Yokes that work better (brain storm thread)

[Hardy Heinlin]

Hardy,
Thank you, i'll check that out.  If the flight hardware isn't calibrated every time, is it normal for values to jump around when they're not touched?  It can't be, because I even have axis' that aren't configured jumping by .1 such as mixture and propeller on the throttle quadrant.  I'm going to do some trouble shooting but I may end up calling Redbird.

The only systems I know that stay at 0.0 when released are hall sensor systems. Regular potentiometers are never 100% stable. I have even seen jumping indications on the real 744 EICAS on the flight control indications during preflight checks; the controls are, of course, hydraulic mechanisms, but the surface position indication system for the avionics uses electrical RVDTs (rotary variable differential transformers).

[Jason M Tutwiler]

thank you, that makes me feel better.  I'm going to re calibrate everything. 

[the mad hatter]

http://bffsimulation.com/CL_Controller.php     Ian suggested that for those that are interested, for those that are serious about adding something like this to their sim let me know and we could do a group thing, if not I will carry on my merry way and go it alone   thoughts?

[SwissCharles]

@ Bernard in Austin 
Whew... 10K... well... for all axes... ain't that much over the top, huh?

Unfortunately I am still too far away - need to be retired a little while longer til I know how to do up/down/left/right as somebody preached to me quite a while ago.. 

Perhaps if interested parties let themselves be known here, from a point on an estimate on the impact per party could be guessed?

[b744erf]

Using all available data i could find everywhere (not a lot), I made a chart of what the stick force should be vs the psi supplied to the elevator feel unit/actuator

http://farrokhchothia.com/slists/elevfeel/FEEL%20TEST.pdf

I have the oem 744 feel unit below the floor, with a variable hyd press valve working on values from psx -

Hi Farrokh, your solution is absolutely fantastic. This is the first‑time I have seen a real‑hardware implementation of the elevator‑feel and centering mechanical system. I feel really fortunate to stumble across this six‑year‑old forum thread today.

I have recently finished the interface and mechanical rework for my full cockpit. The cockpit shell is a retired B165 unit I sourced from an automotive scrap market. It has taken me over a year to restore it. All avionic functions and basic control functions are now working, and I am currently working on the control‑loading mechanisms.

My current implementation:
To achieve the 100 % centering effect Hardy described, my elevator, aileron and rudder axes all use spring‑based mechanisms with a defined mechanical neutral point. The core components are several Boeing 777 aileron breakout‑system linkages I acquired. These OEM spring linkages can stretch and compress, feature a distinct mechanical neutral point, and come with high inherent spring force. I have machined down the springs to lower the breakout force at the neutral position.
The elevator loads are relatively high due to the heavy physical yoke mass combined with large moment arms. Rudder loads are lower, as pedal moment is less than that of the elevator. Aileron loads are very light, which closely matches real‑aircraft behaviour.

‑ Elevator: this 777‑series spring linkage combined with heavy springs plus dampers provides control loading.
‑ Aileron and Rudder: dampers working together with these breakout‑system linkages.

This mechanical setup, however, has significant drawbacks.
In a real airliner, pilots do not feel a hard mechanical detent at neutral. Real aircraft rely on hydraulic or electrically‑driven systems for centering. The full control‑column travel feels completely smooth, with no noticeable click or detent when passing through neutral. Upon releasing the controls, the column slowly drifts back toward neutral under damping.

With my current hardware, the control column produces an obvious click and detent feeling as it crosses the mechanical midpoint, which does not match real‑world behaviour.

I also run into a centering‑accuracy conflict.
After releasing the controls, whether pushed forward or pulled aft, the column should ideally return to the exact same neutral position. Yet the zero‑force point of a spring assembly is a range rather than a precise geometric point. When I set the damper to high values for realistic feel, the residual spring centering force is insufficient to overcome damping, so the column stops short and settles offset from neutral.
For a detent‑free purely‑spring mechanism to reliably return to exact centre, it must operate with zero friction and zero damping. Only under those conditions will it consistently hit the neutral point, since that neutral point is inherently a zero‑force range for the springs.

I do not want to mask this by increasing the dead‑zone inside PSX. Larger software dead‑zones reduce control sensitivity, nullify small control inputs and degrade simulation fidelity.

I watched YouTube videos explaining the Boeing 777 control mechanism, described as a mechanical centering system combined with dampers. This leaves me puzzled: springs without positive detents cannot achieve this on their own. Real aircraft must use external assistance such as hydraulics to accomplish proper centering.

