MRFCS: Some tips

[amalahama]

Yesterday I was testing manual reversion and wohaaa! Aircraft is really hard to fly, impossible to me. It first starts with a heavy pitch-up (which according to "Flying the Hog during Gulf war" book is completly realistic :thumbup:) and then it starts to slip right and left. Stick is extremly sensitive and it's very easy to get in a spin.

 

Can you guys give me some tips to tame A-10 when MRFCS is active? Is really so hard in the real-life?

 

Regards!

[Mud]

Hahaha, after reading your post I figured I'd give that one a go.

Man, was I in for a nasty surprise!!

 

I flipped the switch at FL20 which was probably not my brightest move ever :)

[Bushmanni]

When in MRFCS elevator is kind of trimmed by stick movement. If you push forward it increases the elevator movement speed down. When you center the stick the elevator doesn't move. This is how it works in DCS. You can see it for yourself by flipping the MRFCS switch on the ground and watching from external view how the elevator and other flight control surfaces move.

 

In the real life MRFCS switches control mode to mechanical control. There's a system that unlinks the hydraulic actuators from control surfaces and links the mechanical connection directly to control surfaces. In the case of ailerons the controls are linked to trim tabs instead of directly to ailerons. This results in slightly sluggish control but requires lot less control force. Rudder and elevator are stiffer as a result of mechanical control. This knowledge is based on A-10A operators manual as I have understood it. I think the MRFCS elevator implementation in DCS is wrong but then again if it was realistic there wouldn't be much difference to normal mode which is not true in real life as it's not possible to simulate real control forces even with a really strong force feedback joystick.

[amalahama]

When in MRFCS elevator is kind of trimmed by stick movement. If you push forward it increases the elevator movement speed down. When you center the stick the elevator doesn't move. This is how it works in DCS. You can see it for yourself by flipping the MRFCS switch on the ground and watching from external view how the elevator and other flight control surfaces move.

 

In the real life MRFCS switches control mode to mechanical control. There's a system that unlinks the hydraulic actuators from control surfaces and links the mechanical connection directly to control surfaces. In the case of ailerons the controls are linked to trim tabs instead of directly to ailerons. This results in slightly sluggish control but requires lot less control force. Rudder and elevator are stiffer as a result of mechanical control. This knowledge is based on A-10A operators manual as I have understood it. I think the MRFCS elevator implementation in DCS is wrong but then again if it was realistic there wouldn't be much difference to normal mode which is not true in real life as it's not possible to simulate real control forces even with a really strong force feedback joystick.

 

I think you're not right. I checked yesterday the A10A-1 and it's said that pitch commands in MRFCS mode are directly linked to stabilizator trim tabs. The only direct link is to the rudders.

 

Regards!

[Bushmanni]

I checked my copy of A-10A Flight Manual which I downloaded from the link provided in the DCS A-10C forums and it clearly states (page 58, description of MRFCS) that in MRFCS the yaw and pitch control is connected directly to respective control surfaces and roll control is linked to aileron trim tabs.

[amalahama]

I checked my copy of A-10A Flight Manual which I downloaded from the link provided in the DCS A-10C forums and it clearly states (page 58, description of MRFCS) that in MRFCS the yaw and pitch control is connected directly to respective control surfaces and roll control is linked to aileron trim tabs.

 

Oh yes you're right, sorry. I misunderstood the PITCH MRFCS text. Maybe manual reversion mode has changed in the -C and now elevator tabs are directly controlled by the stick? Who knows? :dunno:

 

Regards!

[elchacal]

Thanks for the advice! :thumbup:

[Aginor]

I tried MRFCS yesterday night and, well....

My first two tries were really bad as well.

 

But then I remembered what I read in the book "Warthog" and I tried it. Just fly VERY carefully. Never raise the nose more than 5 degrees or lower it more than 5 degrees. Don't bank more than 20 degrees. Move the stick very slowly and not too far because the reaction of the airplane won't come immediately.

Be careful with the throttle since it has a huge impact on how the plane flies. Keep speed somewhere between 180 and 250 knots.

 

I managed to land at Batumi after approximately 15 minutes of descending and correcting my course. (i.e. flying huge circles)

I landed without flaps or speedbrake. A shallow approach is recommended because you shouldn't move your stick a lot while landing.

I took the whole runway though, and I think I can do it better. I will try it again now.

I can upload a track as soon as I have resolved my huge problem that all my tracks don't work. (airplanes start to do strange things and crash in all my replays....)

