PSXWeather v1.6 Released

[Hardy Heinlin]

Hardy-

What is the altitude that PSX switches from actual METAR Wx to computed assumptions based on internally modeled data?

There is no switch. It's an interpolation between ground (METAR data) and tropopause (PSX data).


|-|

[cavaricooper]

Hardy-

Perhaps the better question would be- at what point does the PURE METAR data start getting interpolated with tropopause calculations....?

Ta!

C

[Hardy Heinlin]

At one millimeter above the ground.

[Tigh]

I think it was Gary O. who asked for a method to override FMC RTE DATA uplinks. That was some months ago and I think it's implemented now in the BACARS add-on package?

Not yet, they're working on a package to integrate the various apps used in Gary's sim and the FMC RTE DATA information should be part of that.

[cavaricooper]

Craig & Hardy-

With v1.6.1 have you noticed that the injection each 5 seconds is different enough to be noticed by the AFDS system in PSX? 

There is a distinct twitch of the pitch FD bar.... the Wx is a slowly changing medium and would not change noticeably in the 5 second window.  Is the FD behavior simply the reaction of the AFDS getting "new" data and "recomputing" each 5 seconds?

Craig- is there a way to just use ASN data (for AFDS/FMC data and ND depiction) for the FULL flight regime- basically "turning off" the PSX Wx model?

Thank you ALL for your efforts- PSX with VisualPSX, TrafficPSX, BAACARS, PSXonMCP2 and PSXWx have transformed my sim experience totally- and I am much obliged!

Best- C

[Hardy Heinlin]

The FD directs to a target attitude, depending on the engaged FD mode.

[cavaricooper]

Yes sir... the question is more about the "twitching" that is not present without PSXWx, but IS present with it.  I was wondering what the injection of repeated 5 second winds was doing to AFDS to cause this.... a constantly recalculated vertical path every 5 seconds....?  Obviously there would be lateral deviations as well, but these are not enough to cause FD twitching... so perhaps there is a technique that could be transferred from lateral to vertical (in PSXWx)?

Best- C

PS- Craig, hope you are reading all this and that the logs I sent were of use....

[Hardy Heinlin]

The PSX AFDS just does what the real AFDS does: It monitors data from the ADC and IRS.

ADC and IRS sense wind, weather, motion ...



You need to be more specific, Carl, when you describe something. I guess you're in an airspeed-controlled-by-elevator AFDS mode? Airspeed is a function of temperature, among other things. Perhaps there are some microscopic, abrupt OAT changes in your injections, causing a microscopic IAS trend, which is to be compensated by a microscopic elevator input.

[cavaricooper]

Hardy-

Affirmative- FLCH. Perhaps the "smoothing" isn't smooth (enough)? Just trying to provide observations to help speed us along the road to nirvana .

Hopefully this is all helpful to Craig et al....

Best- C

[G-CIVA]

The FD directs to a target attitude, depending on the engaged FD mode.

Hardy,

Its actually the Horizontal Stabilizer that moves rapidly or 'pulses' every 5 seconds or so - in sych with the WX update from PSXWx ... observed from watching the control surface indicator at the top left of my main screen.

The F/D movement is secondary.

Happens in both CLB & DES with PSXWx 'engaged' to PSX.

[craig1231]

Hey guys... Im still here! Sorry its been a while... i've been concentrating on Uni work for a while. I've had a look at those logs you kindly posted, Carl. Some interesting things are happening... Its trying to inject negative values still, which I thought was fixed. I shall attempt to fix when I can, next week.  Are you able to demo the twitching in a video? With PSXWx 1.6.1 I would recommend setting the base altitude to at least 10000ft would fix some issues. Leave the max altitude at -1, for no max altitude.

Thanks for your support

Craig

[Hardy Heinlin]

The FD directs to a target attitude, depending on the engaged FD mode.

Hardy,

Its actually the Horizontal Stabilizer that moves rapidly or 'pulses' every 5 seconds or so - in sych with the WX update from PSXWx ... observed from watching the control surface indicator at the top left of my main screen.

The F/D movement is secondary.

Happens in both CLB & DES with PSXWx 'engaged' to PSX.

On that indicator, top left, that's the elevator.

The causal chain goes like this:

1. There are air mass changes & aircraft motion

2. Sensed by ADC & IRS

3. Data is sent to AFDS (outer loop)

4. AFDS computes target attitude to perform selected guidance mode (inner loop)

5. AFDS displays target attitude by FD bars

6. AFDS computes target flight control positions to achieve target attitude

7. AFDS computes required hydraulic commands to achieve target flight control positions (servo loop)

8. Hydraulic system moves autopilot flight control actuators

9. Go to 1.


This process is a matter of milliseconds and is executed ca. 50 times per second. To your eye, everything seems to occur at the same time.


|-|ardy




When a short elevator command occurs, does it occur when the tropopause OAT has changed (by an injection) on Instructor > Situation > Weather > Planet?

[cavaricooper]

Craig-

No worries mate- Uni absolutely comes first!  I will try and get a video for you if someone else doesn't first .

Are you referring to some .ini file entires re the base and max alt?  I will look next time I'm in my sim

Appreciate your efforts!

