The Roller Coaster technique in Boeing 737's
 

Looking at my February 1, 1982 737-200 Pilot Training Manual, page 04.80.31 under the heading Recovery from Severe Out-of-Trim. It discusses a runaway and manual stabilizer operation.

Quote: In an extreme nose-up out-of-trim condition requiring almost full forward control column, decelerate, extend the flaps and/or reduce thrust to a minimum practical setting consistent with flight conditions until elevator control is established. Do not decrease airspeed below the minimum maneuvering speed for the flap configuration. A bank of 30 degrees or more will relieve some force on the control column. This, combined with flap extension and reduced speed, should permit manual trimming.

If other methods fail to relieve the elevator load and control column force, use the "roller coaster" technique. If nose-up trim is required, raise the nose well above the horizon with elevator control. ( my note: This may require both pilots to use their combined strength.) Then slowly relax the control column pressure and manually trim up. Allow the nose to drop while trimming. Repeat this sequence until the airplane is in trim.

If nose-down trim is required, slowing down and extending the flaps will account for a large degree of nose-up pitch. If this does not allow manual trimming then the reverse "roller coaster" can be performed to permit manual trimming.

In subsequent issues of the Boeing 737 FCTM these vital paragraphs were omitted. Consequently, we were never shown the "roller-coaster" technique in the simulator or even discussed at briefing. The result being a significant number of Boeing 737 flight crew nowadays are unaware of the roller-coaster method of relieving elevator forces in event of a runaway stabilizer trim. Nor were the pilots of the two 737 Max aircraft that crashed

I may be wrong, but I vaguely remember some discussion about this in the context of the Max crashes at the time.

The conclusion (I think) was that neither of those aircraft had sufficient altitude to recover using this technique, had the crews been aware of its existence.

With that and the rudder hardover issue, the chances of an average crew recovering from either of these conditions in IMC on sector 5/6 was pretty remote, so I can see why Boeing took it out. It’s test pilot stuff and a better way to counter it would have been a redesign and the best a totally new aircraft but that still hasn’t happened...

The roller coaster technique first came into being with the DC-8 and the Boeing 707 and was included in their training manuals. In the case of a runaway stabilizer downwards which the 737 MAX experienced, if the crews had switched off the stabiliser trim switches and switched off the AT's, then both pilots could have used their combined strength to haul the nose above the horizon to 20 degrees or more then let go of the control column, the aircraft would have started to lower the nose by itself in a slight negative G state. As that is happening the pilots would have been able to rapidly manually wind the stab trim from its out of trim state towards neutral

. During the period where the Boeing 737-100/200 PTM's (Pilot Training Manuals) were published, Boeing also published its Boeing 737 Flight Instructor Manual. Among other sequences, this manual published the technique to be used for a runaway stabiliser used in-flight - presumably assuming a simulator was unavailable. It stated that in manual level flight use the electrical trim to no more than two units either side of neutral, then switch off the stab trim cutout switches leaving the aircraft out of trim. Then by using the roller coaster technique, manually wind the stabiliser trim to an in-trim condition. There was a warning not to set more than plus or minus two units from neutral when setting up the aircraft in this fashion. As stated, this was an in-flight technique prior to having a simulator available

During my career of flying the Boeing 737-200, we were trained in the simulator the roller coaster method of countering a runaway stabiliser trim. It was considered no big deal and gave one the confidence to deal with what was a potentially serious non-normal if not handled promptly and correctly. There was nothing to be gained by Boeing in deleting this invaluable advice

It hasn't been completely removed, but it's been shortened and reworded.
https://cimg8.ibsrv.net/gimg/pprune....1166c07255.png

I have done the rollercoaster in the sim a couple of times. It's an interesting maneuver. Some points that I kept in the back of my mind:

* from a pilot perspective, you still have elevator, the aircraft is under control when trim runaways are dealt with without too much hesitation.
* from a pilot perspective, the rollercoaster is easy to do in a simulator environment but there is an absolute lack of physical senses (g-forces created in the pull out, or 0g at the top)
* from a cabin perspective, it seems an outrageous maneuver knowing cabin attendants might have been walking around pushing trolleys, especially knowing you are now pretty much letting go of the elevator and that might result in 0 or negative g. Meaning you are sending the trolleys and the cabin attendants, together with everybody that has no seatbelt on, into the air.

