Board index FlightGear Development Aircraft Flight dynamics model

Vertical speed  Topic is solved

Good sims require good FDMs (the "thing" that makes an aircraft behave like an aircraft).

Vertical speed

Postby V12 » Sun Jan 13, 2019 3:07 pm

I found strange VSI property behaviour :
/autopilot/internal/vert-speed-fpm shows -110 fpm
/velocities/vertical_speed-fps shows -1.83 fps
/instrumentation/adc/output/vertical-speed-fps shows 0.3 fps
/instrumentation/vertical-speed-indicator/indicated-speed-fps shows 0.3 fps

/position/altitude-ft rising with rate displayed in /instrumentation/adc/output/vertical-speed-fps

Can someone explain me this behaviour ? I do not understand it :(
I want simulate Cruise-Climb autopilot mode on Concorde, but don't know what source for VSI I should use.

Thank You.
Fly high, fly fast - fly Concorde !
User avatar
V12
 
Posts: 731
Joined: Thu Jan 12, 2017 4:27 pm
Location: LZIB
Callsign: BAWV12

Re: Vertical speed

Postby Thorsten » Sun Jan 13, 2019 3:12 pm

/fdm/jsbsim/velocities/v-down-fps is simulation truth.
Thorsten
 
Posts: 10079
Joined: Mon Nov 02, 2009 8:33 am

Re: Vertical speed

Postby legoboyvdlp » Sun Jan 13, 2019 3:49 pm

I think instrumentation/ADC/ is probably the Concorde's own systems - probably will be a jsbsim system file for that. The vertical-speed-indicator is the best imo but is vulnerable to changes in pressure, if your autopilot uses then you might get bad behaviour. I think the one thorsten said is the best.
User avatar
legoboyvdlp
 
Posts: 5999
Joined: Sat Jul 26, 2014 1:28 am
Callsign: YV-LEGO
Version: 2018.3.1
OS: Windows 10 HP

Re: Vertical speed

Postby it0uchpods » Sun Jan 13, 2019 5:22 pm

I use the GPS one with a lag filter and it acts the closest to a ADIRS sources vertical speed indicator. I've noticed, that some VS sources are "true" I guess not altitude relative, which can cause some discrepancies.

Kind Regards,
Josh
it0uchpods/WTF411

FGFS Aircraft Developer
Lead Programmer at it0uchpods Design Group
Download High Quality Aircraft from it0uchpods Hangar
User avatar
it0uchpods
 
Posts: 3622
Joined: Tue Oct 06, 2015 12:51 pm
Version: 2018.3.1
OS: Windows 10 x64

Re: Vertical speed

Postby Thorsten » Sun Jan 13, 2019 5:55 pm

I've noticed, that some VS sources are "true" I guess not altitude relative, which can cause some discrepancies.


I'm... a bit confused what you mean. What is an altitude-relative vertical speed? Speed is a derivative, it doesn't care what the altitude is. :?:

Generally for a JSBSim craft /fdm/jsbsim/position/ contains the simulation truth for position and /fdm/jsbsim/velocities for the various speed measures.

I believe FG copies them partially for both JSBSim and YaSim under /position/ and /velocities/ .

Now philosophies differ. You can use an FG-side instrument simulation (aptly under /instruments/ ) and if you configure it properly and understand what it does, ideally it gives you something like a real instrument trying to measure truth.

Or you can define your own system to 'blur' simulated truth (which can be arbitrarily complex, the Shuttle navigation state uses a probabilistic estimate of truth given the available sensor readings, their estimated error ranges and the redundancy management procedures to de-select probably failed sensors.

Generally - do not use a value you do not understand.
Thorsten
 
Posts: 10079
Joined: Mon Nov 02, 2009 8:33 am

Re: Vertical speed

Postby it0uchpods » Sun Jan 13, 2019 9:16 pm

Sorry -- did not well talk

I was meaning that I believe that some ones have the actual vertical speed of the aircraft, while the instrumentation/gps one will match more to what the aircraft is doing. By altitude relative, I meant, when the VS reads 0, the altitude remains about the same. I noticed at some times (probably due to metar/weather), the "true" V/S from velotices/ sometimes shows something like -200fpm, while the altitude is staying the same.

