B-737 Speed Trim System
 

I don’t understand the meaning of «return the airplane to the trimmed speed». And if you need force on the controls, in my head, this makes the aircraft anything but speed stable. Or in trimmed speed, whatever that means.
If I let the aicraft STS do it’s job without interference and let go of the controls, the aircraft will pitch up very fast.
It does not make sense to me.

That's self-explanatory, anything that changes is returned to the original state by making a change opposite to the first change (i.e., stable behavior). If you added a change in the same direction as the first change, the state will only move further away from the original (unstable behavior).

I think I understand where you're coming from, and if I'm right, it's due to a confusion between two different things that are meant by "speed stability." They're superficially similar (it's what, um, keeps the speed stable, right? duh!) but actually very different. And the distinction is a little tricky and more than a little subtle, but nonetheless important; and if you're having a hard time applying what I pointed out as self-explanatory to speed and trim, it's probably because you're thinking based on the first definition, while the system was designed (and explained in the FCOM) based on the second.

1. The "speed stability" in some pilot training materials:

This assumes that you're controlling for a constant altitude, be it with autopilot altitude hold, or manual elevator control. Whatever is controlling the elevator, and, in turn, the AOA (be it your arm or the autopilot) is free to do whatever it takes with the AOA to maintain that level flight path. Think of it as altitude-fixed, AOA free. So if a gust speeds you up a little bit, the extra drag will then slow you back down to the original speed. If a gust (or momentary dip in thrust, or anything) slows you down, the reduction in drag will speed you back up to the original speed. This works on the front side of the thrust curve. If you're slow enough already to be on the backside, however, then a momentary slowdown will see more drag, which will slow you down even more, which will cause more drag, etc. in a runaway cycle. So this type of speed stability only exists above Vmd. It's unstable when you're slower than that.

2. The "speed stability" in aeronautical engineering:

For this definition you have to completely forget about holding altitude. The airplane is free to climb/descent at the whims of the other factors at play. However, without elevator changes, AOA is maintained. So, opposite the fist definition, think AOA-fixed, altitude free. (Also for the moment, set aside the effects of thrust-pitch couple like the 737 is famous for, or propwash like in most bugsmashers.) This AOA-fixed behavior is the natural behavior of normal planes. After any disturbance (assuming the elevator and/or trim returns to the original position) the AOA will also return to its original position. It may be after a period of alternating overshoots (phugoid oscillation), but eventually it will.

Let's say you're cruising level in the middle of the speed envelope, and you reduce some power (permanently). The power reduction will cause a speed reduction, which will cause a lift reduction, which will cause the beginning of a descent, which will cause a component of weight to align forward parallel with thrust, which will cause the plane to speed up, and where it finishes speeding up will be its original speed.. (If it finished at any other speed, that speed difference would cause a lift other than the weight, and the plane would go through more repetitions of this same cycle until lift=weight. Since weight is fixed, lift is fixed, and the only 2 remaining variables of the lift equation are speed and Cl, and Cl is a stand-in for AOA. So the speed will settle at that speed which is corresponds only to the particular AOA you're flying at.) This is why there's a fundamental relationship between AOA and speed in vertically unaccelerated flight, and why AOA stability implies speed stability under the engineer's definition. To summarize, we pulled the power, but the speed remained the same (albeit under a descent, but we don't care about that). Vice versa, if we increase power the airplane will begin a climb, at the same speed.

This speed stability (definition 2), by the way, is required to exist throughout the entire certified envelope all the way down to stall speed. Remember, so far no electronic or mechanical trickery is involved. This is the natural, aerodynamic behavior of every stable plane from a hand-thrown model to the 747. To change the speed that is held constant we must change the AOA, which is done either by moving (and holding in a new position) the elevator or moving the stabilizer, i.e., actuating the trim. And in some engineering contexts, they don't care whether you end up having to hold a force with your arm or not. (This can be kind of confusing to us, who are used to thinking of trim as the-thing-that-removes-the-arm-force.) It's all just considered changing the trim point, or trim speed, based on the new position of elevator or stab. Now, this new speed will be maintained (again, maybe climbing or descending now, but that's irrelevant).

It appears that your scenario is based on the premise of the first definition. You're increasing speed while maintaining altitude, (intentionally) so the trim required is forward. (Faster speed = forward elevator and/or forward trim) But the STS, sensing that you're faster than original, trims aft in an attempt to slow you down to the original speed. This causes the plane to tend to climb which annoys you. (Of course, we're used to the premise of maintaining altitude) But this is where the engineering definition kicks in and you have to realize that the airplane doesn't care about the climbing tendency. It only wants to return to the original speed, which, being slower than the new speed, mandates a climb. This is the behavior of a normal airplane, only a little stronger with the augmented speed stability of STS. Understanding that (non-Airbus) airplanes really behave in accordance with the engineering (#2) definition, will hopefully let you come to terms with the "opposite" behavior.