 

EDIT: MAN I make typos all the time.... corrected.

Edited August 24, 2011 by Aginor

[Tango]

I found the stick to be very erratic??? Is this normal? It was very weird in pitch - I had a massive and sudden pitch down for no good reason.

 

Overall, when it was stable it was quite nice to fly. The stick behavior however.... why was it wandering around on its own? Speed was stable at 280 kts.

 

Best regards,

Tango.

Edited August 24, 2011 by Tango

[Aginor]

I cannot recommend flying that fast.

 

The reason for the weird pitch - which is completely normal in MRFCS - is that you are not really using your elevators but controlling the trim tabs which slowly move the elevator where it should go. The trim tabs are like little elevators that control the elevators the same way the elevators control the flight of the plane itself. The rudders jitter in the wind like a flag and every time the AOA changes they jiggle differently, because there is no hydraulic pressure keeping them in the desired position.

 

So the plane might suddenly pitch up or down every time you change direction or change your throttle setting. You have to apply minimal pressure on the stick, it is very sensitive. But it does not always react immediately to your commands.

You might even want to revise your axis settings to have better control in MRFCS mode (and also in normal flight mode although if you are used to it you might not even notice it is has a too steep curve).

[Tango]

Hi,

 

My axis are linear as I have the Warthog HOTAS.

 

If I understand correctly, the stick commands the tabs VIA the MRFCS and not directly?

 

Best regards,

Tango.

[Aginor]

That's right, because your elevators and ailerons are normally hydraulic systems.

If the hydraulic fails (problems with the pump, the engine, the hydraulic lines etc.) then you have absolutely no control.

Your control surfaces drop and you have to control them somehow in order to stay airborne.

The trim tabs are electrical systems (like the servo controlled rudders in radio controlled model airplanes) controlled through the MRFCS which change the flow of the air around the control surface in order to move it.

 

I wonder if that system is simulated correctly, though.

That's because I think I remember seeing somewhere that you have no elevator control via the stick at all in MRFCS mode, but use the trim hat on the stick to fly the airplane...

Maybe it was like that in the A10A and it was changed though. I don't know.

[amalahama]

It's true, right now in DCS:A-10 pitch commands are performed by trim tabs. But this behaviour doesn't seem to be right. In A-10A Flight manual is stated that control stick is directly linked to elevator actuators and only aileron tabs are linked to the stick. But again, I don't know if it's the same in the A-10C. Although I see quite doubtful that they have changed such fundamental backup system. Still waiting to PalkrII, yo-yo or others to clarify :)

 

Regards!

Edited August 26, 2011 by amalahama

[Aginor]

Maybe they changed it when they upgraded the A10As with the LASTE system, some time before the A10C was born.

But I agree, it would be nice to know how it really is like, I'm a bit confused, too.

[Garfieldo]

Just got a MFRCS single-engine landing done with just blowing one tire.

 

Do you really need to trim to change pitch? Since I only used the stick that would explain my more than shallow approach :doh:

[Yo-Yo]

I found the stick to be very erratic??? Is this normal? It was very weird in pitch - I had a massive and sudden pitch down for no good reason.

 

Overall, when it was stable it was quite nice to fly. The stick behavior however.... why was it wandering around on its own? Speed was stable at 280 kts.

 

Best regards,

Tango.

 

I must clear the things up. As we can not model high forces required from pilot on the stick the model simulate it stopping the virtual stick movement if required force is more than 500 N.

 

In manual reverse the joystic deflection from the neutral presents the FORCE applied by the pilot, so at zero speed the constant force drive elevator up or down if joystic is moved from the neutral.

At high speed you must use trimmer rather than joystick keeping it in central (trimmed) position. It's a good idea to use it only to slight changes in pitch/balance.

At low speed as the hinge moment get lower you have more room for joystick control.

 

Uncontrollable pitch movement are encountered if piloting became too agile... but if you can not recover the plane from unwanted attitude USE TRIMMER.

 

For example, you let the plane dive, speed raises - and you have less room to fix the attitude using joystick because of higher hinge moment.

 

 

But trim tab is probably reversed... it's a bug.

Edited August 30, 2011 by Yo-Yo

[Aginor]

Thanks for the explanation, Yo-Yo!

[Bushmanni]

Now that Yo-Yo explained the MRFCS simulation model it seems very logical but it's still somewhat counter intuitive. It would be good to explain the MRFCS simulation model in the manual.