Best- C

Hey guys... Im still here! Sorry its been a while... i've been concentrating on Uni work for a while. I've had a look at those logs you kindly posted, Carl. Some interesting things are happening... Its trying to inject negative values still, which I thought was fixed. I shall attempt to fix when I can, next week.  Are you able to demo the twitching in a video? With PSXWx 1.6.1 I would recommend setting the base altitude to at least 10000ft would fix some issues. Leave the max altitude at -1, for no max altitude.

Thanks for your support

Craig

[G-CIVA]

On that indicator, top left, that's the elevator.

My bad .... that's exactly what I tried to say but did not.

[cavaricooper]

Craig-

I figured out where to make the change (.xml).  WHAT A DIFFERENCE the FL100 made in my flight experience.  That was what I was after with my repeated clumsy queries to Hardy.....

Log here- https://app.box.com/s/6h5g9sl5aj46gz2iiohdif48faaexnyr

As you can see from the videos, the PSXWx injections send data to AFDS and then AFDS is sending corrective commands to the elevator.  What is interesting is that the current "smoothing" results in one UP and one DOWN elevator need that AFDS commands alternatingly....

Here are the videos you asked for... sorry for the rubbish quality, I just made a quick and dirty capture with my iPhone.....

CLIMB video-
https://www.youtube.com/watch?v=6-qMIgiKS8U

VNAV descent video-
https://www.youtube.com/watch?v=pIixR8tGE48

FLCH descent video-
https://www.youtube.com/watch?v=6n7Eg9wMc20
(I allowed the 15t overspeed to regain the profile)

Your PSXWx will allow for RW & PFPX FP use, and for that I am very grateful.... please let me know if there is anything else you need.

HTH

Best- C


PS- Hardy... re. "When a short elevator command occurs, does it occur when the tropopause OAT has changed (by an injection) on Instructor > Situation > Weather > Planet?" I forgot to check... will do so next flight.  Sorry!

[Hardy Heinlin]

Looking at these videos, these "twitches" are extreeeeemely small. I expected something bigger :-)

Look at the IAS. The IAS change is the cause. It's a very short and small change in the IAS trend.

Consider this a normal supermini turbulence in the wilderness of our atmosphere :-)

Sometimes this may also occur without injecting add-ons.

Maybe I can add some more FD damping for these special cases in the next update.


|-|

the PSXWx injections send data to AFDS

Not directly.

PSXWx injections send data into PSX's atmosphere.

The AFDS doesn't read weather reports, it reads data from ADCs and the IRS.

The ADCs have sensors which "read" data from PSX's atmosphere at the current aircraft 3D location.

It's really important to understand this causal chain. There are mutliple sources of ADC and IRS data, and their data differ slightly or may even provide totally wrong data due to probe icing or other malfunctions.

[cavaricooper]

Hardy-

That would be great! 

With PSX getting external injections, perhaps you could look at the overall VNAV and FLCH AFDS calculations and enable some sort of dampening there as well.... With the ASN injected data vs. PSX computed weather there seem to be some tendencies to over correct and then under correct before finally getting the VS correct... resulting in mild oscillations.  This occurring both in VNAV and FLCH descents.

I now realize what  it took for Boeing to arrive at their programming, and why they are so very reluctant to make ANY changes to code without years of testing.  Again...... a multi-billion Dollar industry giant on one hand, and a guy who was never told that he couldn't do it, on the other. I remain amazed .

Still, probably best to concentrate on one variable at a time, so once PSXWx is injecting a steady stream of smooth ASN data......

Best- C

[Hardy Heinlin]

you could look at the overall VNAV and FLCH AFDS calculations and enable some sort of dampening there as well.

Not there "as well". Just there.

It's just about airspeed-by-elevator-control.

No worries, I had a plan already :-)

mild oscillations

See, Carl, a vague description like "mild" says nothing to me. You really need to be more specific. Otherwise all brainwork and forum reading is just a waste of time. I need numbers. Seconds, milliseconds, feet, degrees ...

[cavaricooper]

Hardy-

Once PSXWx is optimized with smooth injection, there will be a baseline for AFDS optimization.  Nothing to look at for the moment. Glad to hear you had noticed the airspeed on elevator issue already.

Best- C

PS- re "a vague description like "mild" says nothing to me. You really need to be more specific. Otherwise all brainwork and forum reading is just a waste of time. I need numbers. Seconds, milliseconds, feet, degrees ..." I will try to capture video AFTER PSXWx is sorted.  My grey matter is merely mortal (read VERY average).... not anywhere HH quality .

[Hardy Heinlin]

Hi Craig,

your add-on injects every 5 seconds. Do you think it's possible to add some random variations to these 5 seconds, to get a random sum of, say, 3 to 8 seconds?

E.g.: Time = 3 + random(5)

I just implemented a function in my AFDS that ignores IAS trend peaks for a fraction of a second when the IAS has been stable over a longer period of time. That is, weather injection related IAS changes no longer cause AFDS counter actions on the FD and elevator. However, such injections may still cause slight movements on the PFD's IAS tape (less than 0.3 knots). These slight movements are actually quite natural, so, per se, they are OK, but only if they occur at random. That's why I suggest to vary the intervals a little bit at random. What do you think?


Cheers,

|-|ardy