Kind Regards,
Josh
Last edited by it0uchpods on Sun Jan 13, 2019 9:39 pm, edited 1 time in total.
it0uchpods/WTF411

FGFS Aircraft Developer
Lead Programmer at it0uchpods Design Group
Download High Quality Aircraft from it0uchpods Hangar
User avatar
it0uchpods
 
Posts: 3622
Joined: Tue Oct 06, 2015 12:51 pm
Version: 2018.3.1
OS: Windows 10 x64

Re: Vertical speed

Postby V12 » Sun Jan 13, 2019 9:30 pm

Thorsten wrote in Sun Jan 13, 2019 3:12 pm:/fdm/jsbsim/velocities/v-down-fps is simulation truth.

Are You sure ? When I fly at FL550 M2 with ALT HOLD, /fdm/jsbsim/position/h-sl-ft sitting on 57000 feets, but /fdm/jsbsim/velocities/v-down-fps shows value something around 3.6. Why ? Earth curvature ? That 3.6 fps is almost it0uchpods's 200 fpm. It is too late for math now...

EDIT :
I can't sleep, it is not Earth curvature, at FL570 and M2 (GS 1100 kts ) is relative descent on the arc trajectory with radius 6378 km + 57000 feets only around 0.08 feets in 1 second...
Fly high, fly fast - fly Concorde !
User avatar
V12
 
Posts: 731
Joined: Thu Jan 12, 2017 4:27 pm
Location: LZIB
Callsign: BAWV12

Re: Vertical speed

Postby Alant » Sun Jan 13, 2019 9:45 pm

The basic VSI instrument has a very large lag. This is due to the internal design of the instrument. Flightgear models this effect.

Have you tried setting a steady climb ( by flying at a constant airspeed and pitch angle) and checking that all vertical speed properties and the VSI all show the same after a time?

The VSI is there to check the rate of climb when in steady trimmed flight. If you want to set a particular climb or descend rate you should adjust the throttle (more to climb, less to descend), and then control pitch attitude to maintain the desired speed. Once you have trimmed throttle and pitch angle then, and only then can you look at the VSI to see the resulting climb/descent rate.

The altimeter does not have this lag, so can be used to fly at zero reta of climb/descent.

Alan
Alant
 
Posts: 761
Joined: Wed Jun 23, 2010 5:58 am
Location: Portugal
Callsign: Tarnish99
Version: from Git
OS: Windows 10

Re: Vertical speed

Postby it0uchpods » Sun Jan 13, 2019 10:16 pm

You misunderstooding.
That VSI actually shows correct at cruise, (even it I sit at cruise for hours)
The so called "truth/true" VSIs are the ones that show wrong.

Kind Regards,
Josh
it0uchpods/WTF411

FGFS Aircraft Developer
Lead Programmer at it0uchpods Design Group
Download High Quality Aircraft from it0uchpods Hangar
User avatar
it0uchpods
 
Posts: 3622
Joined: Tue Oct 06, 2015 12:51 pm
Version: 2018.3.1
OS: Windows 10 x64

Re: Vertical speed

Postby Alant » Sun Jan 13, 2019 11:17 pm

I have just done a quick test on my TSR2 at M1.6, 30000ft with the autopilot in height and Mach lock, and confirm /fdm/jsbsim/velocities/v-down-fps behaves as reported.

At sea level all is as expected.

I checked and the error is not due to aircraft axes against earth axes.

/fdm/jsbsim/velocities /h-dot-fps is the same as /fdm/jsbsim/velocities/v-down-fps.

Perhaps Bertrand on the JSBSim list will know the answer. Do you want me to ask him?

