Pilot Report: My 737 Max Simulator Ride | Hacker News Reader

Pilot Report: My 737 Max Simulator Ride

avgeekery.com

15 pointsThinkingGuy25 comments

> Eventually, the trim wheel made an abrupt twitch forward at the high rate, but only for a fraction of a turn. This, our instructor told us, was the MCAS system becoming active or “waking up”. What was happening behind the scene was the MCAS logic took a “snapshot” of the existing trim position when its threshold AOA was reached. It then calculated a maximum amount of trim that could be added. Should the trim ever meet this computed limit, the speed trim and MCAS system becomes inhibited for the remainder of the flight.

This seems like one of those overly complex behaviors that's indistinguishable from "randomly decides to just stop working" when someone doesn't fully understand (or remember) all the details. Is this common in aircraft, especially for critical flight systems? Does a normal fly-by-wire system do things like this?

the key is that it has to hit the limit for the maximum amount of trim that should ever reasonably need to be added before it is disabled for the remainder of the flight. If it hit the limit it is an indication to the control system that something isn't operating correctly and it should stop automatically operating the outputs (trim) based on the inputs (AOA) as it isn't having the anticipated effect and it should punt the problem to the human pilots.

Presumably there is some annunciation for both the function activating and also being disabled.

Hey, can I ask you to reword the quotation? English isn't my first language and I have trouble understanding it the way it's written. (I know all the aviation lingo and stuff, it just doesn't make sense.)

   All of the training
   events were flown
   from SeaTac airport
   in VFR conditions

What the? Why‽ It is easy to recover from any upset in VFR when you see the horizon. Doing it with no reference is much harder. And often causes crashes, eg: https://en.wikipedia.org/wiki/Flash_Airlines_Flight_604

I guess that's not what they were training for this time. I'm sure the pilot's done hours of IFR training at other times.

Recovering sudden upsets in IMC is VERY different than doing it in VMC. Doubly so if you're not sure which of your instruments you can trust.

Yes, but the purpose of this was not to train how to recover stalls. It was to show to the pilot how the aircraft reacts when flown way out of the envelope and which things happen in which order.

I guess the training was actually positively intended to be flown VFR, to give the pilot the least things to worry about and to allow them to focus on what exactly is happening.

Sudden nose down (as MCAS does) is an upset and thus would probably be best practiced in worst case. Not best. When I go practice emergencies with my CFI (every 3 months) I ask him to make it as hard as possible, not as easy.

Boeing and the FAA did a bang up job hiding the details of their unsafe system behind vague language such as this:

> Specifically, just before stall entry, and well below any normally encountered airspeed, control column pitch forces became lighter when they are required by certification rules to become heavier.

"[not] normally encountered" could mean, once every 10 flights. Maybe once every 1000 flights. 1 out of 10 scenarios for a given flight. Who knows? Certainly not the FAA, because they didn't collect or publish this data in their report. They used similarly vague language most likely provided by Boeing.

This report walks through the stall scenario with MCAS working, and it walks through an entirely different scenario with MCAS malfunctioning. It does not walk through the stall scenario with MCAS malfunctioning. Boeing and the FAA still have not released data on how often MCAS was operating (erroneously or otherwise) prior to grounding, and how safe it might be for pilots to operator aircraft with MCAS disabled.

It sounds like the pilot here is on the take. Such a glaring omission in the simulator is damning in my eyes, and I trust Boeing and the FAA less than I did after the crashes.

>"[not] normally encountered" could mean, once every 10 flights. Maybe once every 1000 flights.

It does not mean either of these. It means "unlikely ever to be encountered by an airline pilot during their career except in simulation".

MCAS isn't an anti stall system. It's a regulatory compliance system. Specifically, it's to comply with a regulation requiring that the force on the stick increase with increasing angle of attack.

The situation that the regulation addresses should never occur on a fare carrying flight. It's at the edge of the flight envelope, and a pilot who deliberately flies at the edge of the envelope with fares in the back should be prosecuted.

