GA Flight Simulator - University Project Question

[bwfly88]

Hey, I'm a student at University and my project is to improve the flight controls on their flight sim. It's simulating A Socata TB 9 (light 4 seater aircraft similar to a PA28). I know with large commercial airline sims they are simulating hydraulic flight controls so don't necessarily employ a feedback system. My question is with a GA aircraft the flight controls are obviously mechanically linked so have more of a feed back than an airliner. I haven't had a chance to visit any over GA sims yet but do they have a system that employs force feedback for the controls? i.e. sloppy controls at slower speeds and stronger controls at high speed? I appreciate GA sims would more likely be procedural trainers and not a lot of emphasis on the feel of the aircraft would be applied but I am interested to know if there are any other GA sims out there that employ this system?

Thanks!

[Mad (Flt) Scientist]

Given you are located in the UK, you may wish to investigate the BAe 146 / AVRO RJ, which is an airliner and should have some full Level D sims around, but has manual aileron and elevator controls, so would need to address the force feedback issue as you described.

It's by no means the only aircraft out there with manual controls on at least one axis that should have a modern sim around ...

[ReverseFlight]

Most if not all GA sims do not have control loading primarily due to cost issues. This generally applies to 3-axis sims (as opposed to 6-axis sims). Even jet-type fixed based sims normally do not have control loading as these are essentially procedural trainers. Only Level D sims will have the capability and necessity to replicate full control loading and aerodynamic characteristics.

[Genghis the Engineer]

What model simulator are you looking at?

G

[Pilot DAR]

Though the control forces do increase with speed, they can also increase with G, and vary near the stall (in the angle of attack - not speed sense). There are also variations due to trim and flaps use, and landing gear position. I would think that this would be a very complex combination of conditions to duplicate in a sim. (But I'm no expect on sims)

If you're taking on such a task, you owe it to yourself to do some flying, and begin to appreciate these forces first hand (pun intended).

[bwfly88]

Thanks for the replies!

It's a Socata TB 9 simulator. I have a PPL so happy with knowing what forces are involved. I just wanted to know if it was a viable decision to include a force feedback system on a GA sim and if there were any others in the Country.
Nothing wrong with me researching this for my project and gathering all the relevant info and specs only at the end to say it is not viable and I do not recommend the addition of such a system.

Regards,

[kenparry]

The only GA simulators of which I have any experience is some Frasca sims that Oxford ATS (as it was then) used for IR lead-in training in the late 70s. My dim recollection is they had some feel system, perhaps only a spring feel of fixed characteristics. I think any sim, whether fixed-base or with a motion system, would be extremely unpleasant and very difficult, as well as unrealistic, if it had no form of force feedback, at least in pitch.

Don't think that airliners with powered controls have poor force feedback; it's not so. For decades they have had q-feel to give realistic control forces.

Pilot DAR makes good points about how real manual control forces vary, and you would be looking at a complex system, as well as a comprehensive aerodynamics data package, to generate fully realistic feedback. However, a simple spring feel system would be simple to fit - and there were some aircraft with powered controls that had no more than that. For example, the Hunter had simple spring feel in pitch (with the addition of a bobweight on the Mk 9 to give some g-related feel). In roll, I can't remember, but there may have been none. (Edit: now I remember, spring feel on the ailerons as well)

[Genghis the Engineer]

I doubt very much it is a Socata simulator - to the best of my knowledge that company has never manufactured simulators. It will almost certainly be a simulator by somebody like Qsim, Merlin, Fasca... which is trying to simulate a Socata TB9.

The reason for my question is that different companies have tried to use different technologies to simulate control forces. Mechanical springs, pneumatic springs, electrical tensioners, hydraulic dampers.... Knowing what's there, and possibly knowing what's been used on similar simulators before, may make it possible to offer some useful advice.

In general terms however, giving approximately representative control forces in a GA sim is useful and achievable. The JAR-STD standards which, if you've not already looked at them, do encourage this, and rightly so - pilots use control force feedback a lot to give them direct feedback of aeroplane behaviour.

What is much harder is a truly accurate representation of control forces - for a start without some good FT data you don't actually know what those forces are anyhow (and whilst a PPL is a good start, you should be aware that light aeroplane control forces and control force harmonisation can vary a good deal between types - as well as with power, speed and configuration within a type. It's an interesting discussion point how accurate forces really need to be however - a literature review may with any luck give you some answers on that.

G

[SimWes]

Only Level D sims will have the capability and necessity to replicate full control loading and aerodynamic characteristics.

It’s not only level D machines that have full Control Loading and the aerodynamic characteristics. Level C as well as Level B use them as well (the ones I’ve worked on anyway)

All modern Simulators (and old) have to have the characteristics of the Controls checked as well how they handle

Control Loading feedback nowadays can be either hydraulically or electrically driven. Older machines (really old ones) had springs or rubber bungies fitted!