Alan
Alant
 
Posts: 761
Joined: Wed Jun 23, 2010 5:58 am
Location: Portugal
Callsign: Tarnish99
Version: from Git
OS: Windows 10

Re: Vertical speed

Postby it0uchpods » Sun Jan 13, 2019 11:57 pm

Alan,
Are you using the GENERIC autopilot and GENERIC Altitude? If so, you won't see this issue. I am talking about using the realistic properties, which don't match the "GENERIC" ones.

Kind Regards,
Josh
it0uchpods/WTF411

FGFS Aircraft Developer
Lead Programmer at it0uchpods Design Group
Download High Quality Aircraft from it0uchpods Hangar
User avatar
it0uchpods
 
Posts: 3622
Joined: Tue Oct 06, 2015 12:51 pm
Version: 2018.3.1
OS: Windows 10 x64

Re: Vertical speed

Postby Alant » Mon Jan 14, 2019 1:01 am

No, I am using my own autopilot , which is based upon TSR2's 1960´s analogue system.

I am seeing the same non-zero JSBSim height rates in level flight that are being reported here. This is with a rock steady altimeter, Mach meter and ASI. In other words the autopilot is not a factor.

I have had a quick look at the JSBSim code.
So far have I found that /fdm/jsbsim/velocities /h-dot-fps is defined as the "current altitude rate" and /fdm/jsbsim/velocities/v-down-fps as "down vel in NED frame, fps". These are both equal and significantly non zero in high altitude, high Mach number trimmed level flight. A hig Mach No, sea level they are near zero.

In light of the different definitions it is strange that both are showing what appears to be the same error.

I have not yet gone into how JSBSim calculates these properties.

Within the Flightgear code I see that the VSI instrumentation has separate code for a basic aneroid capsule VSI and a corrected one as in many more modern air data systems.

I am sure that, when asked, Bertrand will be able to provide an answer to this question.

Alan
Alant
 
Posts: 761
Joined: Wed Jun 23, 2010 5:58 am
Location: Portugal
Callsign: Tarnish99
Version: from Git
OS: Windows 10

Re: Vertical speed

Postby it0uchpods » Mon Jan 14, 2019 1:26 am

Hi Alan,
I do not mean the VSI Instrument, those display correct. I am talking about things like, /velocities/vertical-speed-fps (NOTE: FPS, not FPM!)

Can you say me which EXACT properties which you are talking about for ALTITUDE, and which V/S you are referencing to look at this?

Kind Regards,
Josh
it0uchpods/WTF411

FGFS Aircraft Developer
Lead Programmer at it0uchpods Design Group
Download High Quality Aircraft from it0uchpods Hangar
User avatar
it0uchpods
 
Posts: 3622
Joined: Tue Oct 06, 2015 12:51 pm
Version: 2018.3.1
OS: Windows 10 x64

Re: Vertical speed

Postby V12 » Mon Jan 14, 2019 6:18 am

Another observation - I was on the gate, zero speed, engaged parking brakes, fine weather with zero wind and /fdm/jsbsim/velocities/v-down-fps oscilated from -5 to +4 fps. Where is truth ???
Fly high, fly fast - fly Concorde !
User avatar
V12
 
Posts: 731
Joined: Thu Jan 12, 2017 4:27 pm
Location: LZIB
Callsign: BAWV12

Re: Vertical speed

Postby Thorsten » Mon Jan 14, 2019 6:39 am

/fdm/jsbsim/velocities/v-down-fps

and

/fdm/jsbsim/position/h-agl-ft

seem to be related in that the first is consistent with being the derivative of the second - in orbital flight, I can very easily use the THC to control vertical speed and this relation works out.

That'd mean that v-down-fps (like v-east-fps and v-north-fps) is the speed relative to the local ground point on the reference ellipsoid (i.e. this is not an inertial or earth-co-rotating coordinate 'down' when the coordinate origin is Earth's center).
Thorsten
 
Posts: 10079
Joined: Mon Nov 02, 2009 8:33 am

Next

Return to Flight dynamics model

Who is online

Users browsing this forum: No registered users and 1 guest