That said, sometimes things break, or weather causes things to go wrong. This may result in inadvertently flying near the edge of the envelope.

To a rough approximation, 0 out of a 1000 flights will meet the requirements for MCAS to activate.

You don't know that, and neither does the FAA. You only know what Boeing has said and the FAA neglected to verify even this basic fact.

We do know exactly the conditions required for MCAS activation. And we do know those conditions should never be encountered during a fare carrying flight. Any pilot who deliberately puts their aircraft into conditions that result in MCAS activation should be prosecuted.

Pilots should never come close to stalling during fare carrying flight. Therefore the MCAS should never activate during routine flight.

We do not know the conditions, we know the conditions Boeing said, and we know that Boeing said it's rare. The data from flights before grounding could answer this question, but it's nowhere to be found in the report. It's not even discussed.

The thing that makes me uncomfortable when reading all the details is the fact that Boeing is using trim to move an aerodynamic surface that is not intended to be used for direct and immediate control of the aircraft, to induce an aerodynamic force on the actual control surface (the elevator) which then manifests itself as a force on the control column.

It's basically all backwards. It's like implementing lane assist in cars by rotating the whole transaxle so the driver gets phantom forces in the wheel that keep the car centered, instead of sending some signals to the power steering unit that's already there and already messes with the control forces. (Or going all fuck-it like Airbus and doing drive-by-wire with an Xbox-sourced steering wheel.)

Can I ask if you actually fly planes? Moving the trim is exactly how most autopilots work. I usually fly the 172 and pitch changes made via the autopilot (I am usually flying with a GFC700) are done 100% via actuation of the trim wheel -- and the trim wheel is also the main control for adjusting control forces to maintain a given pitch attitude.

Edit: Reading your comment more closely, I think your concern is more to do with the fact that the trim is actuating the whole horizontal stabilizer rather than just the elevator? There is no separate elevator trim on the 737.

Currently saving up for getting my license. :)

I understand why it's easier to let the autopilot control the trim instead of messing with the primary controls.

What feels weird is the fact that it's using this whole monster of a mechanism in order to provide the pilot with an artifical stick force.

I get it, the control column is supposed to require more pull force the closer to Vs you get. It's not what real planes do, but it's a safety mechanism, alright. But instead of using a simple servo that would actuate the control column and wrestle with the pilot a little, what they're doing is they're actually trimming the plane nose-heavy so that actual real forces wrestle with the pilot.

Instead of a servo pulling on the control column slightly to enhance the natural pitch-up moment of an overspeeding aircraft they trim the plane nose-up and again, let the pilot push against an aerodynamic-yet-artificial force.

The problem is that the 737 was designed in the 60s. The point where all the automation is designed to get hooked in is the trim.

If it were designed today, the 737 would have fly by wire, and all the stick forces would be generated by a servo and the computer anyway.

tl;dr

He thinks the updated aircraft is great.

But he also thought the old aircraft when great in 2018, in the configuration that crashed.

"...I thought it was a great flying machine back then and I think it will be better than ever after its return. Was the aircraft as well designed as it could have been? Perhaps not, but then in no human endeavor is perfection ever achieved."

That's the single least useful thing you could've distilled from the article.

OK then, What's your "better" tldr?

There is no tldr. It's full of information about avionics and control systems. There are descriptions of flight sim procedures. It's an interesting article to be consumed whole.

If you're interested to read a primer on avionics and control systems, yes.

If you just want to know whether it's safe to get on a 737 Max, no. For that, you'd need a white box analysis of the software, access to the change log, and, ultimately, millions of miles in the air waiting to see what happens.

He thinks MCAS is fixed "the odds of a runaway trim scenario are now effectively nil." And stuff.

Also i think pilots are more or less assigned same aircraft models so i don't think you can really trust them saying their assigned model is great since they have conflict of interest.

Reply on news.ycombinator.com