[Aginor]

I admit that I was also a bit confused with how it worked, although I was able to fly around and even land my plane.

 

I think I agree with Hender. It is clear that it is not possible to simulate the system realistically because there is no FFB joystick that provides such a strong force, but simulating the problems the pilot has to cope with in MRFCS mode by making the stick feel more sensitive when in fact you would have to pull more strongly to do anything at all is a bit counter-intuitive.

 

I'm not sure what to do instead, though. But I think you guys from ED will figure it out, since you got real pilots to tell you what feels best. :)

[Yo-Yo]

This isn't right. Don't do it this way. A 20% input should move the in-game stick in position 20% unless the force required is too much. If 100% stick is 500N and 20% is 100N then it is impossible to fly this way. A spring-loaded gaming joystick is a displacement controller. It must be tied to the displacement of the in-game joystick. You cannot map controller displacement to stick force directly or it's hopeless for the user. You can provide limits to displacement based on limited pilot arm strength but joystick position must set the desired position of the in-game control.

 

Besides even in MRFCS the artificial feel device is still connected so the stick forces should be always as strong if not stronger than normal. It should never be "weak spring" even at 0 knots air load. In MRFCS the stick forces are never less than PFCS.

 

No wonder MRFCS pitch control next to impossible! No real A-10 pilot would be able to cope with such a weird control method.

 

Formally you are right... but there are two main reasons to do how it's done. The first reason was, that if we made position oriented method there would be no significant difference between boosted controls and manual regarding the accuracy of piloting. By the way, some years ago I tried Su-27 (or something similar) procedure simulator at MAKS and it has about 16-20 kG (official comment :) ) for full stick deflection... not 50! The FM was very similar to our SFM but my hopes for accurate piloting was broken. There was not a fail but I barely performed 2 or 3 loops and barrel rolls and landed.

Hang a 16 kG weight to your arm and try to write your name on the wall.

 

THe second reason (minor) was that booster model operates with FORCES. The forces from the hinge moment and the force of booster piston due to the metered pressure in cylinders. Of course we could write the gear that converts desired displacement to the force... but see reason #1.

 

We foresaw a lot of complaints that MRFCS is too easy to fly. I think now it meets the main MRFCS function - get home and maybe land if the weather is ideal. If you use trimmer it's not that hard.

Edited September 1, 2011 by Yo-Yo

[Yo-Yo]

Those are bad reasons. MRFCS has real downsides, you don't have to make up false, fake, unrealistic, wrong ones to make it harder. SAS is locked out, limited pilot strength prevents controlling large stick forces, roll control lags using tabs only.

 

Please simulate not fantasize.

 

What to do instead is this:

 

1. Player input is desired stick position

2. Stick assumes desired position if stick forces are normal

3. Stick assumes approximate desired position if stick forces are high

4. Stick ignores desired position if stick forces above maximum pilot strength

 

If stick force is 90% of pilot strength the stick gets close to desired position but jumps around, takes a while to pull, has error because pilot can't be accurate using 90% strength. At 75% it's better. At 50% it's better again. At 10% pilot can move stick exactly where he wants.

 

There is only one person here who are fantasizing... I am afraid it's not me. if you think that we are so ignorant that we do not know how the real stick must behave - you are definetely wrong. The rules you posted here is a kind of blah-blah-blah... no good math to detail it? What does it mean "jump around"? Sinewave? Sawtooth? Any kind of tremor? Do you know the best solution? And do you know that artificial unwanted input will disorient you more violent than the way you handle it now?

 

And what do you know about the stability of high-order nonlinear inertial feedback systems the whole system of plane and pilot is? Anyway, if we make the system close to your desire it won't be close to the real feel because of very complicated coupling with the joystick and virtual stick.

 

Nevertheless, I do not want to say that the current model is the best and is not a subject to change. If we have better solution in terms of real feel of high forces on the stick we will change things. Not really now.

[Speed]

In manual reverse the joystic deflection from the neutral presents the FORCE applied by the pilot, so at zero speed the constant force drive elevator up or down if joystic is moved from the neutral.

 

I thought movement on real joysticks followed Hooke's law, i.e.:

 

F = -k*x,

 

where k is a constant (in units of force over displament), x is displacement, and F is force. It's a very simple equation that applies for most situations of mechanical strain. So basically, what this says is that a force directly relates to a displacement through a simple constant.

