Starship heat shield

[Asteroza]

Tile detachment detection with that kind of granularity would primarily be for the initial test flights. Hopefully wouldn't be needed on a mature launch system.

Why not, in the case of advanced vehicle health management being a normal thing? Knowing you have a busted tile provides at least a chance to deal with it, if you have a means to do so...

[Oersted]

Just a resource and construction issue. Pretty normal to have highly instrumented experimental vehicles and production vehicles less so.

[NH22077]

Open question: would each and every tile carry an RFID chip... Snip

There might be RFID tags that can survive the temperatures... Snip

One thing I could see on the first flight is a simple free flying or tethered 360 camera deployed from the aft skirt after SECO....
Snip

I've brought this up in the past and gotten no traction. Acustic sensing. It probably wouldn't work when the main engines are firing but otherwise a network of mic's and some serious DSP's behind them could conceivably localize to individual tiles and with a good signal library, pin down what happened.

We've all heard how that hull rings when beaten by a hammer. It transmits sounds quite well. Every material has a characteristic acoustic signature. I doubt there is anything on the starship that would sound like a cracking tile except... a cracking tile. There is a lot of other noise but that's what DSP's are for. They pull small signal out of big noise.

Cracking is only one failure mode and even it has subcategories. How fine a discrimination is possible is an open question but at a gross level of tagging failures, I think it would work.

Any sonar operators out there?

Phill,
  Not a sonar op, but a live sound engineer.
The DSP will need to know what tiles cracking while the vehicle is at hypersonic velocity sounds like. Which would require a full size vehicle. Put into a supersonic wind tunnel to get the data the DSP will require. They might get away with putting the top 1/3 of a SS. Nosecone flaps & a straight ring or 2 under it. With the bottom wielded shut with a circular plate.
  And do the same with engine section and lower flaps with the top sealed, the engine skirt open. 
   The problem is I have never heard of a supersonic wind tunnel big enough. And using a 1/10 scale model would sound different. Maybe they could extrapolate the tonal differences. I don't know.

eriblo, I like the idea of deploying a cubesat or two with cameras.

Ned

[edzieba]

Not necessary to know the actual sound, but a calibration session would be noted applying some test sound (ideally an impulse. A hammer would not be unreasonable) in known locations to allow the time difference and reflections recorded by the transducers to be mapped to physical locations. This is a hobby-level operation, many people have DIYed touch-sensitive objects by applying contact mics to objects and calibrating desired interaction areas. It's also something SpaceX themselves have done, using the accelerometers on board CRS-7 as contact mics in order to localise the strut failure.
Acoustic monitoring would give you a map of tile events, through it would not provide information of why a given tile failed without more localised instrumentation (e.g. temperature probes, strain gauges on tile pins, etc). Remote sensing from the test range equipment would only provide exterior temperatures (which can be mapped to flow discontinuities too) but not temperatures of components beneath the surface. Though both would give a map of likely tile loss areas in order to more accurately target future instrumentation installations.

[Tangilinear Interjar]

Y'all are making a pretty big assumption that one of these very light weight, super insulated tiles with a veeeery thin glass layer, and is as isolated from the ship as possible (in most places) with a layer of fluffy insulation that probably works as a pretty good acoustic insulator as well, will actually make much sound at all when it cracks.

I'd be very surprised if the breaking sound signature was even within 20 dB of the ambient background noise.

[edzieba]

Y'all are making a pretty big assumption that one of these very light weight, super insulated tiles with a veeeery thin glass layer, and is as isolated from the ship as possible (in most places) with a layer of fluffy insulation that probably works as a pretty good acoustic insulator as well, will actually make much sound at all when it cracks.

I'd be very surprised if the breaking sound signature was even within 20 dB of the ambient background noise.

Maybe not the sound of a tile breaking, but the sound of the metal insert within a tile rattling around on or tearing off the metal (and welded directly to the skin) mounting pegs will. A cracked tile that remain in place and otherwise in large chunks is likely still mostly effective as TPS (once the vehicle decelerates enough for airflow to slow enough for it to fall off, its done its job) but a tile that fails so catastrophically that it affects TPS effectiveness is of more interest for that sort of vehicle-did-not-survive telemetry.