I would like to ask Farrokh, Hardy and other builders a couple of questions:

1. In home‑built simulator projects, how is high‑precision hardware‑level centering achieved? Without software‑based neutral‑point compensation: how do you get the controls to overcome damping and reliably settle at the geometric neutral position, while avoiding audible clicks and detent‑type feel at mid‑travel?

2. Control forces for elevator change dynamically when the aircraft is out‑of‑trim. I plan to replicate this behaviour with either electric or hydraulic hardware. Which variable / Q‑parameter inside PSX should I read to detect an out‑of‑trim condition?

Thank you.

Here is my latest test video of the simulator  would like to share with you guys

https://www.facebook.com/share/v/1N2cJpbFsK/?mibextid=wwXIfr

Regards

Jack

[Hardy Heinlin]

Hi Jack,

this Qi tells you the pressure in PSI units generated by the elevator feel computer:

Qi211="ElevFeelPsi"; Mode=DEMAND; Min=0; Max=2100;

https://aerowinx.com/board/index.php/topic,2514.msg24654.html#msg24654


Regards,

|-|ardy

[b744erf]

Hi Hardy, thank you very much for your quick reply.

I notice this value matches exactly with the "Elevator Feel Psi" reading under the Miscellaneous page in the Instructor‑Analyze menu.
However, I still have some questions. Since this value reflects elevator‑control‑column feel force: when the aircraft is out‑of‑trim, the physical force required to push or pull the control column changes accordingly. I need to convert this PSI value into actual physical push‑pull force for my hardware‑loading mechanism.

What PSI value corresponds to the mechanical neutral point of my control column?
For example, the aircraft is in‑trim at stabilizer‑trim position of 6 units. Nevertheless, ElevFeelPsi still varies with aircraft weight, airspeed and other flight conditions. It does not have a simple one‑to‑one mapping against stabilizer‑trim position alone.

Could you please advise how I should handle this conversion inside my script?
I would also appreciate any insight from Farrokh as well.

Thank you very much.

Regards
Jack

[Hardy Heinlin]

Hi Jack,

I'm not sure I understand your idea. I mean, when the aircraft is out of trim and you constantly have to hold the yoke outside its center, it's the normal self-centering force that is pulling the yoke to its center. The elevator feel computer will not increase the force for any out-of-trim situation. The elevator feel computer increases the force when the airspeed (EAS, equivalent airspeed) rises or when the CG is more aft, because when these factors increase, the elevator gets more effective. And when the elevator gets more effective, the pilot may more easily oversteer the pitch control. To avoid oversteering, the feel computer increases the centering force; it's the same force on the pull-side and push-side. It's symmetric. There is no center shift. This is not a Cessna elevator. It does not relate to a specific out-of-trim situation in a specific direction. The only relevance to the trim position is this: The elevator feel computer doesn't really know where the CG is. But it does know where the pilot has set the trim. From this trim setting the computer derives where the CG is. Whether the pilot is currently in an out-of-trim situation, doesn't matter. The computer doesn't know the pilot's intention, be it a climb, descent or level flight, or speed change or whatever.

Maybe I'm misunderstanding your idea and you already knew this? :-)


Regards,

|-|ardy

[b744erf]

Hi Hardy,

Perhaps I have misunderstood how the 747 elevator‑trim system actually works. To put it simply, I haven't gone too deep into the theory. Some real‑world pilots told me that when the aircraft is out‑of‑trim, the push‑force and pull‑force on the yoke feel different. For example, if pushing feels heavier than pulling, pilots would naturally adjust the stabilizer trim nose‑down until push and pull forces feel balanced.

This is only verbal feedback from pilots, which I cannot verify by myself, and different pilots describe the sensation in different ways.
Could it be that what the pilots are describing is not truly‑asymmetric push‑and‑pull forces, but rather that, to maintain the desired attitude while out‑of‑trim, you have to hold a larger deflection away from the mechanical neutral point? The larger offset from centre then creates a heavier‑feeling load. Is that all there is to this subjective sensation?

That is one reason why I want to ask real 747 pilots on this forum to share their actual yoke‑feel experience and clarify whether this phenomenon exists.

According to what you explained, the feel computer only varies the overall force level: when airspeed increases, the elevator yoke becomes heavier, and this added force is applied symmetrically for both push and pull directions. Am I understanding that correctly?