Vessbot's description = My Head Hurts! :)

I shouldn't complain though, I don't even have a trim wheel... :ok:

You might want to re-read your FCOM. The STS only becomes active 10 seconds after take off and remains active while the autopilot is disengaged up to Mach 0.68.

It's not trying to stall the plane. The system does not know about you retracting the flaps and it does not need to.

Retracting flaps and accelerating always means you need to trim, with the STS you just need to trim a little bit more.
You are blaming the STS for the need to trim, but that requirement is always there.

So if you assume no flap changes (for which you will always need to trim) the system does indeed try to keep the speed stable.

Accelerating the STS will support the natural pitch up tendency of the plane and the pitch up will reduce the speed increase.
That is what speed stable is, the plane pitches up when you increase the speed.

The system is not meant to assist you trimming.
It is meant to emulate a plane with a bigger elevator in which you need to trim more for speed changes.

As MickG pointed out, takeoff+10sec is where it starts, not ends, being active, at last according to the quoted snippet from the manual. If the trim is moving prior to 10 seconds, I have no answer to that. 737 guys, (I'm not one) is there another system that's supposed to be automatically moving the trim prior to 10 seconds?

As far as the high weight and forward CG... the manual isn't saying that STS is only active at low weight. It's saying that the reason it exists (in general) is to fix some conditions that are more prevalent at low weight, etc. But it's always active. Kinda like the yaw damp, it's there to fix dutch roll, but that doesn't mean it's only active at high altitude; it's always active.

But most importantly, have I been able to answer your confusion about trim and speed stability in general? In your first 2 posts, you noted some really basic misunderstandings, such as not knowing what "returning the airplane to trimmed speed" means. Do you understand that now, and do you understand why a speed stability augmentation system would be expected to increase the stick force you're holding, if you're holding it away from trimmed speed?

The input is "sensing of trim requirement," can't you read? ;)

Overly dumbed down **** in manuals like this pisses me off. That provides zero value in trying to understand how the system works and why it's doing what it's doing... they might as well have saved the ink and replaced that whole section with "The trim is gonna move by itself a lot, don't worry about it it's normal."

To be more succint, "auto trim" means it trims to reduce stick force.

STS does the opposite, it trims to increase stick force.

"Auto anti trim," how bout dat?

This sounds bizarre to me, why wouldn't they just make the performance data tell you to set the correct trim, instead of lying to you by 1 unit, and then having the indicator lie to you by one unit the other way?

Paging Nigel Tufnel...


"Ours goes 1 unit forward..."

I have 9000 hours on the 737. 200/300/400/700/800. I thought I understood the STS. Having read and re read this thread..I am now convinced I know nothing. I must say that in day to day operations..I just let it do its thing..and it never seem to worry me. Perhaps ignorance is bliss!!

For full disclosure I've never flown the 737, and I only learned about the existence of the STS probably the day before I made the post. I only slotted the info from the quoted manual section, as well as Mad (Flt) Scientist's blurb from the Lion Air thread, into the basic framework of how trim and speed stability works in flight dynamics. So if there are any subtleties to the system, beyond what's written here, that might change how it works in a non-obvious way, I wouldn't know about them.

Hi all,

Although the STS demo is captured from flight sim it might be a good description of the workings of the system;


See what you think.....

Cool video! Looks like a bit of a Heath-Robinson setup though...

I suspect Boeing won't/can't because more stability more fuel burn means means less efficiency.

I suspect you're right.

I think it's more complicated than that - this is about stability in response to control movement (which is also why it's off when AP is on). On the R&N Lion Air thread I think there was a post about certification reqs for minimum stick force per knot (of speed change), although I've also seen "stick force per g" elsewhere. As far as I can see the Vessbot's description makes sense, and STS helps meet that cert. req. by increasing stick force (in effect, trimming against you), in certain phases of flight.

Since you're not pulling directly with cables, I suspect Boeing could have met the requirement by just making the control force larger, but that might have made them too hard to move in other parts of the envelope, where STS is inactive. But I'm not sure, I might have that wrong - my aerodynamics is probably past it's use-by date.

At the end of that video STS is disabled when flaps are retracted, that's not something Mr. Boeing put in his FCOM. Any experts to confirm or deny?

I asked myself the same thing earlier, why can't they fix the stick force per knot problem by tweaking the artificial feel unit? That's literally why it exists. Then as I wrote that post I figured out a possible answer... (but, again, this is just supposition.)

