Is it necessary for an elevator / horizontal stabiliser to produce a down force?

If the elevator is producing a lifting motion, then the centre of gravity must be behind the wings. Otherwise it would go nose down.

I imagine on a GA aircraft, with a big heavy engine up front, getting the c of g rear of the wings would be difficult with just a pilot onboard.

Well, if you are able to load the plane with a c of g quite far to the rear, the plane should fly a bit faster and I assume that it is due to the reduce in down force of the stabiliser.

Big jets do have the possibility to adjust the position and such the weight of the fuel in the tanks, and as far as I’ve learned have to pump around some amount of fuel during cruise. This will affect c of g, and presumably will be held in a certain zone which produces either zero downdraft or even lift on the stabiliser.

Yes, an airplane will fly faster as the H stab is aerodynamically unloaded, though having it lift does not increase that benefit. In completely level cruise flight, this would be an efficiency advantage. However, stability, and in particular stall and spin recovery characteristics suffer terribly. Airplanes (well, GA airplanes, anyway) are designed so that as the plane stalls, if you loose elevator effectiveness, the nose will drop, to aid recovery – because the tail stalled, and stopped lifting down. To be able to stop lifting down, it had to be lifting down in the first place.

I’ve done aft C of G flight testing, and it’s really scary recovering spins in that condition. Holding the control wheel fully forward, and watching the horizon go around in front of you, but the nose not really going down, is highly alarming. So, for a long cruise flight, I will load to an aft C of G, within limits. For maneuvering, practicing emergencies, or circuits, I’m happier more forward C of G.

A worse case scenario that must be considered at least for single engine part 23 aircraft is that in a spin the gyroscopic force from the spin, may the cause both the wing and tailplane to remain stalled i.e. flat spin. In that case you want the nose to drop on it own, at least enough to keep the tailplane in an effective range.

Increasing the camber which is what effectively would happen on a tailplane that was designed to normally produced upward lift, may simply be ineffective once stalled.

For something like the MAX you can probably can dispense with all this, but look where that ended.

Simply answer to the question: No, it is not a certification requirement (CS-23 or CS-25). In fact, it is not a certification requirement that an aircraft must have an HTP (horizontal tailplane) in the first place!

The certification worry more about the stability and control response as seen by the pilot. A pitch perturbation must “dampen itself out” rather than “amplify” and lead into an unstable situation. A pull force on the stick shall reduce the airspeed, a push force shall increase the airspeed.
How this is achieved is not so important.

The stick force gradient must not reverse until the stalling angle of attack. This is what happend with the 737-MAX initially under some conditions and the reason why the MCAS was introduced.

There are two things to consider here:

1. The aircraft must be stable in pitch, that is, a pitch perturbation that changes the angle of attack must be “dampened out” rather than “amplified”, the aircraft must return to the original condition before the perturbation.
2. The aircraft must be trimmable, that is, it must be possible to achieve the above point while also flying straight and level, with overall positive lift.

In mathematical terms, the above conditions translate to requiring that the slope of the aircraft Pitching Moment (Cm) versus Angle of Attack (AoA) curve be NEGATIVE, and that this curve intersect the y-axis (Cm0) at a POSITIVE value:

To achieve the negative slope condition, the aircraft CG must lie FORWARD of the Neutral Point (NP). This imposes a constraint to the aft-most location of the CG.

To achieve the positive Cm0 condition, several solutions exist. But because conventional aircraft have wing airfoils with positive camber and hence NEGATIVE Cm0’s, the only way to compensate for this while also employing a horizontal TAIL plane is for this HTP to provide a DOWN force. If a canard (fore plane) is used, then its lift will point upwards.

Tailless aircraft employ sweep-back and wing twist, so that the the angle of incidence of the wing tips with respect to the wing roots is smaller, in effect creating a “virtual tailplane”. They can also employ trailing edge devices, negative camber, or a combination of all the above to ensure an overall positive Cm0.

Peter wrote:

I have spoken to some airline pilots over the years who said this is not the case on a big jet, and that it is quite normal for the THS (trimmable horizontal stabiliser) to exert an UP force (and then the training-edge surfaces provide a positive or negative control around that).

Do you know which aircraft type they were referring to?

Aircraft with a THS are trimmed so that the elevator deflection (with respect to the stabiliser) is zero when in trim.

Peter wrote:

Is this related to “decalage”?

Thanks!

Depends on what you mean exactly by decalage.
To me that’s the difference in setting angles between the wing and the stabiliser.
The setting angle is the angle between the root chord and the fuselage reference line or aircraft longitudinal axis.

Peter wrote:

Standard stuff e.g. a 777.

Part 25 jets may be in a different regime since they don’t have to recover from a spin AIUI. Stall, I don’t know but you aren’t supposed to stall them.

You’re not supposed to stall them but they are still required to behave “properly”, particularly at high angles of attack.
Here is a very good video on the 737-MAX MCAS issue which is related to this:

My opinion is that those pilots saying the HTP would be producing an UP lift were incorrect. I may be missing some other piece of information though.