Avherald mantains it didn't happen as reported:

Bizarrely, the argument that it didn't happen is justified by citing the FlightAware ADS-B log, despite the fact that it contains no data points at all during the 25 seconds of the roller-coaster.

Incident: United B772 at Kahului on Dec 18th 2022, pilots filed safety report

Saw that - but Simon's position is inconsistent with the little bit we do actually (think we) know - including the first hand statements from a passenger about what he (they) sensed on board. The reality is, and facts are, that something happened shortly after takeoff - sufficient that the pilots reported it and were given extra "training". The passengers certainly noticed something was amiss too. Maybe it's not 2.7G - but something happened.

Original FR24 plot (green) plus all of FlightAware's data points (blue), the latter cited by Avherald as "evidence" it was a non-event:

https://cimg9.ibsrv.net/gimg/pprune....d626454df7.jpg

That will save the NTSB a lot of work. :ugh:

to consider:

pax with piloting background saying it did not feel like a nose-dive or push-over is actually consistent with a banked loss of altitude on a steady positive g. One of the reasons it might go unnoticed by the pilots in the first place. Someone sitting in another location of the humongous cabin might get a markedly different perception.

Noted, the tracked open data do not seem to show the expected heading divergence.

Ceterum censeo CVR...

A microburst can certainly upset the flight path without warning, and there is a tendency to blame the crew for mishandling until the data is analyzed - been there.

There are a plethora of data sources flying about. What will count is the fdr or QAR, especially since they offer sub second resolution.

Assuming they kept those records...

I guess I am thinking this really happened. It sounds like the captain coming on the PA and saying "you might have felt a couple of Gs" is indication that there was indeed a steep descent and immediate ascent. I also suspect that he didn't pick that number randomly and likely lowballed it (natural human tendency in a situation like this). The fact that this statement was made along with the FR24 data, and the pilots receiving retraining makes me fairly certain that what has been floated (incorrect flap retraction / loss of lift / recovery) is the most likely scenario. I am curious whether UA preserved the FDR data and/or CVR data.

You can expect the CVR and FDR being unobtainable for all the usual reasons. Some of those are not wrong.

QAR is a different beast, though the data might not be allowed to leave a certain office or even a computer station. That information, alongside honest best crew recollection, will uncover the dynamics in very precise detail.


​​​​​​

Passenger Roid Williams recounts the episode to the Washington Post:

Boarding and taxiing went smoothly, he said. The pilots squared the Boeing 777 with the runway, hit the throttle, accelerated down the airstrip and took off. For about a minute, they climbed normally.

Then, the nose pitched up and the plane started ascending at “an alarming angle,” Williams wrote in a Facebook post recounting the incident.

“That’s when you started to hear folks let out a few screams,” he told The Post.

Williams, who studied aviation, said he feared the plane might stall. After climbing steeply for three to five seconds, the plane went into a sudden nosedive.
. . .
After diving for nearly 20 seconds, the plane pulled up “at a very strong rate.” Williams clenched his teeth, squeezed his armrests, pushed himself into the back of his seat and kept muttering prayers.

Accounts of an alarming pitch attitude should be treated with a certain amount of caution, given that the weather would have meant that no horizon was visible. There doesn't appear to have been any sudden gain in height before the 777 headed earthwards.

I was thinking the same - the QAR can be downloaded in minutes, so if the crew reported anything after landing the QAR could have been readily downloaded during the 2.5-hour turnaround.

Of it wasn't climbing any faster but pitching up higher?

Isn’t the QAR transmitted automatically during taxi in anyway? I know all the data for flight data monitoring is at my companies headquarter usually when the parking checklist is done. And that is without SATCOM, once that is installed it will be transmitted throughout the flight in regular intervals.

I'm having trouble reconciling the increasingly high speed at the bottom of the descent. If it encountered decreasing-performance windshear, I'd expect the speed to stay low during the descent with the nose level or maybe slightly low (close to the sticker shaker at full power to minimise altitude-loss). Even if it then encountered very strong increasing-performance windshear, the speed would start increasing, but should have been controlled enough not to get so fast while still descending. The nose would still have been somewhere near the horizon and as soon as the headwind came back on, the windshear escape manoeuvre would have the nose way up.

This appears to have just nosed over, with power still on, and rapidly accelerated with a big snatch-pull to recover, as if someone has just shoved the stick forward, held it, then pulled up.

If this was caused by a microburst, unless a dry microburst, IMO the convective cloud creating it would have been obvious on the aircraft weather radar.

This is from the FR24 data CSV (Date-Time, altitude, groundspeed):

https://cimg6.ibsrv.net/gimg/pprune....c1778289db.jpg

Surely the QAR would have been downloaded per the FDM programme, and this event detected anyway?

It depends. The hardware varies - some can download automatically via Wi-Fi, some via satellite, and some need a hardwired connection. Satellite is fairly unusual - the QAR records a massive amount of data, so it's rather expensive to transmit all that via satellite (some are programed to download specific unusual events - or small data subsets - to keep the satellite fees down). Others respond to and download when queried from base. When I was still working the 747-8, GE had installed an "EMU" - Engine Monitor Unit - that could automatically download via Wi-Fi at the terminal (assuming the terminal was properly equipped), or send out specific event data sets via satellite. The operators started complaining that the EMU was using way too much satellite bandwidth and it was costing them too much money for routine data reports, so GE revised the EMU software to send fewer specific event data sets.
All that being said, I have no idea what UAL has on their 777s - however the point is that a QAR download would be quick and easy regardless of the download method used. It also would likely contain far more data - in greater detail - than the FDR.

The radar imagery around dive time
 

The area was busy at the time, but not exactly at the plane's position, this is the radar image Reflectivity at 00:54 UTC, other time steps via menu

These are the Doppler speeds at the same time

If I understood the position right, the plane would have been outside of any severe weather. Please correct me if I got the position wrong.

Local time is UTC-10. So 0:50 UTC should be 2:50 PM local. (Edit: was confused with indicated time on the linked page, may depend on location when opening the page)

So these UTC times are wrong? FR24 times

Time of the upset in FR24 is 00:51 UTC (think 24:51 - 10 h = 14:51 = 2:51PM)

Good. Then the radar images I posted are correct. I set UTC for the page. Maybe you have another time zone set on weather.us?