There are two aspects to the problem: First, the stick force per knot gradient, i.e., how much pull does it take to hold away from the trim speed. This aspect by itself is easily fixable with artificial feel setting.

But the second one is changing of the trim speed itself. When you increase the thrust like the 737 if famous for, the low thrustline offset tends to pitch the plane up, which is effectively the same as applying aft elevator and/or trimming aft. Say you're approaching at 140 knots, after you apply TOGA the trim speed becomes 100 knots (these are only example numbers). This is only fixable by retrimming.

There has been a lot of theorising, hand wringing, head scratching, chin rubbing etc on this subject. I quite like Switchbaits post #29... The system works fine, just fly the aircraft. Too much supposition can lead to misinformation and mistakes. What would be ideal is the system description and operation from the Aircraft Maintenance Manual to prove once and for all as to how the system operates and lay this topic to rest. Any engineers out there able to assist this think tank?

B737 pitch trim description
 

Please allow another try at describing pitch trim on the 737. Notice that I chose the word "pitch" trim rather than "speed" trim to be more general. I will address the speed trim system more directly below.

One of the sources of confusion (at least for me and maybe others) is that the word "trim" is often used to describe any motion of the horizontal stabilizer. This gets confusing in that "trim" or "pitch trim" also carries the notion of "moving the horizontal stabilizer so as to offload column forces" as in getting to a state where steady column input is not required to continue flying at the desired steady state condition. It is unfortunate that the word "trim" has been used for both of these.

Now we need to consider the FAR speed stability requirement. In general terms, the FAR states that an airplane should exhibit positive speed stability with a stick force of at least 3 lbs per 10 knots. That means that once trimmed (i.e., no column input required for steady flight) reducing speed by 10 knots (without a throttle or configuration change) should require 3 pounds of pull force to maintain steady flight. Similarly increasing speed (again without a throttle or configuration change) should require 3 pounds of push force to maintain steady flight. In this manner, the airplane (absent a column input) will tend to nose up/down appropriately to resist changing speed. If an airplane does not exhibit sufficient speed stability to meet this requirement, one solution is to have a control system function that moves airplane surfaces as needed to provide sufficient pitching moment as a function of speed changes to require the specified levels of column input to balance moments.

Meeting the stick force per knot FAR is the motivation for the speed trim system on the 737. For a limited portion of the flight envelope the bare airplane without any stability augmentation did not meet this FAR. To compensate, the 737 speed trim system moves the stabilizer in response to speed changes to augment the stability. Note that while this system is only needed to help meet the FAR at a limited set of conditions, it is for simplicity sake designed to respond to speed changes regardless of CG or weight. As a result, the speed stability is increased over a much wider range of conditions than just those where the airplane would not meet the FAR without it. For instance, at aft CG where the function is needed it will require just enough column when speed is changed. At forward CG where the airplane already has sufficient inherent speed stability, the STS will add more and could be seen as being an unnecessary nuisance.

As many have noted in previous entries to this thread, the STS "trims" (i.e., moves) the stabilizer opposite the direction that the pilot has to move to the stabilizer to relieve column forces. In a sense, the STS "un-trims" the airplane thus requiring the pilot to re-trim it. Whether one sees this as an unnecessary bother or a tool providing the flight crew with positive awareness of speed deviations is a matter of opinion. This gets to the heart of the A vs. B differentiation between their respective C* and C*U pitch augmentation systems - another topic that has been discussed in other PPRUNE threads over the years.

It is important to recognize that there are a number of other factors that impact pitch trim for which the speed trim system does not take any consideration. Changes to thrust, flap position, speedbrake setting, and gear position will generate pitching moment changes that must be balanced via the elevator (i.e., column when flying with autopilot disengaged) and followed up via pilot pitch trim inputs (i.e., movements of the stabilizer) to allow releasing the column. The 737 speed trim system essentially adds an increment of stabilizer motion in the positive speed stability direction as a function of a change in airspeed. It does not maintain a target airspeed that it seeks to return to regardless of other pitching moment disturbances that the airplane may experience. Other airplanes with higher levels of pitch augmentation (777 and 787 for example) do provide control surface inputs to counter such pitching moment disturbances and thus return to a specific airspeed, but that is another story for another thread.

Interesting, do you know why not? And if it only increases the stick force without maintaining the trim speed, why it uses the trim system instead of just adding more resistance to the artificial feel?

FCEng84: Good job. Since you are in town, maybe you could zip down to the 737 office and re-write their Systems description? :ok:

Excerpt of 737-300/400/500 MAINTENANCE MANUAL STS logic.....

