Astra is a fascinating rocket company
that has had some great successes and some unfortunate failures. They're bouncing back with a brand new
bigger rocket and improved manufacturing and quality control. In this video we'll get a
detailed look at their new rocket, which they call Rocket Four. We'll
talk to their ceo, Chris Kemp, their vice president of manufacturing
Bryson Gentilly and will try to learn exactly how they plan to manufacture. Rocket four at scale and manufacture at
scale is exactly what they set up to do. They've built an assembly line
designed to churn out a rocket a day, which is nuts, but is that even possible? What has to be done on a company level
in order to enable such rapid production of such a complex thing?
Let's take a look. So. Today we're going to show
you a couple of cool things. We're going to show you how raw material
becomes a rocket through the factory. So last time you were here, we
started from our front lobby. This time we're going to
start from our receiving door, which is where the action
in the factory begins. Excellent. So parts go in that
door, rockets go out another. Door, raw materials and parts come in
the door right behind us over here, and then rockets go out the other
side of the factory. Awesome. So what you see on the rack behind us,
this is a whole bunch of raw materials. So one of the nice things about being
a vertically integrated company is we actually understand the sizes of
parts that we traditionally make. So we stock a lot of material that's
in common diameters or shapes in long bars. And these bars are actually wider
than the door, the receiving door is. So we store them outside and we
actually bring them, if you follow me, we bring them over here.
We have a cutting area. So the factory actually begins outside
of the walls of the factory. I like. It. And this machine over here, it's called
the cold saw. So we take these bars, we bring 'em over here, we forklift them over and then
we ch cut off little sections, little slugs of material. And then we take those and we bring them
in a machine shop and then parts start their journey here from these large bars, which we produce at a much lower
cost than buying little small bricks. So we'll hop actually over this barrier. This is not normally a tour friendly path. So we'll hop over this barrier and
then we'll walk into the factory here. So we walked up the ramp past the
aforementioned receiving door and into the factory proper. I was immediately blown away by the
increased scale of everything at a. Very high level. Since we've
been on our last tour video, a number of things have changed here. Inventory has doubled the
size it was before the shop, which we're about to jump into is about
triple the square footage of what it was before. And this factory overall is almost tripled
the square footage of what that last tour contained. Well, last factory tour you said We have all
this space to expand into and now it's very clear you have
done exactly that. Yes. All going according to plan since the
early days of looking at this building and kind of having a vision of
how it could come together, we've been executing as we've expected. A lot of these pieces of equipment
are very specialized for rockets. Actually many of them are not as well. We prefer the ones that are not because
they're easily available. But yeah, I can walk you through some of this
as we walk through the place. Cool. Should we put on our handy
dandy eye protection? Yes. Thank you. So let's put on our safety
glasses and jump right into the shop. Safety first. So we're entering the shop right now. The shop is controlled to a tighter
temperature window than the rest of the facility and that allows us to make
sure that as we're machining parts, the machines and the parts themselves
are expanding and contracting with temperature at a known rate.
So we can always hit very, very tight tolerances here in the shop. So the strip doors are actually mostly
just to keep a temperature controlled environment. Makes sense. We're going to pop into the sheet metal
shop and we'll start the journey for a sheet metal part here. So very similar
to the last time we came through, parts show up on the sheet metal rack
or raw material shows up on the sheet. Metal rack then gets
loaded into the water jet. We cut flat patterns and then those, maybe I'll just grab a prop here
instead of waving my hands around. I love that you have a handy prop case. Oh yeah. So I came through
here enough times that Scott, who is right over here speaking with a
couple folks in the shop actually decided he was going to make me a display case
so that I didn't have to keep asking him for things. Nice. So
that's us iterating live. But these are a couple examples of parts
that started their life as sheet parts. So these were flat sheets that we then
cut and then brake formed on the brake press there. So these are
parts essentially formed to
be able to get stiffness. This particular part started as life
originally as a machined part and is now a brake form part that is lighter and higher
performing than the machine part get replaced. And if you're not machining it, you don't have to lose a bunch
of material to that process. You have the material, you're using it, you're not throwing away
a bunch of material. Exactly. Machine parts I used
to think were quite cool. They're now deeply offensive to me because
you throw away generally most of what you paid for when you're
making a machine part. So these sheet metal parts you use a lot
of the material you paid for so you get low cost parts and they also, these particular styles of
manufacturing break forming is very high utilization and very low cycle time and
means you're putting not that much money in the time that you invest into the part. So you can make these very easily
to be scaled up at very low cost. Alright. The machine shop was pretty
similar to the last time we had a tour, but that's where the
similarities end. As we moved on, Bryson once again underscored how big
the changes to the factory have been. As we get into the next part of the tour, we're passing by basically a bunch of
active machines in the machine shop. So please forgive any
less than perfect audio. So now we're entering kind of the main
part of the machine shop and what we would've seen previously is a curtain at
that first kind of set of columns when we were here last. Time. Right. It's like
double the size now. Yeah. So the shop actually used to end
right over that vtl by column b2. Yep. We've actually expanded all
the way up to that last wall. So the overall square footage of the
shop is almost triple what it was before. And we'll walk you through how we've
expanded and how that's changed the shop and its arrangement here.
Cool. So let's keep going. And as we get into the shop, what you
see first is a lot of manual machines. So we do a lot of quick turns sort of
development work and the manual machines and the semi-manual, semi-automated
machines are really handy for that. They don't take a lot of programming
time, not a lot of setup time. This is an area that's really important
to us, particularly for development. And we need to do really,
really rapid iterations. Maybe a part turnaround in a couple of
hours rather than days or weeks. Yeah, that makes sense. The manual shop is kind of the
first entry as you walk in here. What are we doing over here?
What are we doing over here? That's a good question. Hey
Dave, what are you making. Bearing simulators for
balancing locks rubber. So bearing simulators for
balancing the locks rubber. That's the first stage
engine, correct? Yeah. So Dave's working on some first stage
engine parts. That's cool. Cool. Thanks Dave. Oh, we have different levels of
automation within the factory. So we had the pure manual machines. We talked about some of
the semi-automated ones. That's where you program certain
features in. Those are repeatable. And as you kind of scale up in
higher and higher volume production, you actually want to change the
style of manufacturing you're doing. So these machines are all CNC
controlled, computer numeric controlled. So they're all programmed,
they're highly repeatable. We have very specific fixtures,
very specific tools. Cool. And then as we go to the
production shot of the shop, I'll get into the details of what we
do differently there to enable another level of scale. So we're going to peruse through here and
we're going to kind of cut a nice tour path that might interrupt some
of our machinists a little bit, but I think it'll give you
some good visuals of what
we do in the factory. Sure. So Steve is working on a part
here in this machine. Steve, what are you working on? This is so lock vein housing. So we got a lot of first stage engine
parts going through the machine shop right now. We got Dave working on
one, Steve working on another. This is one of our more
challenging five access parts. So thanks Steve. You're welcome.
