Flightglobal article from 2010:

"Rossiya says incidence of technical failure in its fleet occurred once every 344h in the An-148, every 5,355h in its Airbuses, and 2,824h in its Boeing 737s."

From (remove 2 spaces) www.flightglobal. com/ news/articles/rossiya-slams-an-148-reliability-347983/

Also interesting to see the ~50% difference between Airbus and Boeing, could it be difference in fleet type/age or is this a general picture?

Many thanks for the usual excellent information, this clears up a lot of initial speculation. That was some ground impact to totally destroy an engine like that and propel the part that far...

It does now appear that there was no in-flight breakup, the plane came down in one piece, in a very high energy state with at least one engine on full power.

CFIT? It certainly supports suggestions of an eroneous IAS - Pitot?

https://cdn.aviaforum.ru/images/2018...511f97952.jpeg

https://cdn.aviaforum.ru/images/2018...d8c1f852a.jpeg

AndrasZ
I don't see how you can exclude the same linear traces of debris across a field with some random wide scattering from a possibility of an inflight breakup.

I have yet to see a mangled turbine shaft

many possibilities are still open

You actually commented on it, saying you only saw such thing once... :)

In an inflight breakup one would typically expect to find pieces well before the main wreckage (hence my questions to locaton). In this case it seems to be confirmed that all the pieces we saw initially are within the debris fan ejected from the main impact point.

I've been following the Russian PPRuNe equivalent site since the first reports,

They reached page 40 while we still hadn't hit page 3 here and they are just about to reach page 100.

A vast amount of dross to filter but some pertinent points to pass on.

The AN 148 aircraft were from the Russian production line rather than the Ukraine.

Rossiya bought them but were unhappy with maintenance and reliability.

They later dry leased 4 airframes to Saratov.

The helicopter mid air claims were generated by a single news source and immediately debunked by the Russian Post Office.

Hours of in flight breakup conjecture for the usual host of potential reasons crushed by the cctv video. The aircraft rode its wings and tanks all the way to the ground and it had a tail to enable impact at such a low angle.

FR24 and other software fans claimed immediately that the entire 148 fleet never produced ADSB speed data and ground interpolation of position and time stamps is the only way speed could be displayed. For various reasons Saratov were restricted to internal flights plus a small handful of former Soviet states. Thus the lack of knowledge outside Russia regarding data integrity and absence of any usable granularity.

That final point above is the most important of all for the FR24 junkies out there already building their maps and theories.

Rob

Enlarging a frame from the video just after impact, it appears that the burning fuel plume was associated with the engine launched forward from the main impact point.

It might appear like that, but there is a building obscuring part of the view. I don't think you can draw detailed analysis from that video.

If my understanding of the russian PPRuNe equivalent is correct the Voronezh (Russian) built An-148 have issues with the speed data in ADS-B, that's why FR24 interpolates the speed between datapoints - which naturely provides unreliable data (especially in a turn). If I understood correctly the Ukrainian built An-148's don't exhibit this behaviour.

I think it is trees that block the view ob the lower part of the fireball. But the general statement is true: one can only see a small (upper) part of the fireball.
What else can be drawn from the way the fireball and the parts spread is that the angle of arrival was rather shallow, maybe somewhere around 30° and the forward velocity was very high.
So no lumbering in a stall and no lost tail feathers either.
Looks generally like a loss of control at altitude for whatever reason followed by an (aerodynamically) controlled (but from a piloting perspective apparently uncontrolled) dive with insufficient altitude to recover.

I've just deleted a post from a data driven FR24 junkie so lets do this again.

The Antonovs from the Russia production line at Voronezh have never transmitted speed whatever the requirement and specification docs say.

All the Ukrainian produced airframes do work.

The SSJ had the same problems to cause embarrassment on its world tour until it was 'modified.' I can't say for sure but I suspect Western equipment had to be bought and integrated as overseas sales were inherent to the SSJ's reasons to exist.

Further, I have an Excel spreadsheet of the raw ADSB data. Three columns have zero for every sample entry for three fields.

Altitude
Speed
Vertical Speed

So, a big air data black hole.

Irrespective of the flight following software you favour it is making a huge proportion of what you see up. Good luck working out what each version of software used as its algorithms and data crunching to guess the data you are looking at.

Rob

What was the source of your spreadsheet data?

Be vague if you wish...

FR24 published their ADS-B data in csv format.
Airspeed isn't there, but GPS lat/long is, and that's what I used to get a ground speed of 300 kts until nose down. Even that data is weird, though,since not only does the same data repeat, sometimes it throws other numbers in there.
So I found that I could get acceptable data by taking the values that are repeated and using the time stamp of the first time that value was reported. After all, we know where the plane ended, and the points selected draw a line there over the time elapsed.

But if you want to do fine -grained analysis, the data won't work. Something about the laws of physics.

Flight profile from cleaned-up FR24 data:

http://www.avgen.co.uk/SOV703.jpg

Last 60 seconds of recorded data (ending at 2475' AMSL, give or take a QNH adjustment) suggests a ROD of around 3800 fpm and a GS of about 315 kts.

Ground Speed from GPS Data
 

The Haversine formula will give distance between two nearby lat/long points. Dividing by the interval yields groundspeed in level flight and is close enough in normal climbs and descents.

In the case of steep climbs and descents you also need to include altitude change.

GPS accuracy can be affected by extreme attitudes as satellites come into and out of view.

New version of the spread sheet with more data.

No Altitude until the alt kicks in passing 425' on departure at 11:21z and then the data shows a steady climb.

A series of climb data points indicating 1200 - 1500 feet until a maximum of 1900 at 11:22z. The transponder remained squawking 1571 and didn't ever change through the whole flight. There were several data missed captures with no squawk or time code but altitude was read

Climb continued but with great amounts for time hack discards and replaced by SOV073. There are fairly regular misreads from single data points usually indicating spurious altitude in the teens and as erroneous as FL290. These were individual points

By 11:25z it was operating or or near 6000' but climb continued.

11:26z 6700' drifting back down to 6000' but continuing descent.

VSI data show the usual odd, individual rogue numbers slipping in as I'd expect but the climb 6000 to 6700 has rates suggesting it had been cleared higher. However the descent rate becomes pronounced but leaping from sample to sample from positive to negative numbers in the course of a single second of data.

The attempt to level again at 6000 dips below and here the time hacks are more regular suggesting there's a fair bit of heave ho going on to reach 6000 from 5800' and continues up to 6275'.

The aircraft remains around that altitude for about 5 seconds but the pitch to regain 6000' and remain there is very brief and returns to descent and remains that way until the lowest recorded 2475' and beyond. Below 4900 feet there are no time stamps at all but the claimed vsi data is very extreme.

I've taken the time to describe this in words so that pilots can analyse what was happening and run it through the things they've experienced in the sim over the years.

The pretty illustrations while dramatic are compressed and don't show the data we see in the time periods we have to analyse and react.

Rob