The maneuver is a theoretical solution to a "might come with broken bones and possibly necks as well" reality.
So if you want to focus on the rollercoaster in a training syllabus, I would at least double the amount of time working on the stab trim runaway scenario. Avoidance creates much bigger safety than having to correct afterwards.

I'd be inclined to agree; it's an ambulance at the bottom of the cliff reaction. Note also that:

• They removed the separate cutoff switches for manual and automatic electric trim (so pilots could not use the thumb switches to get the aircraft back in trim if the computers got it out of trim)
• They reduced the size and leverage of the trim wheel on the NG compared to the classic
• IIRC the simulators did not adequately simulate the forces on the trim wheel at high speeds and out of trim - it would have been essentially impossible to get back in trim even with a rollercoaster.

There's surely more critical, more general training that needs to be done.

But Why, Mr. Boeing (uh, McDonnell)? What did that improve or solve?

Boeing's response was it doesn't matter because the trim runaway procedure is turn off both switches together. But wasn't the reason for the original 2 switches that the pilots can at least try to use a simple motor to fix the trim without going straight to pointing the plane down to trim it up? When it could be just a command problem rather than a control problem?

Didn't one of the Max8 crash crews try turning the switches back on? And if they could have used just one without it force-pairing the functions, they would have saved the plane?

(somewhere said if they'd held the yoke trim switch all the way it would eventually have trimmed up even with MCAS and both cutoff switches on, but I assume not as quickly as earlier 737s?)

How I miss the simple time of travel by hot air balloons, you couldn't control the direction but at least the altitude... except when the vent stuck open... or something caught fire... birdstrike? Well... OK, I prefer a 737. But fix it!

The following extract is from the 737 pilot guide dated November 1975 FCT 37(IP)-1
Initial Conditions
Altitude: 5000 ft above terrain (minimum)
Airspeed: 210 to 280 knots
Configuration: clean
Stabilizer trim: Note.
Perform on ground first
Do not combine with jammed stabilizer
Do not use electric trim on ground to engage brake
Do not continue any runaway stabilizer trim action beyond 2 units from in trim.
Procedure
Cut out stabilizer trim cutout switches
Oppose trim with smooth control movement.
Retrim manually.
Out of trim Recovery Methods
Use manual trim
Accelerate or decelerate into trim
Use roller coaster (least desirable)
Bank airplane for nose up out of trim
Clean up
Stow stabilizer trim handles
Set stabilizer trim sitches to normal
DISCUSSION
Cutout switches to cutout even before brake engages
Use manual trim on ground to demonstrate brake.
Stabilizer trim wheel brake and release knob
If stabilizer brake or cutout switches fail to stop runaway, grab trim wheel and hold
Discuss manual trimming against electric motor
Smooth elevator; do not jerk
Show "apparent' runaway.
Stabilizer trim warning light, drive motor and clutches
Electric trim in opposition to runaway (not in training)
Isolate fault and use other systems
Manual trim wheels located 90 degrees apart (two-man effort
Trim speed: normal vs autopilot trim, flaps up and down.

It is interesting to note that Boeing state: Use roller coaster (least desirable). My comment: The whole procedure is complicated and fraught with the danger of getting out of sequence. All the more reason for a competent simulator instructor to lead students step by step through each procedure - including the roller coaster method to recover from a severe out of trim condition

I would generally argue the opposite - how well is it going to stick 5 years after training? Are people going to jump to the rollercoaster procedure even when it's not a sensible choice? Can a sim even effectively reproduce a rollercoaster?

The necessity for it could & should have been engineered out (and basically has been, it appears).

You have the same questions and answers previously discussed on other threads about why we don't simulator train pilots for no-engine landings and ditchings, except that those are significantly more common and significantly harder to prevent occurring in the first place.

Something else would have to come out of the sim to put rollercoaster in.

I would also argue that the sim time would have been better spent training people how to fly NPAs without hitting something, given the extremely remote possibility of a trim runaway going undetected until it hit the stops and one of the commonest causes of CFIT back in the day...

To demonstrate the roller coaster method of regaining control of runaway stabiliser in the simulator is a five minute exercise at the best. That five minutes might one day be the most valuable exercise on an emergency the average pilot might see in his career. Just one demonstration is all you need and I guarantee you will always remember it. It does not have to be practiced every year. Yet we spend 15 minutes wasting time on engine start, push back and taxy to the holding point without any qualms.