[OTV Booster]

Open question: would each and every tile carry an RFID chip... Snip

There might be RFID tags that can survive the temperatures... Snip

One thing I could see on the first flight is a simple free flying or tethered 360 camera deployed from the aft skirt after SECO....
Snip

I've brought this up in the past and gotten no traction. Acustic sensing. It probably wouldn't work when the main engines are firing but otherwise a network of mic's and some serious DSP's behind them could conceivably localize to individual tiles and with a good signal library, pin down what happened.

We've all heard how that hull rings when beaten by a hammer. It transmits sounds quite well. Every material has a characteristic acoustic signature. I doubt there is anything on the starship that would sound like a cracking tile except... a cracking tile. There is a lot of other noise but that's what DSP's are for. They pull small signal out of big noise.

Cracking is only one failure mode and even it has subcategories. How fine a discrimination is possible is an open question but at a gross level of tagging failures, I think it would work.

Any sonar operators out there?

Phill,
  Not a sonar op, but a live sound engineer.
The DSP will need to know what tiles cracking while the vehicle is at hypersonic velocity sounds like. Which would require a full size vehicle. Put into a supersonic wind tunnel to get the data the DSP will require. They might get away with putting the top 1/3 of a SS. Nosecone flaps & a straight ring or 2 under it. With the bottom wielded shut with a circular plate.
  And do the same with engine section and lower flaps with the top sealed, the engine skirt open. 
   The problem is I have never heard of a supersonic wind tunnel big enough. And using a 1/10 scale model would sound different. Maybe they could extrapolate the tonal differences. I don't know.

eriblo, I like the idea of deploying a cubesat or two with cameras.

Ned

With an array of mic's it should be possible to pin down the physical source of transients. Combined with high speed video/LiDAR would that be enough to start building a library?


Would the tile-bracket-pins-hull acoustic path give a signal independent of external atmospherics?

[OTV Booster]

Not necessary to know the actual sound, but a calibration session would be noted applying some test sound (ideally an impulse. A hammer would not be unreasonable) in known locations to allow the time difference and reflections recorded by the transducers to be mapped to physical locations. This is a hobby-level operation, many people have DIYed touch-sensitive objects by applying contact mics to objects and calibrating desired interaction areas. It's also something SpaceX themselves have done, using the accelerometers on board CRS-7 as contact mics in order to localise the strut failure.
Acoustic monitoring would give you a map of tile events, through it would not provide information of why a given tile failed without more localised instrumentation (e.g. temperature probes, strain gauges on tile pins, etc). Remote sensing from the test range equipment would only provide exterior temperatures (which can be mapped to flow discontinuities too) but not temperatures of components beneath the surface. Though both would give a map of likely tile loss areas in order to more accurately target future instrumentation installations.

I think you just answered my questions. Thank you.


It sounds like a combination of IR cameras and mics would touch a lot of bases.


Monitoring the tiles is important for development and ongoing operations. For development learning the reason for failure is top priority. For ops, it's most important to know you had an oops.


Hmmm. A name for the SX softball team? Noopsoops!

[OTV Booster]

Y'all are making a pretty big assumption that one of these very light weight, super insulated tiles with a veeeery thin glass layer, and is as isolated from the ship as possible (in most places) with a layer of fluffy insulation that probably works as a pretty good acoustic insulator as well, will actually make much sound at all when it cracks.

I'd be very surprised if the breaking sound signature was even within 20 dB of the ambient background noise.

My musicians ear tells me that cracking ceramic is a sharp transient with little or no ring. Not much area under that waveform but quite a bit of amplitude.


The pins give an acoustic path but in the end we will never unless it's tried.


The sound you hear is the frantic scurrying of SX lurkers digging out SS accelerometer data.

[adrianwyard]

I'll admit this is an interesting discussion, but to point out the obvious: job #1 should be demonstrating an attachment mechanism that can keep the tiles in place for repeated static fires and/or wind tunnel testing. No harm developing a fancy way to detect when/where they fail in flight too, but if they stay on during ground tests this may not be needed...