We're going to carry on. It's a good looking part. Ah, yes it is. What's up Daniel? What's up Brett? Daniel's working. What are
you working on here? Daniel? These are brackets, these are
for the thrusters. It's like basically the gas tank for the thruster. This radius is what kind of locks it in. So how many of these are you making? 30. Cool. Yeah, they asked for
30. I might give 'em 32. Nice. If I screws you up. So Daniel's working on some
astro space engine parts here. And this is an example of maybe one of
the higher batch quantities that we've got that goes through the
development side of the shop. So we're making 30 of this part, right around 30 is kind of the maximum
that we would be running in this side of the shop once we start passing about
30, we want to go to maybe one of the, that we'll talk to in a few minutes. It's so cool to see it move
around on the different axes. Xi Xi. Yeah, this part is running on five
different axes right now. Saves. Saves a lot of time.
Versus the three xis machine. The three axe machine. These
would all be different operations. See how the part here, the part
has tapped holes like all over. So if it was a three axis part,
here's one operation from the top, here's another operation, here's another,
here's another and here's another. So that would've been five different
operations we were able to do in one. Nice. Yeah. Fewer operations the
better. It means fewer tools, it means less setup time overall.
It means we get parts faster. Very cool. It also just helps you
hold really tight tolerances too. Yeah, you don't have to reaff fixate and
get the fixture just right every time. And even. Say if you were trying
to hit this operation, you just figure out how to hold it to
make sure that the top is super flat, you dial in and the X and the Y. Cause you don't have any flat edges
and it ends up saving a lot of time if you're able to do it for five access. Very cool. Cool. Thanks Daniel. All right, thanks. Let's carry on. Ah, I see there's some of the 30. Yes. That is the rack of those,
they're, most of 'em are through op one. I think he's got, I see four billets left
that haven't gone through op one yet. Very cool. So as we'll walk down
this aisle, we'll go little faster. We're getting to lunchtime here, so folks are starting to
peel away from the shop. No. Worries. But you see, we run some of these machines when
they're doing simple operations, you can basically run even
some of these unattended. That's so cool to see. I imagine it gets boring after
a while to see the machine run, but I'm like a kid in a candy store right
now. It's like, oh look, it's, it's. Milling. Yeah, the machines off
pretty sweet. It's like the most, it's almost like an art form, right? You're taking a brick of raw metal and
you're turning it into an actual usable part with real shape. So yeah, the machine
shop never gets old. It's very good, very amazing capability to have for a
lot of reasons on top of the fact that it's just cool. Yeah. Nice. So let's keep going. We're walking past
some blades now. We kind of started, we looked at a couple of the
mills that are in the shop, a lot of machines that we have here.
But as we transition over here, we're now entering the chip hallway and
the chip hallway divides the development shop from the production shop. So yeah, we even think about how we're going to
get rid of these many, many, many pounds, many thousands of pounds of chips
that we make on these machines. Do they get recycled? They do get recycled. So all of
our chip hoppers are basically, not all of them, most of the chip
hoppers are aligned with this hallway. And then those chip hoppers travel all
the way down to a roll up door at the end of the hall where they get dumped into
a large chip hopper or a number of chip hoppers and they get recycled
as specific materials. Nice. Cool. Let's head down the hallway here. So we're going to talk a little bit more
about the production shop right now as you're we're walking down here you
see a few different types of machines. These are three machines concentrated. These are wire edms
electron discharge machines. These allow us to do extremely tight
tolerance cuts and hole drilling. We use these for very specific features
on both the spacecraft engine and the rocket. We won't talk at
great length about these ones, but we do have this capability now
and it's a recent capability as we've expanded the shop. Very. Cool. Yeah. Let's keep going over here. So on this end of the shop, this is
the far end of the production shop. What you're looking at are a couple of
lathes that do have live tooling and they're bar feed lathes. So theses are pretty cool in that
you can automatically kind of keep a bar loading itself as it makes a
part and it parts it off and it drops it into one of those little buckets. And. It'll. Just keep making parts. Just keep,
once it's made apart out of the bar, it'll separate it from the bar, then feed
more bar in and then make more parts. Exactly. And even when that bar is done, it'll grab the next bar and
then start feeding the next bar. I feel stupid, but it's
like a hot glue gun. When you need more hot glue you
just put the next stick in. Yes. This looks very fancy. Hot glue gun. Yeah, I think the machinist might be a
little offended by that. Yeah, I. Hope I didn't say that too loud. So Nick and Matt here are
a running machine. Nick, what do we got going on right now? So right now we are cutting turbo
pump components for our first engine. What makes this machine really cool is
that it combines both burning and mailing capability in a single package
and add an automation element. So while our development side of
the shop is more focused on rapid turnaround, the production side is more
focused on efficient use of resources. This machine that, for example, has bar feeder down on that end
parts catcher and the milling capabilities. So really what that gets us is the ability
to run the machine further into the night unattended. Nice. Cool. Thanks Matt. Alright, so Nick mentioned what sort of
capability we get out of these machines. They're called lights out machines. Typically it means that we can basically
program a part or program a job into the machine program, exactly
what material we're going to use, what tooling we're going to use, and
then the machine can run unattended. Nice. So you're, when you're
leaving for the night, you give it a whole bunch of bar
stock, you tell it what to do, you come back in the morning,
you have a whole bunch of parts, you have a hopper. Full. Parts. Love it. Yeah, it's
beautiful. I love that we got. A lot more of these machines
in the production shop. That's kind of what the
production shop is all about. We actually organize the shop differently. I mentioned before we have a machinist
that takes a job from beginning to end in the production shop. We have
several different roles. So we have a programming role, kind
of a set up role operator role. And so that really enables us to
ultimately prep for where when we might run or run multiple shifts, but also it divides the rolls up so
we can keep these machines fed nice. They want to be fed, they have a lot
of capability and we want to keep 'em. Moving. And I imagine they're
at least some of these, if not all of 'em are
significant investments. So the more time they're
working the better. Exactly. So right now we're in front
of a five axis pallet loading machine. So as you heard earlier from Daniel,
he was making some five axis parts. It's very handy, particularly when you're making engine
parts which have maybe complex thermal machinery or complex geometries
to be able to use five axes. And so a lot of our engine parts
will go through this machine, which is a lights out
version of a five axis. Wow. Okay. So there's a place where you feed in
bar stock and it just does its thing. Yeah, this will see receive billets. So maybe you take the bars and you chop
those up so those will become bricks and you'll feed those into this machine. So it'll start from slightly different
thing from, it'll won't start from bars, but it'll start from maybe
usually rectangular bricks. But basically you have a way to input
a whole bunch of raw material and when it's used the raw material,
it moves onto the next piece. Yeah, exactly. And we'll show you another machine
that'll make it super clear how that all works. Oh cool. Very cool. And just
actually let's walk over to it now. All right, so over here
we've got a horizontal mill. This has a similar operating system
where it has pallets that you load up. We'll show you what those load
on those stations look like. But I wanted to point this area out
because you can kind of get a good idea of what some of the pallets look like. So each one of these pallets has kind
of a vertical set of fixturing on it. We call those tombstones. They have lots of different
options for how you would fixture, you can make big parts on them, you can make lots of small parts on them
and those would hold raw materials or in process jobs as they're running.