Two instances I can think of - one a 737 at Amsterdam and one a Thompson at Manchester. For both, the pilots were distracted with very low thrust set and the trim rolled nose up due to the dropping airspeed. Then a Go-around is initiated and the autopilot disconnected.

As the trim is more powerful than the stab, the aircraft pitches up, beyond the ability of the elevator to correct it - i.e., full foward on the controls and still pitching up. Now if the autopilot had been kept in, the trim would be running forward, but in manual flight the startle can mean that the pilots don't think to apply the forward trim needed.

In my experience, there are very few pilots who when faced with an ever increasing nose up change of trim, remember to roll the aircraft to the nearest horizon to get the nose to drop to regain lost speed.

I was shown this technique on my conversion to 737-300 in 1998. It did not strike me as a difficult or particularly hazardous procedure in itself and certainly wasn't described as reaching the point where cabin crews and trolleys get airborne. Indeeed it was a pretty measured affair, not overly complicated and not needing any unusual degree of co-ordination between pilots. As demonstrated it seemed to work very well indeed. I was surprised to discover that Boeing subsequently removed it from the FCTM as it seemed a useful fallback 'just in case'.

meleagertoo, I agree, I also was trained in the procedure when i converted to the 707 back in 1978. No big deal!

On the first-generation 737s, there were 2 trim motors. One was controlled by the trim switches and ran at a higher speed. The other was used by the autopilot and ran at a lower speed. Each motor was controlled via a pair of wires, one for nose up and one for nose down. On later models, the 2 motors were replaced by a single motor that could run at multiple speeds. The rest of the control system was left unchanged, so there were still high-speed inputs and low-speed inputs.

During MAX development, it was found that in order for MCAS to serve its purpose, it would have to be able to trim at a high speed. Therefore the Flight Control Computers were connected to both wire pairs. They would still use the lower speed for the autopilot, but MCAS would use the high speed inputs. If the switches were left unchanged, moving just the AUTOPILOT switch to cutout would've still allowed MCAS to operate. To allow the AUTOPILOT cutout switch to disable all automatic trim, like before, would've required an additional pair of wires. By that time, the procedure already specified that both switched should be used in tandem, not individually. So there was no benefit to adding more wires and instead, both switches were changed to have the same function (disabling all electric trim).

Autopilot issues were common in the 50s and 60s, so it was considered useful to be able to disable autopilot trim inputs while retaining pilot-controlled electric trim. In case of a runaway, the column-actuated trim cutout switches would stop the motor long before the aircraft could become severely out of trim. The rollercoaster technique wouldn't be needed. As far as I can tell, the only scenario in which the rollercoaster technique could've become necessary would be a trim brake failure. In that scenario, the motor wouldn't be operating but the stabilizer would move freely and would react to aerodynamic loads. The only way to regain control would be to grab the trim wheel(s), bring it to a stop, and then hold onto it for the remainder of the flight (or at least be ready to grab it again if needed). This is the reason for the final step on the Runaway Stabilizer checklist (Trim Wheel...Grab and hold).

MCAS was the only input that could bypass the column cutout switches, so it did create an additional risk. It was still assumed that pilots would use the trim switches to interrupt MCAS and keep the aircraft in trim. The Lion Air captain did exactly that. He flew successfully for about 5 minutes, interrupting each MCAS input within a second or two and not allowing the aircraft to get out of trim. He called for the Unreliable Airspeed memory items, but the FO didn't know them. The FO started flipping through the QRH, but paused several times respond to ATC calls. After 5 minutes of that, the captain handed over control to the FO, presumably so he could look for the checklist himself. The FO didn't interrupt the MCAS inputs. He let it trim nose down for the full 9-10 seconds each time, then he would make a single 1-second nose-up input... just enough to re-arm MCAS, but not nearly enough to get back into trim. He did express that he was having difficulty controlling the airplane, but the captain appeared to ignore him until it was too late to recover.

That was the second crash, Ethiopian Airlines. Yes, if they'd actually made significant nose-up inputs after moving the switches back to NORMAL, they would've regained control. Pilot input always overrode automatic inputs (including MCAS), and each pilot input would disable all automatic inputs for 5 seconds. I'm not sure what you mean by "not as quickly as earlier 737s". The trim would've run at the normal flaps-up speed. It would've taken around 20-30 seconds of continuous input to get back into trim.

The Ethiopian Airlines captain essentially repeated the actions of the Lion Air FO. Even after re-enabling electric trim, he made enough nose-up inputs to reactivate MCAS but not enough to get into trim.