The only real main difference between level D & C Simulators are that you are required to run more tests, but these are Motion and Sound related, not the Controls

Then of course there are differences between level C & B, which I’m not going to go into

Just to complicate matters, Control Loading systems can be classed either as Primary Control Loading systems (yoke, wheel, rudder etc) as well as Secondary ones (toe brakes, auto throttle etc).

Hope that’s of some help

[what next]

Hello!

If you are interested in the JARs that define the criteria for flight training devices, google for "JAR FSTD-A". This document specifies (among many other things) the accuracy with which control forces need to be modelled for certain types of training devices. (Only relevant if these are used for JAR-FCL licenses of course).

The FNPT II device that our FTO uses for instrument and multi-engine training uses a pneumatic system to generate control forces. This is quite complex and uses expensive components (dual actuators for all three axes with linear encoders and continous pressure monitoring), far too much for your university project.

But I have seen some amateur-built simulators (based on MS Flight Sim or X-Plane) that use springs to simulate control forces. The springs are tensioned by modified RC-aeroplane servos with increasing airspeed. Out-of-trim conditions can be generated by tensioning the spring only on one side of the control. Simple, cheap and accurate enough for a project like yours I think.

Regards, Max

[Pilot DAR]

bwfly,

IO-540, who frequents the Private flying forum would probably have some helpful thoughts for you... He has very helpful, and very experienced with that type of aircraft....

[ZFT]

The only real main difference between level D & C Simulators are that you are required to run more tests, but these are Motion and Sound related, not the Controls

And if we are going to be really pedantic, it's not motion but vibration tests that differentiate between C & D

[bwfly88]

Sorry should of been more clear, it is a real Sacota TB9 plane with it's tail and wing tips chopped off and it's based on MS Flight Sim supported by a 180 degree visual and simulated instruments so it isn't a certified sim or used for 'actual' training. It's just used by students at the University to support their studies. For my project I have been tasked with looking at how to improve the flight controls. It uses software alongside Flight sim and labjacks to power the motors. I need to look at writing some code using the speed variables outputted from flight sim and research some motors powerful enough to drive these forces.

But I have seen some amateur-built simulators (based on MS Flight Sim or X-Plane) that use springs to simulate control forces. The springs are tensioned by modified RC-aeroplane servos with increasing airspeed. Out-of-trim conditions can be generated by tensioning the spring only on one side of the control. Simple, cheap and accurate enough for a project like yours I think.

What Next, That sounds like what I'm looking for. Thank you for your info!

[John Farley]

To replicate the V squared control forces af a specific type from your starting point (very dubious aero package and probably no real aircraft data) is simply going to be impossible so I would suggest it would be very much more useful (for you and your colleagues) to mod the sim so that you can use it to teach yourselves about and experience some of the important aspects of all aircraft manual control systems.

To do this you would want as a starting point any benign simulation that is easy to fly (there are plenty of those on MSFS). Then you provide the ability to select/superimpose a range of the control system problems that plague real aircraft.


In no particular order:
• Breakout forces
• Friction in the linkage
• Sticktion in the linkage (a very different animal)
• Mechanical backlash between you and the control surface movement
• Control sensitivity that is too high
• Control sensitivity that is too low
• Cable stretch (in real life the same effect can be caused by the flexing of pulley mounting brackets)
• Different types of control force feedback (spring centring and V squared related)

See what your Prof thinks.

[SimWes]

And if we are going to be really pedantic, it's not motion but vibration tests that differentiate between C & D

OK, let’s go halfers on that one. According to the JAR-FSTD-A it’s “Motion Vibrations”

Not pedantic, merely trying to point out that the Control characteristics are not limited to level D machines!

[airpolice]

bwfly88, Can you share any more details on the visuals used in this?

MS Flight Sim supported by a 180 degree visual and simulated instruments

[ZFT]

SimWes,

I fully agree(d) with you and 50/50 it is.

Cheers

[Mechta]

There was an RAeS lecture by Dr Guy Gratton in Farnborough very recently in which he described his research into the relationship between stick forces and accidents. His conclusion was that low stick forces, which are usual seen by pilots as desirable, are actually a major contributing factor to accidents in light aircraft, as pilots inadvertently stall the aircraft whilst engaged in other cockpit tasks.

http://bura.brunel.ac.uk/bitstream/2438/3255/1/SETP_LA_JOLLA_Mar20_final.pdf

Flight test: Supporting the investigation of factors affecting loss of control of light aircraft | SFTE European Chapter (http://www.sfte-ec.org/node/633)

If you could create variable force feedback, you may be able to demonstrate this in your simulator.