You can actually see the pallet changer is operating,
it's moving around, it's grabbing different pallets right
now and it's moving them from station to station. You can see
it's placing a tombstone, I believe it's placing a tombstone. I can't really see from my
angle placing a tombstone into a transfer station, which is
a 180 degree kind of rotation. So we're loading up a next job right
now while it's running a job in the machine. So let's walk around this way
and I'll show you what the operator side, we're kind of on the backside
of the machine right now. As you walk along the backside, you can also see we have a
number of different tools. Again, because we're vertically integrated, we actually can kind of control what
sort of features we put in our parts. And so we have common radii, common
tapped hole geometry, common drill depths. And so we actually stock standard tools
in most of our machines to be able to hit the same sort of features. So it's really easy for us to take a
job from one machine and then re-hit the same sort of tolerances
and shapes on another. Makes sense. We have about a hundred tools that we
use that's kind of standardized tools. And so this machine has more than twice. That's about three times that many tools. So we can do a lot of custom
tooling on this machine. And this is a lights out sort of machine. This is the biggest workhorse
we have in the factory. So we do expect the most tooling
to be in the changer here. I'm noticing a trend where a lot of
these machines are lights out machines. We're in the production shop. Yeah. Exactly. I like it. So let's keep going and I'll
show you the business side of this. Oh, we caught this at a good time. Oh, that is neat. See you see his tombstone being loaded
into the load on unload station. So this is a little different from the
CNC machines you saw in the development shop. We interact with
them in a different way. So typically we'll have a set up and
operator folks that'll be interacting with these two areas here. So essentially we'll load raw
materials or work in process parts onto a tombstone with custom
fixturing for those jobs. And then we will tell the computer what
fixtures and what raw material is loaded and what jobs we need to produce. So those will then go into the pallet
rack where they'll be stored into our job queue, which is essentially how
we can run throughout the night. So that job queue will be fed
into the machine here and we'll essentially run apart just
like a regular CNC machine, but there's a whole bunch of
automation around feeding that machine. Okay, so you said tombstone? Yeah. Is that the black vertical part or
is it the metal part that it's on? That's the pallet, that is the. Black vertical part. So the
pallet is kind of the base. The tombstone is a style of
fixturing that we use here. Got it. And it allows us, we essentially have four sides available
to us around the tombstone fixtures. So we can index them and we
can actually make lots of, you see Sean actually rotating it around. A good example of how we
use the tombstones here. So we can make lots of parts
with four axes on this machine. It's a horizontal mill, typical ally four. So you don't even necessarily need to be
making a whole bunch of the same part. You could just put something
on this side of the tombstone, something on that side of the tombstone, and then the machine just knows on this
side I'm going to do this and this on this side, I'm going make
something completely. Different. That is exactly correct. Very cool. Saves time. It does. And we have a standard ways of attaching
our fixtures to all of the tombstones. So it's very easy for us
to very repeatably change
those out with very little setup time. Question. Yeah. Why? What's
with the green light? So all the lights mean different things.
A green light essentially means it's, it's good to go. It's operating each
light color means something different. A blue light for example,
means there's no activity. Like it's. Safe, it's in kind of manual mode
and it's not in automated mode. So if you walk up to these and they're
red, it's like there's a problem. The computer's. Mad. So if you get a red light and any of
these stations actively a problem, you'll see on many of the machines,
there'll be colored lights. An andon light is what that's typically
called and it indicates status of some of the machines. Very cool. Yeah. One more thing I want to point out. So this is basically a big cell and we
have a lot of custom tooling on here and sometimes we don't want to run the
horizontal mill to make our tools. So we have a regular three xxi CNC
machine right next to the machine. We make tools for this
machine right next to it. Have a big machine that makes the tools
for the bigger machine. Yes, I love it. Exactly. Can I look in this window right
here? See I saw something spinning. Hey Shawn, you want to talk to
us about what's going on in here? So we just switched out pallets right now
and then we're starting a new program. So what's nice about this machine is
it'll automatically load a pallet in the back while the part's running inside
here. And as soon as that's done, those pallets will switch out for at
least amount of downtime as possible. So right now we're just running
spindle, warm up, with's, some cooling going on and as soon as
that program is done we'll switch out to the next pallet and for the next. One. Very cool. Yeah, thank you. Cool. No problem. Thanks Sean. Absolutely.
Yeah, we're going to head this way. What's up Dan? So this
is our maintenance hub. This is where we command and control
maintenance for the entire factory. So we just saw Dan over there. Dan actually maintains
most of the equipment here. This station right here
is our finishing station. So this is where we might install things
like threaded inserts or do little finishing work to some of our
machine parts. Makes sense. Before they come into the QC lab, you'll notice that this is in
the production side of the shop. So this is a recent thing
that we've added in the Astra. As we've been able to invest in the
facility and our ability to go produce things at scale, it's very important that we have parts
that are going to work when they're on the production floor. So we
don't actually stop production. And the QC lab is how we
ensure that those parts. Just like you guys said in space
tech day, reliability and scale. Reliability and scale. You got it. Wow.
So maybe I can grab Melissa. Melissa? Yeah. You want to describe to
us what's going on in here? Hello. Hello. This is
the quality control lab. We inspect all of our machine parts
from our internal machine shop and we also do supplier validation for purchase product. As well. Makes sense. So we sort of have a two-way
stream going on here. Sometimes we'll get something from a
supplier that will come through our lab that we check and then we do some
post machining on it. So for example, we have these ca here that
we'll inspect from the casting supplier and then they'll get
a secondary option operation on the Michigan. Nice. What sorts of
equipment do you use? Huh? What sorts of equipment do you use? Stuff we have a ton of hands tooling. We particularly use that
for sort of in process inspection for the machine shop. So each operation we validate
the machinist programs and then we also have sort of our shining star here. Yeah. What is this guy?
What do you even call this? It's a direct computer control
coordinate measuring machine. Neat. We say DC dcc, cmm, and frequently just DCC
because it's a lot of letters. This one in particular is pretty special because frequently the probes on the end, which is where most of the
business happens are three axis. So this fun spinny movement that
is doing most CMMS don't do. And this allows us to
do get two things. One, it's incredibly accurate. Makes sense. It actually gives us more data than
we actually need for certain things. So we actually have to whittle
things down a little bit. And it's also in addition
to that incredibly fast. So right now the machine isn't
even running it 100% speed. It's only running at about 25%. Work harder machine. No, I'm kidding. So we use it for two sorts of things. So we have our spacecraft engine product, we use this to do acceptance
inspection mostly, but for Rocket four on system 2.0, we actually work really closely with
engineering and the machine shop to provide data back to engineering and to machine shop so that we can have a tighter loop on
the design and development process. You can just iterate
faster that way. Neat. And. We can do, so this for example is a cam shaft for our first stage engine
flight termination system. Oh, that's cool. Yeah, it actually has some
pretty tight and pretty complex geometry. We're able to import the 3D
model that the engineer is designed and the machine shop has machined into the software and do an inspection
and compare it directly to the cad. And so if it's off the machine says, Hey, this isn't what you thought it was going
to be. Or if it's just right. Yeah. It says hey, it's just right. And for the design and development loop, sometimes it makes it so that we know
that as we're going through testing, engineering has that data. If something is looking kind of weird
or say something looks really great, we can either tighten
tolerances on things to really work in and get something smaller. But sometimes it also means
that we can open them up, which makes it much easier for us to
manufacture things much more quickly. Makes sense. So it's beeping right
now. Is it mad at me? Did I offend it? No, it's beeping because it's actually
taking individual points right now. It's also capable of, so one of the other sort
of special things about it. Oh, it's going to switch tools. It's going to switch tools is, so typically these probes use kinematics to take points. So actually it takes pressure and does
displacement. It's like cosine error. This one actually has a laser, the stylus here, this carbon fiber shaft
is actually hollow. And there's a laser in it. Laser. There's a laser that wow. Shoots
to the end of it and then back. And that's how it takes the. Points. That's cool. So it's actually able to, I mean it can take tens of
thousands of points in a matter of seconds. So neat. That's why I the I need my space. No out of the way it needs its space. I'm fortunately short enough I
can get every year this time. But yeah. Very cool. Yeah,
it's my favorite. Yeah, you do all kinds of stuff here too. What's going on over here?
Is this like a microscope? A microscope, but it's also capable
of doing dimensional inspection. So we use this for some of our non-contact things that we
need to get into inspect. So this is a boar that gets, this is one of our manifolds for
our spacecraft engine and the boar on this is very sensitive
to scratches and things. So we don't want to touch it. So this allows us to take
non-contact measurements on this particular feature. Neat. Yeah. Cool. Thanks Melissa. I think we're going to get back out the
door into the main hallway and yeah, appreciate you showing us around. Yeah, thanks so much. Cool. Oh
really quick. Who drew the unicorn? Not me. Well it's. Great. So are you seeing a trend here? Lots of automation and I love
the term lights out machines. Having the quality control lab right in
the shop is a smart and logical thing to do and it helps Astra iterate and
ensure quality in reliable parts. Now let's take a look
at the star of the show, the rocket assembly line
and talk to CEO Chris Kemp. So we're back in the main aisleway.
This is where raw materials come in. So among those raw materials, actually a coil of aluminum comes in
that door and goes all the way to this line. So we're going to get into a little bit
more of the factory and I think I see Chris is actually walking up here. Oh, a wild CEO has appeared.
What do you guys do with. Cameras in here? Hey Chris. Hey Kristy. Nice to see you guys. So Chris, what do you
think of the factory here? Awesome. You guys have
reached my favorite part. This is our new rocket production line. So we just covered that. The coil comes through the door and
we haven't talked about the machine. So basically a coil of aluminum gets
inserted in one side of this machine and we make rocket stages out the
other side of the building. So we can turn one coil into about
three rockets and the coils cost $25,000 or so about that. So we're really working on applying
automotive manufacturing to rockets. Nice. On this. Line. But it's the new rocket
production line to Rocket four. It's cool that it just goes from
coil to rocket all in one device. Okay. It's got different stages,
but if we can want to walk them. And why don't we just show it off? Cool. So the first thing you
see here is de coiler. So that 5,000 pound coil gets
loaded onto this loader here. Yep. Brings it into the machine
and then it actually gets fed. This machine is largely automated so each
major piece of equipment on this line is automated and some of
the transfers are manual. Eventually we can always automate
those transfers and then increase our production rates. Makes. Sense. But what you see is the coiler gets
decoyed and feeds into this straightener. So it comes with a kind of curvature
as it comes out of the coil. We flatten it with a number of pinch
rollers and then we take it into a tensioning loop. So this allows us to deal with a
little bit of non straightness. If there's a little bit of non
straightness coming out of the coil, they're not always perfectly trimmed. And it also allows us to decouple the
mechanical motion of two machines. That makes sense. So it allows us the flexibility to have
a control mechanisms that don't have to be timed to the nanosecond. Chris, you said this is your
favorite part of the factory. Do you have a favorite part of
your favorite part of the factory? I imagine it's when the finished
piece comes out the end. It is. Well, in a lot of ways this whole machine
was designed alongside Rock four and every single aspect of this,
such as the width of the coil, the ability to kind of
cut and clean up the has been designed with intention.
So imagine not having this, you'd have to have all of these
pieces come in on pallets. There's shipping costs that get
damaged, you can't make adjustments. So this gives not only the team an
opportunity to design and iterate more quickly during development, but also it dramatically reduce the cost
of finished product reward production. Makes sense. Yeah. I mean you
don't just build the rocket, you build the machine that builds the
rocket and everything efficiency wise follows from that. Yeah. This is a true rocket production line
and it's designed in the very spirit that an automotive production line is
designed when you have ram fuel up on the one side and you have the
primary structure of the
vehicle from out the other. Side. Nice. You'll see the. Whole thing. Different. Did I hear you
right on space tech data? This is supposed to make a rocket a
day or can make up to a rocket a day. That's right. It was designed to produce
a rocket a day out of this facility. And some of these stages can actually
produce even more material than that for development currently. And then in some stages we would add
another module and of course you have more labor to accomplish the higher perion
rate and all of these machines are designed to reduce the total
labor content of the vehicle. So that's really what drives costs. Nice. I like driving costs. Cause that
means cheaper access to space. Indeed. That's exactly what it means. So
yeah, when Chris says a rocket a day, these pieces of equipment are specifically
designed around a rocket a day in one shift as our Wow. Total capacity. That's so cool. So. We're not making this stuff up. It was, it's really marginally more expensive
to go from a rocket a week to a rocket a day if you're designing the
vehicle and the factory and the line. And so we decided this is one area
that we want to invest in the factory, that we want to go all the way to
rocket today and just do it once. And so we've basically sized
up the diameter of the rocket, we've sized up the equipment on the
line and everything was all designed in concert so that as we go onto
scale up and potentially have future versions of the rocket, we already have the equipment in place
and we don't have to redo this work. Make sense? And think about the utility of this. A rocket a day is not really
a crazy amount of capacity. And we think about all the small
satellites that all want to be placed in particular orbits on particular schedules
and all these mega mega constellations where you can replace failed
satellites, you can do gap filling, you can do better capacity management. This really fills in a gap much like
small airplanes fill in a gap where large cargo planes operate kind of the bulk. And the, yeah, I want to nonstop, but I don't have to go to a
hub and then go to my stop. And then you guys are making the
nonstop versus the large hub bound bigger. Yes. It's very much like a bus, you know, can take the bus but the bus is going to
leave at the wrong time and it's going to drop you off at the wrong place. You're going to have to spend some time
getting ready to go. Yes, that's great. But it's all about economics, right? If the Uber is only a little
bit more expensive than the bus, you'll take the Uber, but
it's a lot more expensive. You'll just wait for the bus. And so we're really trying to drive the
economics to the point where choosing a dedicated launch is always the right
choice for the majority of payloads. Very cool. What do you guys
say? Daily space delivery. That I believe was our wifi password on
the first day of the company we set out with this. Wow. This goal. Yep. Very cool. Worked tour it. Ever since. I mean it's clear the progress, the difference in the factory since
the last time we were here. You. Guys were here and then
you came back and this was. Here. This is amazing. I mean this was all cubicles last time
and now it's a rocket building machine. So it talks about the tensioning
loop, what that is for. So we come into the main laser bed. So this is sized so that we can make
a continuous cut on a laser cutter. It's called blanking is the term
that's used in automotive. Nice. So we do laser blanking here. And the reason this machine
is the size that it is, is because we can do one
continuous rectangular, mostly rectangular cut that will bring
a barrel together at rocket forest diameter with only a
single friction ster weld. Scene. Oh, that's satisfying. Yes, very satisfying. So it really reduces the amount of
times you need to be on a machine, which is part of the way that we get
to a overall time of a rocket a day. I got to say it too, I love the
aesthetic. It's an Astra looking machine. Very astra looking
machine. Yeah, we designed. It was a custom machine. Yes, yes. It took several years. In fact, it arguably took longer to build this
production line than it did to develop the first iteration of the rock. Up. Yeah, because you guys, I remember even on our last factory tour
and when we did our interviews with you guys, this is something you've been
working towards for a long, long time. Very long. It must be very satisfying
to see it come to fruition. Yeah. It's incredibly satisfying. This Chris and I have been talking
about this for six years, since. Day one. Yeah. I mean you're
either all in or you're not. You're either going to do something
that makes the rocket expensive and then you're really impaired to the Starship
and some other rockets that are coming online. Or you're going to go all in, you're going to invest in the scale and
you're going to invest in all with the liability and the vertical
integration that we've invested in. And it's a big bet. It's worthwhile though because you know,
you guys say improving life from space, I really, really appreciate that mantra. Improve life from space because so often
when people maybe aren't space fans or rocket nerds or what have you, or
maybe they just don't know better, they're like, why are we going to space? And it's like you guys say
it right on the box. Yeah. We're going to improve life
on earth from space. Yeah. I mean you look at so many of these
killer apps are planet I think is a killer app. You take a picture of earth
every day, what's going on? Starling killer app, right? The
ability to provide global connectivity. And these are all focused on
providing people earthlings benefits. And that's not to take anything away
from becoming a multi planetary species, these rockets that we're going
to need to build to explore. The solar system. Of course not. This is a really unique mission. It's a distinct mission that
we really are inspired by here. Yeah, no, it's like we can do all
the things. We can be multiplanetary, we can help people on
earth. We don't have to. It's not about forgetting earth or
leaving earth, it's about helping earth. Love it. I really love that.
Cool. More machine. Yeah. Let's keep, so let's keep walking
this thing. It's rather large. I see Nate's hanging out over
here. Maybe Nate, maybe Nate, you want to talk us through
what happens after we laser cut? Sure. This is our new laser cutter. This machine will produce the large
format blades we need for our first stage and upper stage tanks.
After the pattern is cut, some smaller scrap will
fall away and exit here. Then the material will
transition out onto this. Now here we'll remove larger scrap as
needed before the material then moves towards the four bar. After the material transitions from
the alfi conveyor in this area, we clamp the material to
square it with respect to the. Four bar. Roller. Here we roll it into the shape we need. This machine was designed to roll the
correct radius right up until the very edge. Of scenario. So it comes out of the
laser blinker, it heads over here, you square it and then you roll
it. Yep. Very cool. And then. Onto the friction, sir welder,
where we weld these linear seams. Together and you only need one seam.
That's right. Love it. Nice and simple. Yeah. Cool. Well thanks
Nate. Yeah, no problem. Cool. Yeah, so Nate mentioned we use a four bar
roller here so we can roll all the way to the ends of the barrel. So we have a
nice perfect cylinder, typical roller. He uses three bars. Those the bars we're
seeing on the bottom there. So you have two pinching and
then two that you use to form. So because we have a four bar
roller, it means that we go, you can see the radius of that barrel
is all the way to the very edge of the barrel. So we don't have a flat section.
And when we go to weld the tank up, we actually offset our
friction, sir welds. And so it means we have a nice cylindrical
circular pattern that we can weld easily with simple fixturing
on the next machine. Love it. So you don't have to cut off a flat
strip on the end all about optimizing the time that the things spend in each station
and going faster through the process so that you can get to that rocket a day. Exactly. And so it takes a little less than 10
minutes to actually run through a roll on this machine. Wow. 10 minutes. Yeah, it's pretty wild. It's all
automated. Each part of this is automated. In Rocket three, we had to do this at various stages
and the material would come on pallets, one cylinder at a time, and a lot of
times it would be damaged shipping. And then of course you've got a huge
amount of floor space being taken up by all of this inventory of
giant aluminum cylinders and so on many dimensions, it's a lot better just to
turn a coil of aluminum into exactly the material that you need. And then you're not sitting
on a lot of inventory, especially if you have things that change
perhaps then you have scrap issues. So it's like just in time production,
not even just in time. So. This line follows a lot of
lean principles. Love it. We've got single piece flow essentially
going from workstation to workstation, really, really minimizing work in progress whip
we can do just in time sort of which barrel do you need? It gets blanked and then 30 minutes
later it's a barrel from the start of a laser cut. So that's crazy.
It's a pull type system, kind of like a Toyota production system. Those are some of the lessons
that toto off the world. We use a lot of those
lessons here on this line. I love it. Yeah. Cool. So what's next? Yeah, so next is the friction,
sir welder. So we take the barrel, we unload it onto these rollers, then we pick it and we
drop it into these rollers. This is rocket forest
friction, sir welder. We actually did a little video on
this, a little bit of a deep dive here. And so if you want to check
that out, maybe we can, I dunno, maybe we can post it in the link or
something. Yeah, that'd be great. So people can do the deep dive
into the machine for sure. So this machine is unique really
because what it does is it characterizes the weld
performance as it's running. So it's equipped with a lot of
different sensors, force feedback, position feedback speeds,
feeds, light curtains, all sorts of things that our automation
engineering team has come up with. So it's kind of like it's condensing
the manufacturing of the part and the quality control of the part
exactly into one machine. So it's measuring everything. So we've actually quantified what
a failed weld looks like as well. And we put boundaries on actually
the weld performance as it's running. And so Owen and the team have actually
developed a little bit of code that maybe you can tell
us about it instead of me. Yeah. So there's a couple
different layers to it. So one is actually running on the machine
in real time and what that's doing is it's looking at where in the weld we are
and looking at each of the parameters that are collected and setting basically
downs on them. So for here we go, for any reason it goes above or below
some of our critical parameters, the operator immediately gets
a warning saying welds fail. If at the end of the well we've stayed
within our boundaries that we've proven to be a good weld, get well
passed. So it's very simple. But then after the weld is completed, there are some code that runs
but then goes to take a look, makes them plot so that we can kind of
take that data correlate to some of the physical testing that we have done
and will do on end of the barrel to continuously build up our data set. We're constantly getting
better and better data. Right now this is really quiet because
I actually have it off the work piece. Okay. It's currently not
doing it. I was about to ask, is it friction stir
welding right now? Because. It is currently not friction stir
welding. It backed off the work piece. And the reason for that is I want to
make sure that everything is perfectly aligned. So this machine controls
the position as well as the force, the RPM and everything,
a very, very smalls. And so any time any changes made, I'd just like to verify that we're
still running in Rocket three. Cause. We didn't have any of the automation
we had to test every well. So when the barrels would come
off, we had to have a coupon, we had to pull the edge of the barrel and
make sure that the well performed well enough to meet our requirements. And all the data is going back
into a giant data lake as well. So we can attach all the data for
every serial number of the rocket. So there's something that occurs
later in the integration process. We can go back and look at the
data that came off the machine. Nice. Our data lake and our data collection
system is a whole big massive piece of Astra and we should definitely
do, there's a video on that. Topic really interested honestly,
because the larger of a family of data, the bigger the pool that you have, the
more you're able to quantify things. And I always say it on streams,
more data more better. So yeah. I like it. So every single bit of data from every
sensor on every test that's ever been conducted on any piece of hardware can
be read by any engineer at Astra at any time and it's all in one giant data link. So it's an incredibly powerful platform. We've invested a lot there. I mean it shows there's just layer after
layer after layer of investment here in scaling things up, making things more reliable and getting
it so that you can produce rapids rapidly and cheaply. Exactly. I love. It. Affordably. Yeah. Affordably. Thank you. I should
use better words affordably. Cool. Thanks Owen. Next stage.
So yeah, one of my fun fact, my favorite thing that
I like to tell folks, we're in a hallway right now and this
is actually designed around being one continuous line. It's about 200 feet long. The line is so long that we had to
put a hallway in the middle of it. We debated at length a year or so ago
where we were going to put the hallway so that our employees could
walk from their desks, which are generally over
there to the kitchen. Very important that we have a path
between those two places. Yes. So we had to divide the factory up with
a hallway in the middle of it because this line is so long. So this is the Astra snack
hallway if we're going to. Name Freeman? Yeah, pretty much. Yeah.
Got it. So you see the green stripes, you do not need safety equipment here
other than the areas where we striped them off. Very cool. And when we
do a transfer across the line, we bring up safety barriers and then
we will bring parts across from one station to the next. Nice. All right. What's next? So the Ctig. Holy cow. Yeah. Wow. So this is version two of our
circumferential TIG welder. There are a couple of parts of the process
of the line that we haven't covered, we won't cover today. But essentially the circ TIG receives
barrels from the friction sir welder over there and barrels with domes already
in them from another station. And it goes ahead and attaches
those barrels we with or without domes to effectively grow a tank. So you see a first stage in the process
of being manufactured right now, you can actually kind of see one of our
domes that looks like the forward lock dome. So forward end
after end I believe. Is. There a common dome in there or is it. Common dome in here? Yes. So we use a different style of
domain manufacturing for Rocket four. I'll maybe get you to a sample there in
a sec so we can do a little quick dive into the domain manufacturing method.
But yeah, really simple machine, it incorporates a lot of lessons
that we learned during Rocket Three's production on a similar style of machine. A lot about how we fixture our weld
torch and our wire feed and how we have operator inputs that
are active during the weld potentially for EOPS
or anything like that. And then also how we tune
in our development weld. So
a lot of lessons learned. We essentially took a very,
very similar style of machine, adapted it to rocket forest diameter,
gave ourselves a local gantry, cran dedicated to transferring to a lot
of really good lessons that we learned. And it's kind of that iterative flavor
of why we've invested in these different rockets over the years to learn lessons
like this so that by the time we actually want to produce
the one that scales, we've already learned a lot of those
lessons and we don't have to learn 'em while we're also trying to get
straight into a ramp from zero. I like it. One of the things. About this rocket is at this diameter, you're approaching the limit of what you
can put in a shipping container. Makes sense. And so one of our value
propositions is mobility. So we can take the system, we can
deploy it anywhere in the world, and we can conduct a launch with a mobile
launch platform and the rocket can be integrated in a container. So actually think about the limit
to the size and the capacity of our vehicle. It's constrained by, do we want to go beyond what fits in
the shipping container? Right? Yeah. So if you go beyond what fits
in the shipping container, then you've got a lot higher
transportation cost and
logistics cost and things like that. So this allows us to basically
push the limits of the energy in the fuel to the Maximo. So once we optimize everything on this
design and the engines and everything, we can probably push about a ton
of capacity from this platform. And so this whole production line is
designed to support a rocket that has that one ton capacity in the
end of the lifecycle. I. Love it. And like you said, you're one of your value propositions is
being able to launch from basically any concrete pad that you can get access
to and have approvals to launch from, et cetera. And that's so important
for a variety of reasons, including, like you said, it's space, tech day, reconstituting critical infrastructure
in the event that becomes necessary. Yeah, it's about resiliency,
not just mobility. So if Cape Canaveral or Vandenberg or I
guess now Brownsville isn't available, well what do you do? How do
you reconstitute a capability? What if there's a hurricane of the
cape? What if there's an earthquake, a vandenburg, or what have
you? We need to account for. That. And most of the new satellites
are 500 kilograms or more. So we need a rocket that has the capacity
to be useful to reconstitute critical capabilities like the SDAs and your
Constellation that they're building. So this connects all of that. Very cool. I'm loving this. How does it feel when you stand here
and you look at an entire barrel section like this? Like a grown rocket as you say? Yeah, it's pretty freaking cool.
Seeing it all come together. Seeing the factory come together
has been amazing. Really. We love it when we have a lot of
hardware, when we're hardware rich. So it's just a really cool, it's a really, really special feeling
to see all the hard work. There's so many people here that have
invested so much time and so much effort into all these processes. They're
all qualified production proximately. We go through very rigorous
campaigns of testing. We talked about purchaser welding. We do that for all of our
kind of manufacturing methods
that are very important for the performance of the vehicle. There's an incredible amount of work
that happens to get to this point. So really it's like just, it's
the reward for all that work. If. You go back in time and you look at
the 1.0 vehicle, the 2.0 vehicle, yeah, that was a time where we
were incredibly scrappy. We didn't have any of this right. And we really had a very small team that
just threw things together that allowed us to prove out the software and the
overall approach of mobility and the overall system. As we got to Rocket
Three, it became more mature, but we realized we had to dial it up
one more notch and have a whole mission assurance organization, reliability lab, and a lot more automation in
the manufacturing process. Nice. So this represents kind
of the third swing. What is the appropriate
level of investment in the
rigor around the engineering processes and also the qualification
of all of the different manufacturing processes that ultimately come
together in the final launch system? Very cool. One more notch.
You guys took it to 11. I gets. You don't want to overdo it, but you definitely don't
want to underdo it. Yeah. Love it. So this is an entire first
stage with that piece attached. That would be the entire
first stage Tankage. That's. Great. That would be it. Yeah. I'm not sure if we're about
to get another barrel. If I had spent the time to count them, I might be able to tell you off the top
of my head what is being inserted. Yeah, it looks like we're short one barrel.
Okay, got it. Rough top of my head. So cool. And it's just immediately
obvious compared to Rocket three, how much bigger this thing is.
I, and like you said, you're, you built it so it's the biggest
you can fit in a shipping container, which is interesting because a lot of
our viewers might know that Falcon nine was made so that it's the maximum
size to be road transportable. Right. So it's interesting, like
a similar philosophy. You have the maximum size you can fit in
shipping container while still getting the most fuel in a rocket tank, but also being still mobile
and still responsive. And I think Starship is trying to
be the biggest rocket possible. So I think that'll drive your cost
per kilogram down as much as possible. We're trying to drive the
price per launch down, and so you do that by really driving all
of the trades around the overall cost of production, automation,
automation during launch operations, and also transportation. Nice. Yeah, very cool. All about you then economics. Love it. So I think that's the first time I've
ever said I love economics in my life, but I truly do. I love it. Yes,
because again, it's all in the service. Of all economics. Of making access to
space more available. So. If we want to do this for a
really long time, which we do, it means we must make a very, very
sustainable business out of it. So what will this be used for? Is this a structural test thing or what
are we looking at here when it's built and assembled? Is it more of like to qual the
machine that's building it or It's an. Actual qual tank. So they're taking this out next
week to our other test facility. They're going to fill it up with liquid, nitrogen and water and they're going
to do a whole bunch of pressure tests. So we're going to basically cycle the
tank a lot with cryo and then we're going to probably. Burst it. Well, we love test tanks here. We call our viewers Tank Watchers for
a reason because we spend a lot of time looking at test tanks and waiting
for 'em to burst or what have you. You have to make sure the big steel
balloon can hold pressure. So. Yeah, you want that thing to be as light as
possible and you want it to do the job all the way though. Yeah, makes sense. Why don't we keep going a
little bit? Yeah, absolutely. So what we're looking at
here is a formed dome. So I think we've, I'm not sure what the destination
of this dome is there. There's probably a traveler
on the other side of this, but essentially what you're looking at
is our fixture for how we insert domes into barrels. I won't give
away all of the secret sauce, but we are using a totally
different manufacturing than
we did for Rocket three. Rocket three we started
with large forgings and then
we machined those down to very thin material. You probably saw in some of you the folks
that are watching might have seen in the last video where we talked about how
we took a very heavy forging and made it a very light. You take an entire brick or bill it,
whatever you is, bill it the right term. Yeah. It was a forging in this case it started
this bill and then was forged into a profile and then you. Machine all the extra off between. 99% Of what we paid for, which is again,
like we talked about in the beginning, deeply offensive. It's offensive to you. Yes. But that's that in the previous video
is the dome that you lifted with your pinky. Yes. This is on a fixture, so I don't think you can
lift it with your pinky. It's also a little heavier.
It's bigger. It's a lot bigger. So this dome is actually stamped. So we talk a lot about how
Rocket Four is designed around manufacturing methods. The diameter was also sized
by the practical limitation
of what shape we could stamp based on the sheets we can
get formed into the dome. Nice. So this is a one single dome
that's stamped out of one sheet and the production rate of these
guys is an order of magnitude. I think it's two orders of magnitude.
Wow. Faster than rocket threes. So you're not tying up machines in the
machine shop to CNC this and you're not throwing away a
whole bunch of material. So there's multiple
levels of improvement in. Rocket three. One of these domes would
take several days to machine. Yes. Out of a large, heavy, expensive
aluminum that was forged, we can probably make several hundred of
these a day because they're just stamped out the fender of a car or. A hood of a. Car. To Chris's point, the stamping operations you can do
hundreds of times a day for overall dome production. Our output, I think
we can caps out around 20 a day, which is far more than we. Need six of those per rocket. Nice. So. Again. That supports several rockets. A
day worth of these can be produced. So in five years when we come back and
there's two or three of these rocket production lines, you have
plenty of domes to go with it. Yeah. Hopefully next time we'll
get this will be old news. Right. Our goal is for this to be old news and
you'll see other things we're working on that are clever. I love it. And again, there's a whole bunch of the factory that
we probably won't get a chance to see today. There are a lot of pieces of
hardware that we can't show on video. These are some of the simpler ones. So we've kind of tried to
focus on what we can show. No, we appreciate you guys showing
us everything that you possibly can. I know there's iar R and there's all that
sort of stuff and there's secret sauce too. So we NSF and our
viewers definitely really appreciate the openness and the
just information because how do you get people excited about space
and excited about rockets? You tell them about them.
Yeah. So thank you for that. I'm going to say goodbye.
Great to have you guys. Out really quickly. Yes. If you need a payload for
your first Rocket four, I have a sticker, so I don't think
it should affect the mass very much, but it's a Thomas Promise sticker. If you can stick that anywhere on
a rocket four, that's going to fly. Preferably the first one. I think. We have the capacity to. I think you do find some space for this,
so we'll make it happen. Cool. Well, we'll let you go. Thank you so much
Chris. Thank you for coming out. Appreciate it. Cool. So before we wrap up,
I just want to show you one, I think kind of cool area that gives
you an insight into the way that Astra works. So we've got our kind of
production development area. And so this, let me open the section up for you. We've been iterating
on the factory layout. So what you just walked through isn't
actually the first version of Rocket Forest Factory. We actually constructed the factory with
a few different big pieces of equipment locked in place. But if you were to walk around and
take a look at more of the other areas, the sub-assembly areas, you notice
that almost everything is on wheels. And when you look at the ground,
there's tape on the ground. There's not paint on the ground. So we also want to make our
factory so that it can be dynamic. As we iterate the product, we can actually tune the layout of the
factory. So what you're seeing here is a pretty clever thing. The team came up with a custom made
magnetic whiteboard that is actually the factory layout and we've got little
magnetic strips that you might see the size of a first stage tank here stuck
to the whiteboard so that we can, we've gone through several architectures. You see all the different types
of aisleway where you need ppe, where you don't what each area is.
We talked about harnessing avionic, sub assemblies, the engine
sub assemblies, clean room, which we didn't go into
today, engine final assem. We do all the way through the shop
today and we kind of walked up this aisle. So you saw this portion of
the factory up to right here. And so the remaining parts of the factory, there's a lot of work happening
in a lot of these different areas. There's development work. So. This is the snack aisle
that's very important. That's the snack aisle. Yep. That's
the snack aisle. Most important aisle. But also development is important. Also. Development is important. Yes.
Yep. Yep. So we've got mission controls, we've got our thrust structure bill. That's a pretty clever thing that
maybe we'll get to do another video on. We're very proud of that. We've got all the well booth and
stations that are all kind of roped off. We've got a lot of functional test
work here. Hardware in the loop. It's in a little building
behind us. Fairing assembly, that's all the parts of the
rocket. They're all covered here. Well, so you have a mock, which is a mock rocket that you can
do testing on all the parts together. This is a mock factory that you can test
a factory out with on a board before you go spend a bunch of time moving
stuff around, rearranging things. Dare I say it's a factory.
I'm sorry. We'll work on. That. We'll work. On that. Oh yeah. We'll we'll work.
It's an Astra rocket factory simulator. Yeah. Yeah, exactly. I mean, maybe one of the biggest lessons that
we learned as an example of something is that this main big wide aisleway that we
actually came down as we walked through and talked here, that wasn't an
original part of this layout. And we learned that as we're moving
some of these bigger pieces of rocket, we had kind of given ourselves
an aisleway that was wide enough, but we had traffic moving in multiple
directions and we have kind of stations that come off the main
aisle. We learned actually, we want a pretty wide aisle here, make
sure we can do forklift turnarounds. And so we increased the size of this
aisle, we restructure some of the layouts, we change some of the aisle, we peeled
up the tape and we put more tape down. Nice. Super quick, easy for
us to make changes like that. And we're really proud of
that style of how we execute, which I personally haven't
seen anywhere else. And I think it's really cool and
the team's done a great job. Yeah. No, I really dig the flexibility
and just the ability to constantly reassess. Are we doing this the
best way? Yes. Okay. Continue or no, let's do it. Let's figure out
a better way to do it. Yeah, like I think in the last video you
said in engineering, if you have a very complicated solution to a
problem, maybe you're asking, you're solving the wrong problem.
Yes. So this seems like a very, it's a whiteboard with the layout of a
factory on it. It's very simple. Yes. But it's a very good way to solve the
problem of how do you lay out the factory. Yeah. Simple scales, right? That's
still one of our values, right? Dang. He said it. He said, simple scales D. This is my favorite. Yeah. Love it. Cool. So I think we'll wrap here.
Maybe go get some lunch. Cool. And then if we can afterward, maybe we'll get you over by
the next thing, which we will, we'll probably just show on video. Oh, okay. Awesome. Well thank
you so much, Bryson. Yeah. Can't wait for that next thing. All right. I guess I'll just spill the beans here. The thing Bryson was talking about is
a detailed walk around we got to do of rocket four. So let's
just get right into that. Yeah, this is Rocket four. This is essentially us proving out our
production processes for all the major primary structure pieces of the
rocket. We've got the engine bay, the thrust structure, we've got
the skirt, the covers, the engines, we've got the tanks using all
production-like processes, all qualified and a couple of
secondary structures as well here. So there are a bunch of learnings that
were coming up to this rocket here, and then there are a bunch of learnings
that we're going to get from it. So we've qualified a lot of
processes for production as we go into welding or machining or anything that has very specific
tolerances to make this thing actually come together. When
we bring the pieces together, this is us proving that all of those
processes as we bring them together from cascading sub assemblies, they actually sum to a rocket
that mates together correctly. So this rocket is not match drilled or
it's not made as a suite of parts that are specifically only
fitted to each other. It's made on a bunch
of sub-assembly tools. So any sub-assembly should go
into any other sub-assembly. So we can produce this rocket more like
a car is produced than the way rockets are typically produced.
This is us proving that out. That's what this vehicle is for. So we're changing manufacturing
methods for the faring of rocket. We used to use water jet kind of
pedals and a cylindrical section that made approximated an O jive contour
at the beginning of the faring. I don't think we've released exactly
how we're making this faring for Rocket four. We might save that
for maybe a future video. But this faring is going to be
a pure O jive, a nice smooth, I mean you see it.
It's a pure O jive, nice smooth contour. So it should be a little bit
higher aerodynamic performance. And we've actually simplified
the way that we manufacture this. There was a lot of labor that went
into Rocket three s fairings and we've eliminated the large amount of
that labor going into Rocket four. So we'll maybe save that specific
technical detail for a future video, which will be cool. But suffice to say it's
different performance should
be a little bit better than Rocket Three's performance from like
a mass and structural perspective, and it'll take far fewer hours to
manufacture from a personnel standpoint. After the faring, we've got a few things that are different
In Rocket four versus Rocket three one, we now have a
payload attached fitting, which is kind of a more traditional
thing you see on a lot of other rockets. It's an adapter to adapt the outside
of the second stage. We call it the upper stage to the payload itself. So we can now we have a universal kind
of interface that we can make several different payload adapters as opposed
to emitting those directly to the upper stage. I can now with my arms show you
the length of the upper stage. It's a lot shorter than our last supper
stage, which I couldn't bear hug, so I can't bear hug around anymore,
but I can do it lengthwise. So we've gone to a command dome
architecture for rocket fours upper stage, a little different than we did for
Rocket three where we had two spherical tanks. We're no longer pressure fed.
So we're running at lower pressures, which allows us to use
these elliptical domes. And you can see some witness marks for
actually where we have welded these domes into this stage. So you
have the lock tank up here, which operates at a bit higher pressure, and then you have the common dome and
then the kerosene tank at the bottom. So we switched from the pressure fed of
rocket three to a pump fed of rocket four going for that higher
performant upper stage. And actually the failure that we had on
our last Rocket three launch was related to the fact that we chose
a pressure fed upper stage. So this actually eliminates that failure
mode in addition to having higher performance from the stage
itself. Yeah, we chose Ursa, really fantastic partner. You can see their engine featured
here attached to the upper stage. We have a lot of great
friends at Ursa. They're very, very good at designing
and building engines. They have a similarly iterative approach
to increasing performance over time, which we love. They're all backward compatible with
our designs and they've been really, really great to work with both technically
and from a business partnership. So we love working with Ursa. We're
looking forward to having many, many more engines here. Basically
seeing an inventory ready to integrate, ready to integrate on the stages. So I mentioned some of the manufacturing
processes that we validated. By building this, we've al also validating like we'll
specifically use the tanks to go through several test campaigns and this
vehicle we wanted to do fit checks of a lot of these major components coming
together to make sure that in the real world they actually came
together the way we expected. So you can see kind of the
tolerance stack of where the nozzle skirt lands with respect
to the first stage dome. That's something that's really important
as we integrate the stages together. We want to know where that nozzle is
going to need some support from the first stage. So you're looking at the inner
stage here. This is inner stage, which is not completely
representative, but we'll fly. We'll have a slightly different
interface with the upper stage there, an actual separable interface and
some separation fittings and then some stiffening structure. But from a skin and manufacturing
method of producing the skin, this is essentially what you're
looking at for Rocket four. The first stage tank architecture is
essentially the same as the upper stage tank architecture. So we
have a common dome tank. They actually share manufacturing
methods for the domes. They're a few different tweaks we make. They have different holes
in them that we cut, but they're stamped using
the exact same process. This diameter was chosen set essentially
around the raw material availability to be able to stamp those domes. So we use that diameter for both
upper stage and first stage the tanks. You'll have seen this on rocket three, very similar heritage to what
we've learned on those rockets. Friction sir welded longitudinal seams. And then TIG weld circumferential
seams. The domes are now, they're not post machined out of
forgings like they were for rocket three.
They're now stamped out of sheet. And then those sheets are then
welded into barrel sections. So much lower cost to produce
and much lower tack times. So we can ultimately scale all of these
processes up to a rocket a day should the market support that. So you're seeing essentially the most
of the rocket that you're looking at is the first stage tank. It's built the same way as the
upper stage tank command dome, just with a larger transfer
tube or down comer, depending on the word
you're familiar with using. Some of the big changes that we have
going on here are in the thrust structure. So the thrust structure now
has nodes, it has nodal points. Those nodes are castings
and then we braise a tube tube and casting braise joint. So
it's a strut system, you can't see it. It's all behind this shroud. But there is one here and you can
see evidence by one of the nodes actually poking out from
beyond the thermal protection. So we have the two engine configuration. We basically take the loads from these
two engines and transmit them through struts to these nodes that you
can see all the way around. There are six nodes and
those nodes are also, they double As how we support the
rocket on the launcher itself. Well there you have it. What an absolute treat to be able to
wander around a rocket factory and learn about it from the people operating
the machines and the people in charge. Thank you so much to Astra and CEO
o Chris Kemp, VP of manufacturing, Bryson Gentilly, and all the other employees we talked
to who not only shared their time, but also their experience with
us. If you like the video, be sure to check out our
previous factory tour of Astra, which will really blow your mind in
terms of how much things have changed. Alright, let us know what you
thought of this one in the comments. We want to do some more of these, so
hopefully you like it. And don't forget, be excellent to each other.