One feature of both crashes was that none of the pilots seemed to note the trim position and how far it was from the normal location nor that the reading was changing. The first, safe flight, appears to have been regulated by the PIC feeling the excess load to hold the nose up rather than the reading of the indicator.

Perhaps this is a problem that the difficulty of using a dial to produce a value rather than the digital display of a number has moved from children being unable to read a clock with hands into the cockpit. Pilots know what V1 is and what V2 is and what flap settings to use, but somehow the most powerful control, trim, is not given the same concern? It must happen as it controls the relation between weight and flight speed?

Not enough people understand that the most powerful, and most deadly, control on most Boeings, is a little switch you move with your thumb, as a secondary task.

The absence of that switch, is one of the most deadly things about several other designs.

I remember when I flew B707s way back when that if the stab ran away that full up elevator control was equivalent to four units of stab trim so if you didn’t catch the runaway before that then you were left with little or no pitch control and the only option was the roller coaster procedure that we practiced in the sim.

Cheers,
BH.

"ignorantAndroid" thank you for that information

Which brings on another topic of another recovery procedure that is completely forgotten and where things might get really confusing if you start with these recovery procedures.

If the THS is "free-floating" in the wind, you can "control" the trim through the elevator, but the weird thing is that the actions are completely opposite to the "normal". If I recall correctly, if the trimwheel is continuously trimming nose down, you have to push the elevator full down and it will stop the runaway. You don't use the elevator to "control the nose movement", but you use the elevator to "control the THS movement". And in that case, the elevator actions are completely opposite to the rollercoaster. You create a force that moves the THS in the direction you want it to move and you don't even need to touch the trimwheel to retrim. It will rotate by itself, you just have to know how to steer the elevator. If others have more information, feel free, as this is really long lost in my memory... I'm not aware of any sim that is capable of simulating this recovery procedure.

In the end, all these procedures are nice to know if you know what you are doing, but can lead to gigantic disasters if you don't really know what you are doing. Just stop the runaway in time and continue with what you have as you should have plenty of controllability. You shouldn't be doing testpilot recovery procedures.

It certainly does seem to be lost in your memory. I've already explained we trained in the rollercoaster manoeuvre in (a Gatwick area) sim in the late '90s. It didn't seem unrealistic at all, far from it. There was no mention of THS being "free floating", this is a technique to recover from a runaway trim that's not been reacted to fast or decisively enough (as, apparently, in both MAX accidents) and enables recovery of ability to re-trim after the force exerted by the control column to maintain level flight has rendered the trim wheel so stiff as to be immovable.
My immesiate reactions on learning details the MAX accidents were 1) why the digamma didn't they react to what was obviously a runaway trim by operating the cutout switches and 2) if control was hard to maintain by control column alone whey didn't they roller-coaster - not realising that this (far fron "test pilot" manoeuvre) had been removed from the book and maybe they didn't know it. I was astounded to learn thst it was no longer included - along with a whole bunch of other potentially useful but rather esoteric stuff like variations on battery starts and in-flight FMC resets that we had in our FCTM and did in the typre conversion course, admittedly with a very old-school instructor who thought this stuff important.

If it's free floating as in if the trim brake failed, I think was what he meant; someone had mentioned the brake.

Boeing shackled to the past - man or machine
 

RCTechnique - an old, ancient view of safety certification which claimed a piloting contribution in meeting requirements.
Long past its best-before-date, but with Boeing's policy of Grandfather Rights in the 737, and their design philosophy it was been retained.
Hindsight reminds us that human involvement cannot be quantified or prejudged.

Pre 1980s

https://www.dropbox.com/scl/fi/7wivz...=z97cebal&dl=0

"Boeing shackled to the past - man or machine."
Others consider the future - man and machine, together.
But who will embrace the next change; not a step jump, but a slow evolution so as not to disturb the current levels of safety.

Thanks, yes. If it's a "trim system failure", the yoke cut-out switches stop the runaway. If the brakes on the jackscrew fail, the yoke cut-out doesn't work and you need to grab the trimwheel and stop it immediately. In the first case, if manual trim is required but unable, the rollercoaster is an option. However, doing a rollercoaster with a free floating THS seems complicated to do from a CRM perspective. The easier solution shown to me was to "use the elevator to control the THS" to get in trim which requires opposite elevator movement to the rollercoaster.