand, and to prevent too slow a response on the other hand. When you have 50 knots to go, you don't want to wait 10 minutes; and when you're approaching the speed bug faster than planned, you want to counteract this new trend earlier than planned. Otherwise the control will oscillate. Knowing the trend is essential to dampen and  stabilize the control loop.

These trend parameters need to be fine-tuned very carefully. It's not a question whether they are used, but how much influence they have.

This influence also depends on the trend direction: Like on the real aircraft, PSX's A/T reacts faster upon a tailwind increase, and slower on a headwind increase. When increases of tailwind and headwind alternate nearly equally in time and intensity, the A/T will therefore keep the average speed some knots above the speed bug (a desired side effect) because the thrust reduction after a headwind event always takes longer than the thrust increase after a tailwind event.

In PSX 10.0.6, the A/T's thrust decrease command is slower than the spool-down capability of the engines, while the thrust increase command is slighly faster than the spool-up capability. I.e. if I would make the command faster, the engines wouldn't spool up faster anyway. But you asked for a slower thrust increase command. I'll check if I can fine-tune this again in the next update. A hair too slow and you will unnecessarily lose some knots in the next tailwind shear -- that's what I want to avoid ...


Cheers,

|-|ardy

[Britjet]

Hi Hardy

That's an interesting explanation. Thank you! I certainly didn't know that the trend arrow was a factor.
Common 'cultural' knowledge is that the autothrottle is relatively unresponsive in cruise as a fuel-saving measure, but if course it is a difficult thing to check on the aircraft. Certainly the sims don't model cruise turbulence at all well, so I can't experiment.

Of course another factor is that if the thrust changes in a few seconds from nearly max to nearly idle, this is a huge pitch-couple change, and also an aircraft at max thrust that suddenly encounters an increasing speed trend would have to throttle back quickly. For this reason I know that one school of thought ( on all jets, really), is to disengage the auto throttle, and make small manual adjustments about a datum.

I have noticed on the sim that if you allow the speed to get low on the approach (say, 10 kts below Ref) without the autothrottle engaged, and then engage it, the autothrottle immediately certainly puts on a LOAD of thrust. Cruise might be different, however. I have no idea if there are cruise and approach modes. I wouldn't have thought so..

I appreciate your thoughts about changing it slightly but I would beware of changing anything just based on my 'feeling'. I could be wrong...(as usual). It's just that in PSXAloft the auto throttle seems more active than I remember.

Cheers!

Peter

[cavaricooper]

FWIW, I remember reading that approach mode triggers a difference response curve in AT logic (from TOGA, Cruise, Climb etc), with particular attention to low speed attenuation....however, John and our other gurus can no doubt shed more illuminating light...

@Jean-phillipe-

I just flew a SHORT route segment of the last flight with MODERATE turbulence (PSXAloft version 2.0.2)... I usually smiled through my outside loops in the S2C, but this bit had me hanging on for dear life.  I was particularly glad the trolly-dolly hadn't brought my cuppa as my attention was fully on the PFD and 'tis a shame to waste a good Darjeeling

Thanks for working on this immensely important part of long haul simulation!

Best- C