It may not be that you misunderstood me; rather, I myself do not fully grasp this mechanism, so I am trying to get clarification from everyone here.

For my hardware implementation, I would use both electric motors and mechanical springs and dampers to simulate these force sensations, rather than hydraulics. If push‑pull forces are indeed symmetric, implementation will be much easier than expected, because detecting whether the aircraft is out‑of‑trim is quite challenging for me.

I also have one more quick question: On the 747, the control cables do not link directly to the elevator surface; they connect to hydraulic actuators. Unlike the 737, does that mean aerodynamic loads, wind‑induced effects and airframe‑vibrations only act on the output side of the hydraulic actuators, and these disturbances are not fed‑back to the pilot's yoke input side? If so, would that make the 747 yoke‑feel easier to simulate than that of the 737?

Regards,
Jack

[Hardy Heinlin]

"Some real‑world pilots told me that when the aircraft is out‑of‑trim, the push‑force and pull‑force on the yoke feel different. For example, if pushing feels heavier than pulling, pilots would naturally adjust the stabilizer trim nose‑down until push and pull forces feel balanced."

Yes, and when balanced, you can put your hands off the yoke. For example, you want to lower the pitch and keep the pitch there. So you push on the yoke and hold it there. In order to relieve your muscles, you trim the nose down, and while the trim moves down you see on the flight instruments that you need to relax your muscles, otherwise the nose would go below your target pitch and this is not what you want. Now, while your muscles relax more and more in order to maintain the target pitch, the yoke goes more and more back to the center. When the yoke has reached the center, the aircraft is trimmed for your target pitch and you can put your hands off the yoke.

On a Cessna 172, for instance, there is no stab trim. The 172 has an elevator trim which is not the same as a stab trim. An elevator trim is a small extra panel within the elevator panel. If the small trim panel goes up, the wind stream will push the large elevator panel further down causing the entire elevator force center to be further down. And because the elevator panel is mechanically linked with the yoke, the yoke's force center in the cockpit shifts forward. This won't happen in the 737 or 747 because those types have a stab trim instead of an elevator trim. The stab trim rotates the entire horizontal stabilizer surface. And this rotation will not change the center point of the elevator control. Whatever angle of attack the stabilizer is set to, the elevator surface (in neutral position) is always inline with the stab surface. When not, the aircraft is out of trim. Then you need to trim it until there is no knee anymore along the line. When a Cessna is trimmed there may be a knee along the line because the Cessna's stab cannot be rotated; this causes some drag, and that's one reason jets use stab trim instead of elevator trim systems.


"According to what you explained, the feel computer only varies the overall force level: when airspeed increases, the elevator yoke becomes heavier, and this added force is applied symmetrically for both push and pull directions. Am I understanding that correctly?"

Correct. And this is no contradiction to what the pilots told you. The one thing is not related to the other thing.


"On the 747, the control cables do not link directly to the elevator surface; they connect to hydraulic actuators. Unlike the 737, does that mean aerodynamic loads, wind‑induced effects and airframe‑vibrations only act on the output side of the hydraulic actuators, and these disturbances are not fed‑back to the pilot's yoke input side? If so, would that make the 747 yoke‑feel easier to simulate than that of the 737?"

I think so.



Regards,

|-|ardy

[b744erf]

Hi Hardy,

Thank you for your detailed reply. Now I understand. When I fly the Cessna 172, I can clearly feel that the yoke‑neutral reference point shifts forward or backward. I already know that for the 737 and 747, the mechanical neutral position of the yoke itself never moves. My confusion has always been about the handling‑force sensation. Thanks to your explanation, I have a much clearer picture now.

Following your description, I will set variable damping according to the ElevFeel‑PSI value to reproduce the noticeable change in control‑stick heaviness. I plan to implement this variable‑damping effect with electric motors.

Another challenge is the self‑centering behaviour. On the real aircraft, the yoke returns to neutral extremely smoothly without any stiction or jerky movement. Achieving such silky‑smooth recentring is quite difficult using only mechanical springs and dampers, so I will most‑likely rely on motors for this part as well. I will keep experimenting.

If other forum‑members have proven‑good solutions for smooth centring‑mechanisms, I would be glad to hear your suggestions. I have heard that FOC motor‑control technology works very well for simulating control‑yoke feel. Does anyone here have practical experience? I would appreciate any guidance.

Regards,
Jack