 

Even if the force on a joystick didn't follow Hooke's law exactly, it would still have to follow some kind of deflection-based equation, maybe something like F = -Ax^2 - Bx

Otherwise, it would be implying that there is no restoring force on control sticks to try to return them to center. Just like a box you move across the floor- it takes force to move it somewhere else, but when you release it, it doesn't slide itself back to the point where it was before you started pushing it. Unlike sliding a box across the floor, there has to be a restoring force of some kind on real flight control sticks, so that when you release them they return to center, right? (unless it's a Ka-50 and you hit the trimmer ;))

 

I'll have to give MRFCS a try sometime, I've actually never had to use it in combat! I agree with the need to make the MRFCS harder to use than regular control though! What about a forcing a joystick curve on people that makes it so they, at max, only get like 10% of the normal control stick deflection that they normally get? Is that possible, and would it be an adequate solution?

 

Anyway, personally, I wouldn't want the ED team pulling their hair out and wasting too much time on such a rarely used feature unless it was easy to fix/tune... there are a lot of more important things to do :)

Edited September 2, 2011 by Speed

[Speed]

I think I agree with Hender. It is clear that it is not possible to simulate the system realistically because there is no FFB joystick that provides such a strong force, but simulating the problems the pilot has to cope with in MRFCS mode by making the stick feel more sensitive when in fact you would have to pull more strongly to do anything at all is a bit counter-intuitive.

 

I disagree, it is possible to simulate this somewhat realistically. You are forgetting force-sensing sticks. I am not sure at what level of force it would take to damage my X-65's force sensors, but in its least sensitive mode, I have to exert very powerful forces to get it to deflect fully. I'm sure that the forces I apply are not even remotely as high as in a real MRFCS situation, but if you made it so that full deflection on the X-65 at its lowest setting would only move the simulated plane's control surfaces like 15% of their normal movement range, then I suppose you could simulate MRFCS for the likely deflections a pilot would be using. A pilot would be saving his/her strength, moving the plane in very slow turns and rolls.

Edited September 2, 2011 by Speed

[Bushmanni]

How about changing the axis response curve continuously in random fashion but only the parts where the stick isn't currently. And this random variation would be zero at and near the center position and gradually increase towards maximum deflection. If you pull 50% back, the curve would stay where it is at 50% and randomly fluctuate around it. When you pull more to 60%, the 60% position on the curve would then stay where it was when the stick moved there but all other points around it would fluctuate. This would mean that when you keep the joystick stationary, virtual stick won't tremble but when you move the joystick there's some randomness to how it relates to virtual stick. This would make it so that the player input is somewhat the desired stick position but you can't make accurate virtual stick movements when the control force is high. You could also make it so that when the airspeed increases and control forces get stronger the axis curve saturation would be reduced accordingly.

 

The most strangest thing with the current MRFCS implementation is that the controls are more sensitive at high speed and get less sensitive at slow speed while it should be just the opposite. When the required control force is high you should need more joystick travel to achieve the same virtual stick movement.

[Yo-Yo]

How about changing the axis response curve continuously in random fashion but only the parts where the stick isn't currently. And this random variation would be zero at and near the center position and gradually increase towards maximum deflection. If you pull 50% back, the curve would stay where it is at 50% and randomly fluctuate around it. When you pull more to 60%, the 60% position on the curve would then stay where it was when the stick moved there but all other points around it would fluctuate. This would mean that when you keep the joystick stationary, virtual stick won't tremble but when you move the joystick there's some randomness to how it relates to virtual stick. This would make it so that the player input is somewhat the desired stick position but you can't make accurate virtual stick movements when the control force is high. You could also make it so that when the airspeed increases and control forces get stronger the axis curve saturation would be reduced accordingly.

 

The most strangest thing with the current MRFCS implementation is that the controls are more sensitive at high speed and get less sensitive at slow speed while it should be just the opposite. When the required control force is high you should need more joystick travel to achieve the same virtual stick movement.

 

Try to compare virtual stick and joystick movements at various speed. At high speed you have less room for stick movement but the plane is less sluggish becouse MOIs are the same and the moments increase.

By the way , there is no special plane control for MRFCS. If you have the same deflections of controls you will have the same behaviour of the plane.

 

 

I disregard that external random inputs can simulate the desired effect because when you move something under a pressure there is no fluctuation during the travel. You just can not stop your movement where you want exactly overshooting or undershooting.

If you try to replace real pilot in these condition it must be feedback system, nonlinear with strange kind of dry friction... very strange because in RL you can get the desired position of the stick after some iterations.

 

Regarding saturation - you have saturation now when you can not overcome stick force.