The Speed Trim system provides automatic stabilizer trim for positive speed stability during low speed, high thrust conditions. The speed trim
system is only operational while the autopilot is off, and on 737-300 airplanes only, the speed trim system is automatically terminated when
the flaps are up. Either FCC can provide the speed trim commands and only one channel will be engaged at any one time. The system which will be in command is alternately chosen as a function of the squat switch input, i.e., if FCC A was in command on the last flight, the FCC B will be in command this flight.
The engine No. 1 and 2 N1 inputs are used to control the gain of the trim command signal. The two engine inputs are summed, divided by 2 to get the average and then summed with stabilizer position gain control. The gain, as a function of the N1 signal, will go from zero at 60% N1 to a gain of 100% at 80% N1 and above.
The stabilizer position input is used as the reference at the time that the speed trim system was engaged and also to provide position inputs as it changes as a result of the computed airspeed input. The stabilizer reference input also provides a gain control function which is combined with the N1 input to provide overall circuit gain.
737-300 AIRPLANES;
The Digital Air Data computer provides computed airspeed and altitude rate inputs. The computed airspeed is used to generate a stabilizer
command input as a function of airspeed. The altitude rate input provides an in-phase signal referenced to the CAS command to increase the
nose down command as the altitude rate increases.
ALL EXCEPT 737-300 AIRPLANES;
the digital air data computer provides computed airspeed. The computed airspeed is used to generate a stabilizer command input as a function of airspeed. Inertial vertical speed from the Inertial Reference Unit provides an in-phase signal referenced to the CAS command to increase the nose down command as the altitude rate increases.
The flaps input provides a reference angle of airflow program for any particular flap position which is then compared to the actual alpha angle
of the airplane. If the airplane approaches the stall condition, switch S1 will open, which will remove any more inputs from the speed trim
system until the condition has been corrected.
To enable speed trim system operation, the following conditions must all be met:
Autopilot not engaged
Flaps not up (For 737-300 airplanes only)
Flight Control computer, speed trim circuit valid
More than five seconds elapsed after cessation of manual trim
More than 10 seconds after unsquat.

Prior to speed trim operation, STAB trim, CAS and altitude rate (or inertial vertical speed on ALL EXCEPT 737-300 airplanes) are synchronized. The synchronized signal is used as a reference and any differance from that reference is the output signal. When the speed
system is operational and the airspeed increases, there would be a command generated as a result of the difference between the airspeed
reference and present airspeed. This is compared to the stabilizer position along with any assist from the altitude rate input. This signal
passes through relaxed S1 to the gain amplifier. The amplifier has a variable gain which is dependant on stabilizer position at the time the
speed trim system engaged (hold logic generated). The amplifier gain ranges from 100% at zero units of trim to zero gain at 5 units of
stabilizer trim.
The polarity detector determines whether the command is for nose up or nose down. The trim detector will not put out a command until the error signal is at least equal to 0.072 degrees of stabilizer command. This also provides the other required input to either gate 2 or 3 which
provides the nose up or down command. The output of the trim error detector remains until the error gets down to less than 0.070 degrees of
command error. The output then goes to zero and the stabilizer trim stops. Gates 2 and 3 also require that the speed trim system is in operation and that the local FCC has been selected for command of the system. The signal then passes to the nose up or nose down relay in the stabilizer trim servo.
Selection of FCC A or B is altered at each squat or selection of the operational system, if one system fails.

Why move stab and not adjust column feel to spd stability?
 

First let me thank downwind for the STS details that show that I had a mistake in my earlier description. I had stated that STS operation was not a function of thrust setting. This is not correct as noted in the details. I will correct my earlier entry to align. In response to Vessbot's comment about why not adjust column feel instead of move the stabilizer to provide missing stick force per knot I see two reasons.

First, if the unaugmented airplane were neutrally speed stable it would not require any steady elevator for a speed change and thus stiffening the column feel would not help as none would be needed to fly faster or slower. Further, if the unaugmented airplane were actually unstable with regard to speed it would require a push force to keep the nose from rising after slowing down and a pull force to keep the nose from falling after speeding up. These are clearly not the desired situation. Stiffening the column feel in that event would actually make the speed stability handling characteristics worse.

Second, if the column feel were stiffened, it would not only change the resulting stick force per knot (the measure of speed stability) but also the stick force per g (a measure of maneuver stability). The problem that STS was introduced to address is limited to speed stability and thus stick force per knot. There was neither need nor desire to change the stick force per g maneuver characteristics. The implementation of STS that moves the stabilizer in the direction to increase speed stability does not impact stick force per knot and thus leaves the maneuver stability characteristics essentially unchanged.