[Hog]

After viewing a few of the latest videos of working with and removing TPS tiles, it appears as a labour intensive system.  It will be very interesting to see what Space X goes with when SS/SH enters it's operational phase.  In the meantime, some flight testing of the TPS will be very exciting!

[Nomadd]

 Paid the old homestead a visit today and cleaned up the mudflats some.

[BT52]

OMG very nice.

So how u can sinter with metal inserts? Got no idea how they do that...

Seems to me texture implies not machining is needed. Also that would be preferred methodology to ease mass production.

 

[matthewkantar]

Nice that SpaceX’s mom wrote their name on it.

Edit: it seems like the metal insert is steel, but not a stainless alloy?

[InterestedEngineer]

Nice that SpaceX’s mom wrote their name on it.

Edit: it seems like the metal insert is steel, but not a stainless alloy?

It is stainless.  It looks stainless, and it's been sitting in a salty estuary for a while.

Aluminum would have had a telltale corrosion, as would almost any other metal besides stainless.  The tiny streak in the faster looks like very mild stainless corrosion to me.

[RamsesBic]

OMG very nice.

So how u can sinter with metal inserts? Got no idea how they do that...

Seems to me texture implies not machining is needed. Also that would be preferred methodology to ease mass production.

I have been told the brackets are pushed inside after the tile has been sintered. On some images you can see the edge of the bracket sticking a mm or so out of the side. 3 brackets per tile.

[eriblo]

OMG very nice.

So how u can sinter with metal inserts? Got no idea how they do that...

Seems to me texture implies not machining is needed. Also that would be preferred methodology to ease mass production.

I have been told the brackets are pushed inside after the tile has been sintered. On some images you can see the edge of the bracket sticking a mm or so out of the side. 3 brackets per tile.

By someone who would know or who is speculating? This would seem like the absolutely worst way to do this from what we know about the tiles..

[RamsesBic]

OMG very nice.

So how u can sinter with metal inserts? Got no idea how they do that...

Seems to me texture implies not machining is needed. Also that would be preferred methodology to ease mass production.

I have been told the brackets are pushed inside after the tile has been sintered. On some images you can see the edge of the bracket sticking a mm or so out of the side. 3 brackets per tile.

By someone who would know or who is speculating? This would seem like the absolutely worst way to do this from what we know about the tiles..

By one of the regulars here.

[Nomadd]

OMG very nice.

So how u can sinter with metal inserts? Got no idea how they do that...

Seems to me texture implies not machining is needed. Also that would be preferred methodology to ease mass production.

I have been told the brackets are pushed inside after the tile has been sintered. On some images you can see the edge of the bracket sticking a mm or so out of the side. 3 brackets per tile.

It does look like there's a path through the material where the bracket was pushed in.

[eriblo]

OMG very nice.

So how u can sinter with metal inserts? Got no idea how they do that...

Seems to me texture implies not machining is needed. Also that would be preferred methodology to ease mass production.

I have been told the brackets are pushed inside after the tile has been sintered. On some images you can see the edge of the bracket sticking a mm or so out of the side. 3 brackets per tile.

It does look like there's a path through the material where the bracket was pushed in.

Is there a visible difference between the original and fracture surfaces on the tile? What does the tile "channel" surface look like? Are the brackets free to move or are they rigidity attached to the tile? Can a metal object like a nail be pushed into the tile without excessive fracturing?

The reason I have been sceptical of the "pushed into the sintered tile" claim is that I thought the tiles would be too brittle/fragile. Most people would not buy the explanation of glassware with embedded objects to be "you just have to push them in carefully...".

If the brackets are embedded before sintering they have to withstand temperatures of 2300 °F for silica tiles and up to 2600 °F for something like AETB. The former should be doable with some steels or nickel alloys while the latter would be difficult.

If the brackets are added after sintering I would expect the channel to either be cut separately by a saw or vibrating tool or by vibrating the bracket while it is inserted. It would then be free to move unless cemented in place afterwards. Independent movement is not a drawback with regard to thermal expansion as long as it is small enough for the holes line up.