[jhb] Re: Pirep Grrrr!
- From: "Fossil" <fossil@xxxxxxx>
- To: <jhb@xxxxxxxxxxxxx>
- Date: Tue, 27 Apr 2010 16:13:16 +0100
The difficulty for aircraft designers is that they need skills in too many
areas these days. Most start out just wanting to design a nice aircraft so
they get to grips with GMax or FSDS and eventually (after much torture) end
up with a very nice visual model. That alone can take months - it did for my
Tiger - but then FS created Virtual Cockpits so the poor designer now has to
create the instrument panel and gauges - and partly build the cockpit
internal detail too.
Next step is to find out about texturing (another learning curve) as it can
be tricky wrapping textures to part of the aircraft that are compound
curves. Texture wrapping nearly had me throw the whole project away.
Next comes sounds but these are complex to record and to apply in FS so most
farm out this work to sounds guys like Mike Hambly. Even after he sent me
the Tiger sound set and I looked at the CFG file values I still couldn't
make sense of what he'd written.
Now you finally have a nice looking (we hope) visual model with working
instruments and decent sounds - but the hardest part hasn't yet begun.
Turning the grounded aircraft into something that remotely flies like a real
aircraft requires two disciplines. First is that, in the designing of the
aircraft, you set the empty Centre of Gravity in the right place. Oddly
enough this is embedded in the GMax file and if you've got it wrong when
designing the aircraft it is a right pig to change. In fact you can't move
it as it is the reference point for the visual model and the only solution
is to move the whole aircraft model around it.
The .air file and the aircraft.cfg file then need to be built to reflect the
handling and performance of the aircraft. Given that the .air file is a
shadow of the earlier FS98 data set and is now quite crude the first shock
for a new designer is that even if they have all the necessary book figures
to type into the aircraft.cfg file the sad fact is that the aircraft will
not fly anything like the real aircraft. In truth you start off with the
book figures and see how the FS aircraft flies and then slowly start
adjusting all the parameters to get the aircraft as close to the real thing
as you can get.
The trouble here is that you really need to know something about the real
aircraft and the way it flies otherwise your changes are guesswork. With the
Tiger the first changes were in basic stability - get the aircraft in level
flight and then see what happens with a brief control displacement. This is
crucial because if you don't have correct response in yaw, pitch or roll
displacement the aircraft will never handle correctly. This test has to be
run through the full speed range of the aircraft too so it can take some
time.
Once this is determined the next step is to tune the aileron, elevator and
rudder responses. Again this must be done across the whole speed range - but
slow speed testing has to be treated with caution as FS is awful in this
area. Sadly FS isn't good at high AoA lift/drag calculations because it
models both wings as a single surface and the result is an aircraft with
very docile handling compared to the real thing. Only Real Air have pushed
this boundary back but only by careful manipulation of the .air and config
values.
Once basic handling is tuned the designer can move on to gear and flap
operation, tuning these for correct drag and pitch effects. Next is engine
power changes on pitch and yaw but before doing this the correct power
output has to be checked in case the FS model is over or underpowered.
Whilst on the engine the fuel burn at various altitudes must also be checked
as this will have an impact not only on burn rate but also on range and rate
of climb/descent.
And so it goes on, checking every config value and testing again and again
because some changes will impact on other values already tested. It's hard
work and you sometimes find the changes result in unacceptable behaviour and
you have to start all over again.
I know a lot of designers are quite good at the external model and panel
design but spare a thought for performance. Some of the config settings are
serious aerodynamic theory and a lot of people will quickly get well out of
their depth. A decent aerodynamicist might make a good stab at a design but
even here it needs some actual knowledge of the real aircraft's handling to
get it anywhere near being accurate.
bones
bones@xxxxxxx
-----Original Message-----
From: jhb-bounce@xxxxxxxxxxxxx [mailto:jhb-bounce@xxxxxxxxxxxxx] On Behalf
Of Gerry Winskill
Sent: 27 April 2010 11:50
To: JHB Restricted
Subject: [jhb] Pirep Grrrr!
Just come back into the room, after hanging a couple of pictures for
SWMBO. I'd landed at FACT at1037 our time. Back in to receive a fizzer
saying I'd been idle too long, so days times cancelled.
Anyway, the flight was to check the changes I'd made to the Gulfstream
550. It now flies as it should and covers the claimed 8000 nm, instead
of its out of the box 6000 nm.
It's difficult to justify criticising Freeware designers but I sometimes
feel they are happy to produce a nice looking aircraft, with the correct
numbers entered into its Aircraft.cfg. Flying the things, better still
getting someone else to test it, seems to be overlooked. Pity because
the Gulfstream 550 looks more like a commercial airliner than most
bizjets, whilst having great range and performance.
Whilst up in the loft, looking for frames, I came across a mini TV, that
we used to use on the canal boat. Since it's analogue, it's merely a
candidate for the bin. Pity but that's progress foryou.
Gerry Winskill
Other related posts:
