Author Topic: FAILURE : Astra R3.3 - LV0010 - TROPICS Flight 1 - CCSFS SLC-46 - 12 June 2022  (Read 38497 times)

Offline kdhilliard

  • Full Member
  • ****
  • Posts: 1082
  • Kirk
  • Tanstaa, FL
  • Liked: 1572
  • Likes Given: 4080
A few more screen captures:
T+07:14 -- Flight Director Chris Hofmann abruptly removes his mask, presumably in frustration at the premature shutdown.
T+07:21 -- Shutdown visible on our feed.
T+07:22 -- Tumbling.

Link to T+07:10.

Our view of the Velocity/Altitude telemetry seems to be synchronized to that of the downlinked video, as velocity peaked at the same time we saw shutdown: 6575 m/s at 531 km.

Offline Celeste_El

  • Member
  • Posts: 28
  • MA
  • Liked: 53
  • Likes Given: 1
Also, that last screenshot shows an altitude of 359km…wasn’t the planned altitude 350km ?
Perhaps the first stage guidance was lofting the rocket and the 2nd stage ran out of fuel trying to obtain orbital velocity?


Planned was 550 Km. They probably over shoot and null out vertical velocity as they approach target apogee. I think Manley's take on the matter is the very plausible. Rocket noses down to kill vertical velocity. This causes enough slosh to cause the engine to shut off. Possibly the nose down was too violent, would be inline with the poorly tuned control seen with the wiggling on ascent. Then, without engine power, the nose down becomes a tumble.

Either way, Astra is good at trouble shooting and getting back to the pad. Last failure was met with a successful launch within a month.

https://twitter.com/DJSnM/status/1536048958429990914
« Last Edit: 06/12/2022 07:11 pm by Celeste_El »

Offline DaveS

  • Shuttle program observer
  • Senior Member
  • *****
  • Posts: 8526
  • Sweden
  • Liked: 1199
  • Likes Given: 65
Will interesting to know if any Low Level Cut-Off (LLCO) sensors were tripped. They're there for this very reason, to shut down the engine(s) before a complete loss of propellant occurs.

LLCO sensors on the shuttle was known by the faulty term of "Engine Cut-Off" (ECO) sensors and malfunctioning sensors caused by a badly designed external feedthrough connector on the External Tank side of things were the causes of launch scrubs of STS-114 (July 13 2005), STS-115 (September 8 2006) and STS-122 (December 7 and December 9 2007).
« Last Edit: 06/12/2022 07:20 pm by DaveS »
"For Sardines, space is no problem!"
-1996 Astronaut class slogan

"We're rolling in the wrong direction but for the right reasons"
-USA engineer about the rollback of Discovery prior to the STS-114 Return To Flight mission

Offline kdhilliard

  • Full Member
  • ****
  • Posts: 1082
  • Kirk
  • Tanstaa, FL
  • Liked: 1572
  • Likes Given: 4080
Tracking map, yellow vertical line is 1st stage MECO point with the green vertical line being the planned SECO and orbital insertion point:
MECO was back much closer to Florida. I think the yellow line was the planned Stage 2 cutoff and the green line was planned payload deploy.
That doesn't make any sense unless the vehicle was seriously underperforming and flew long as the tracking line was well beyond the yellow line before the early SECO.
...

No.  That tracking line wasn't shown past the yellow line until, T+09:06, nearly two minutes after the early shutdown.

(Complicating things, that tracking line didn't appear to be updated regularly, as seen by its significant jump at T+05:47.  Link to T+05:45.)

Also, the timing of the shutdown, 75 seconds or so before the planned SECO of T+08:30, doesn't suggest a longer flying underperformance.

Offline kdhilliard

  • Full Member
  • ****
  • Posts: 1082
  • Kirk
  • Tanstaa, FL
  • Liked: 1572
  • Likes Given: 4080
...
So what does this mean?  Reentry somewhere?

My guess?  Reentry before the coast of Africa.  (A 1 km/s deficit is a lot.)
Anyone care to do the math?
6575 m/s at 531 km for T+07:21, with that speed including Manley's estimated 700 m/s of loft.

We get to see another minute of telemetry (contradicting what I said earlier), with final values of 6518 m/s at 570 km for T+08:20, so a more precise calculation of angle could be made.

Just remember that with a significant portion (~ 85%) of orbital velocity achieved, it is no longer a "flat Earth" calculation.

Offline edkyle99

  • Expert
  • Senior Member
  • *****
  • Posts: 15391
    • Space Launch Report
  • Liked: 8566
  • Likes Given: 1356
That's five failures and two successes over the course of Rocket 3's flight history.
Two failures during first stage flight, two during second stage flight, and one due to fairing separation failure.  A little bit of everything.  Frustrating to have failures after successes, but note that the two most recent failures were both out of the Cape while the last two successes, including a success between the two failures, were both out of Kodiak.

Something like -1950 x 570 km, so very suborbital.  Probably Atlantic reentry.

 - Ed Kyle
« Last Edit: 06/12/2022 10:50 pm by edkyle99 »

Offline niwax

  • Full Member
  • ****
  • Posts: 1422
  • Germany
    • SpaceX Booster List
  • Liked: 2040
  • Likes Given: 166
...
So what does this mean?  Reentry somewhere?

My guess?  Reentry before the coast of Africa.  (A 1 km/s deficit is a lot.)
Anyone care to do the math?
6575 m/s at 531 km for T+07:21, with that speed including Manley's estimated 700 m/s of loft.

We get to see another minute of telemetry (contradicting what I said earlier), with final values of 6518 m/s at 570 km for T+08:20, so a more precise calculation of angle could be made.

Just remember that with a significant portion (~ 85%) of orbital velocity achieved, it is no longer a "flat Earth" calculation.

https://twitter.com/planet4589/status/1536069497395810307
Quote
Slight revision: I estimate reentry about 400 km west of Dakar. I am now a bit more confident that the debris wouldn't have got quite as far as the African coast.
Which booster has the most soot? SpaceX booster launch history! (discussion)

Offline Michel Van

  • Full Member
  • ***
  • Posts: 310
  • Liege, Belgium
  • Liked: 148
  • Likes Given: 170
So what went wrong ?
some speculation

it look almost redo of flight Rocket 3.2 were was issue with propellant mixture ratio
also issue with Second stage software flaw in the thrust vector system on flight LV0008

but this here look different
the engine stop combustion then spit gas out  first orange then white

Possible reasons:
a Combustion instability either slouch propellant or issue feed in propellants
or issue with valve in Helium pressurisation system or valve in propellants feedline to Engine.

some people in Twitter poiting out that stage was not flying stable and was wobbling.
if that was reason, i don't know the Video and Data feed was "Sluggish"
i hope Astra bring soon new data on this

Astra has for moment success rate of 22% !
and lost in total 6 satellite of paying customer,
they have to fix that issue fast or is Game Over !

Offline Star One

  • Senior Member
  • *****
  • Posts: 13999
  • UK
  • Liked: 3974
  • Likes Given: 220
Scott Manley’s take:


Offline OneSpeed

  • Full Member
  • ****
  • Posts: 1588
  • Liked: 4930
  • Likes Given: 2078
Here is a plot of the webcast telemetry from Astra LV0010.

Astra do not appear to use a throttle down for MaxQ, but there is an unusual artifact at around T+104s, where the acceleration appears to increase dramatically for about 6 seconds. From the telemetry, first motion is not until T+11s, so perhaps they are allowing the telemetry to catch up by a few seconds for separation?

The quality of the Astra telemetry is improving all the time, to the point that I can make a rough estimate of the Velocity Y component. From the plot, at SECO, it was around 750m/s.

Offline FutureSpaceTourist

  • Global Moderator
  • Senior Member
  • *****
  • Posts: 48174
  • UK
    • Plan 28
  • Liked: 81684
  • Likes Given: 36941
https://twitter.com/jeff_foust/status/1536793959434813440

Quote
Zurbuchen, on Astra/TROPICS launch failure: after it happened, wondered if we should have done something different; concluded absolutely not. Mission costs $30M, three launches $9M, to get a new capability into the field. #AAS240

Offline king1999

  • Full Member
  • ****
  • Posts: 440
  • F-Niner Fan
  • Atlanta, GA
  • Liked: 306
  • Likes Given: 1280
Zurbuchen, on Astra/TROPICS launch failure: after it happened, wondered if we should have done something different; concluded absolutely not. Mission costs $30M, three launches $9M, to get a new capability into the field. #AAS240
I understand they had to say that in public. But that's not a good attitude to have in engineering. You can ALWAYS have done something better, be it more analysis, more testing or better modeling etc. If it was a problem you didn't anticipate, you would need to have a better fault tree.

Offline Redclaws

  • Full Member
  • ****
  • Posts: 750
  • Liked: 860
  • Likes Given: 1047
Zurbuchen, on Astra/TROPICS launch failure: after it happened, wondered if we should have done something different; concluded absolutely not. Mission costs $30M, three launches $9M, to get a new capability into the field. #AAS240
I understand they had to say that in public. But that's not a good attitude to have in engineering. You can ALWAYS have done something better, be it more analysis, more testing or better modeling etc. If it was a problem you didn't anticipate, you would need to have a better fault tree.

Zurbuchen is the NASA administrator in charge of the acquisition, so not part of the engineering team.  He’s saying he’s happy to have bought launches in this risky way.

Offline TrevorMonty


Zurbuchen, on Astra/TROPICS launch failure: after it happened, wondered if we should have done something different; concluded absolutely not. Mission costs $30M, three launches $9M, to get a new capability into the field. #AAS240
I understand they had to say that in public. But that's not a good attitude to have in engineering. You can ALWAYS have done something better, be it more analysis, more testing or better modeling etc. If it was a problem you didn't anticipate, you would need to have a better fault tree.

Zurbuchen is the NASA administrator in charge of the acquisition, so not part of the engineering team.  He’s saying he’s happy to have bought launches in this risky way.

TROPICS was good choice for risky launch. Being small constellation easy enough to build couple replacements.

Best save more expensive one off satellites for more reliable LV.
« Last Edit: 06/16/2022 12:18 am by zubenelgenubi »

Offline Steven Pietrobon

  • Member
  • Senior Member
  • *****
  • Posts: 39218
  • Adelaide, Australia
    • Steven Pietrobon's Space Archive
  • Liked: 32738
  • Likes Given: 8196
Astra do not appear to use a throttle down for MaxQ, but there is an unusual artifact at around T+104s, where the acceleration appears to increase dramatically for about 6 seconds. From the telemetry, first motion is not until T+11s, so perhaps they are allowing the telemetry to catch up by a few seconds for separation?

Maybe its a change from relative velocity to inertial velocity.
Akin's Laws of Spacecraft Design #1:  Engineering is done with numbers.  Analysis without numbers is only an opinion.

Offline OneSpeed

  • Full Member
  • ****
  • Posts: 1588
  • Liked: 4930
  • Likes Given: 2078
Maybe its a change from relative velocity to inertial velocity.

Perhaps, but why is the elapsed time from first motion to MECO about 6 seconds longer on the video feed than for the telemetry?

Video Elapsed      02.56
Telemetry Elapsed   02.50

Offline Jim

  • Night Gator
  • Senior Member
  • *****
  • Posts: 37441
  • Cape Canaveral Spaceport
  • Liked: 21451
  • Likes Given: 428
Zurbuchen, on Astra/TROPICS launch failure: after it happened, wondered if we should have done something different; concluded absolutely not. Mission costs $30M, three launches $9M, to get a new capability into the field. #AAS240
I understand they had to say that in public. But that's not a good attitude to have in engineering. You can ALWAYS have done something better, be it more analysis, more testing or better modeling etc. If it was a problem you didn't anticipate, you would need to have a better fault tree.

NASA is only buying a service.  It doesn't manage the launch vehicle.   He is only referring to the risk of using Astra

Offline edzieba

  • Virtual Realist
  • Senior Member
  • *****
  • Posts: 6105
  • United Kingdom
  • Liked: 9330
  • Likes Given: 39
Maybe its a change from relative velocity to inertial velocity.

Perhaps, but why is the elapsed time from first motion to MECO about 6 seconds longer on the video feed than for the telemetry?

Video Elapsed      02.56
Telemetry Elapsed   02.50
Video delay may be variable over time, or even at the same time between shots. Add to that the latencies of receiving the telemetry, generating the pretty-printed video overlay, overlaying that onto the muxed final video mix, and adding any additional intentional time delay to the final mix, and there's no guarantee a video feed will resemble real-time.
This is not unusual: we see on SpaceX's broadcasts that sequence callouts occur 'before' video of those activities (e.g. fairing sep) due to video delay, or landing shots from the droneship showing the vehicle has landed on one angle and still descending in the angle shown next to it.
For the public feed getting something out is prioritised over timing precision. Accurately timed and synchronised video footage can be reconstructed from embedded timecodes offline at a later date if needed.

Offline OneSpeed

  • Full Member
  • ****
  • Posts: 1588
  • Liked: 4930
  • Likes Given: 2078
Maybe its a change from relative velocity to inertial velocity.

Perhaps, but why is the elapsed time from first motion to MECO about 6 seconds longer on the video feed than for the telemetry?

Video Elapsed      02.56
Telemetry Elapsed   02.50
Video delay may be variable over time, or even at the same time between shots. Add to that the latencies of receiving the telemetry, generating the pretty-printed video overlay, overlaying that onto the muxed final video mix, and adding any additional intentional time delay to the final mix, and there's no guarantee a video feed will resemble real-time.
This is not unusual: we see on SpaceX's broadcasts that sequence callouts occur 'before' video of those activities (e.g. fairing sep) due to video delay, or landing shots from the droneship showing the vehicle has landed on one angle and still descending in the angle shown next to it.
For the public feed getting something out is prioritised over timing precision. Accurately timed and synchronised video footage can be reconstructed from embedded timecodes offline at a later date if needed.

Telemetry is often offset from the video, but for each camera, the offset is constant. The telemetry stream might pause due to a loss of signal, but if it resumes, it always catches back up to the same offset.

That's not what we are seeing here. There is a six second period where the rate of the relay of information has roughly doubled. After that period, the offset has been shifted by about six seconds, and remains that way.

If you've seen something similar in another webcast, please point me to it.

Offline edzieba

  • Virtual Realist
  • Senior Member
  • *****
  • Posts: 6105
  • United Kingdom
  • Liked: 9330
  • Likes Given: 39
but for each camera, the offset is constant
The offset is constant across all cameras (post-MUX), not for each camera, as there is a delay offset between cameras.
e.g.:
- Telemetry shows launch at t-0
- Pad cam shows launch at t+5 (5s delay for hardline pad feed)
- Pad cam switches to on-board cam, on board cam has a 10s delay due to relay via local compression, direct to ground telemetry, then decompression
- Telemetry shows shutdown at t+100
- On-board cam shows shutdown at t+110 (10s delay)
- Difference between launch and shutdown times based on video feed is 110s - 5s = 105s
There is now a 5s phantom mismatch between telemetry time and video feed time, due to the change in video delay between feeds. No mismatch exists in reality, it's an artefact of variable video latency. Adding to the headache is the same physical camera may be sending video through multiple routes (e.g. a direct to ground link local to the launch complex, a direct to ground link well downrange that uses its own satellite backhaul link to get back to the LCC, or relay via TDRS) at different points in the launch sequence.

We can make a reasonable estimate of velocities and timings form the on-screen telemetry repeater because it is a reasonable assumption that telemetry is processed in real-time to align timestamps and compensate for varying link latencies. That same assumption does not hold for the video feeds.

Tags:
 

Advertisement NovaTech
Advertisement Northrop Grumman
Advertisement
Advertisement Margaritaville Beach Resort South Padre Island
Advertisement Brady Kenniston
Advertisement NextSpaceflight
Advertisement Nathan Barker Photography
1