2024 年 6 月 5 日,星舰(Starship)第四次试飞的前一天,蒂姆·多德(Tim Dodd,频道名 Everyday Astronaut)站在得州博卡奇卡的发射台上,问马斯克对明天有几成把握。
马斯克的回答很马斯克:
胜算在我们这边,但火箭在发射台上炸掉的概率不是零。(马斯克)
不到 12 小时后,那枚起飞质量 5000 吨的大家伙在窗口开启的第一分钟点火升空。再入大气层时,直播画面里一块前襟翼被等离子体烧成了骨架,像一只被扒掉皮肉还在动的爪子。飞船最后稳稳落进印度洋,落点偏了约 6 公里。助推器溅落之后在海面上漂了一会儿,然后倒下去,从 70 米的高度砸进水里。
马斯克给这一天的评价是:最佳结果日。
这期节目分两段。前半段是发射前 12 小时的发射台参观,两个人从火箭脚下走到塔架底下,一边走一边聊设计。后半段是发射后几小时的采访,聊哪里成了、哪里没成、下次改什么。两段连起来看,能拼出星舰这套设计里那些反直觉的决定是怎么做出来的。
站在 5000 吨的火箭底下,尺度感会失灵
多德问马斯克有没有从海滩那边、从沙丘上远远看过这枚火箭。马斯克说看过,那景象疯狂。
多德自己最喜欢的是隔热瓦那一面,认为那是整枚火箭最酷的部分。两人一路走到塔架底下,多德说他拍了这么多年火箭,真正钻进底下,还是拿不到那个尺度感。周围什么都没有,没有建筑,没有参照物,一座巨大的塔架旁边立着一枚更大的火箭,孤零零杵在荒野里。
马斯克说是的,那些网格舵有汽车那么大,而且高高挂在上面,所有东西都对不上常识。
5000 吨起飞质量,人类造过的所有飞行物里最大的一个。而 SpaceX 打算用一双机械臂把它从天上接住。
这双机械臂的名字叫 Mechazilla。马斯克承认,名字和灵感有一部分来自电影《金刚大战哥斯拉》(Kong vs Godzilla),所以才有“机械哥斯拉”这个叫法。
为什么用筷子夹,而不是给火箭装腿
回收火箭的老办法是装着陆腿。猎鹰 9 就是这么干的。星舰不用。
马斯克把理由拆成了三条。着陆腿本身有重量,这是一笔死重。其次,再入时要保护这些腿,飞船尤其麻烦。第三条最容易被忽略:火箭落在哪儿,你还得把它从落点搬回发射台,这个工程量比外人想的大得多。
他的思路是反过来问:极限情况下,最省事、复用最快的架构长什么样?
答案就是让把火箭吊上发射台的那套设备,同时负责把它接住。同一套机构,两个用途,不用重复建设。
飞船怎么被夹住:腋下、凸起,还是直接抓襟翼
现在飞船的侧面有一些凸起。多德问,将来是不是就像夹着胳肢窝那样把它夹住,还是干脆直接抓住襟翼。
马斯克说这个问题争论很大,解法很多。具体选哪种,取决于飞船回来时的着陆速度范围有多宽。如果回来时动量很大,接住机构就得做得非常壮。
这里藏着一个很漂亮的权衡。如果想让飞船落得非常准,就得留一部分推进剂去做最后的修正,这些推进剂本身就是质量。于是要选:
- 静态质量,把机械臂和飞船上的抓点做得很结实
- 动态质量,多带推进剂换精度
马斯克的说法是,需要两者某种组合,才能既让机械臂接住飞船,又不至于把前襟翼从根部撕下来。
星舰在天上是跳伞运动员,不是飞机
这一段是整场参观里信息最密的部分。
马斯克说,这件事困扰他很久。低轨任务的飞船回来时货舱是空的,重的东西都在后面,重心很低。所以飞船天然想发动机朝前飞。后襟翼的任务是抵消重心和压心的偏移,本质上是在平衡压力分布和质量分布的关系。
多德接了一句,那它就是个跳伞运动员。马斯克说,完全正确。
前襟翼就完全是另一回事。如果前襟翼没有完全收起来,它就还存在迎风面积,会把火箭往后推,让它更想发动机朝前。
理想情况下,前襟翼收起来的时候要对气流不可见。(马斯克)
多德问,那是不是要尽量最大化减速。马斯克的回答点出了核心:你要的是一个可控的配平角,从高超音速一路控到超音速、跨音速、再到亚音速。在亚音速之前,真正的难题是别让发动机跑到前面去,所以要保持大约 65 到 70 度的攻角。
后襟翼展开,增加阻力,把飞船往一边推。前襟翼其实是配平襟翼,主要管滚转和偏航,不用来改俯仰角。俯仰角主要由后襟翼负责。
多德说这跟飞机正好差 90 度。马斯克说是的,就是反着来的。
还有个反直觉的细节。在高超音速、气动加热最狠的那一段,空气非常均匀稳定,没有风,没有扰动。飞船工作正常的时候,你看直播会觉得它几乎什么都没做,襟翼只是微微在动。等到进入低层大气,扰动才来,控制律也要大改。马斯克说,从高超音速到超音速再到亚音速,控制方案变化非常大。
多德感慨,这种模型太难建了。马斯克承认,最终还是要靠真飞一次才知道。
整段听下来,星舰再入的控制逻辑可以归成一句话:它靠四次襟翼在不同方向上做加减法,把自己维持在一个固定角度上往下掉。这个角度不能太大,太大就烧穿;也不能太小,太小发动机就会扎进气流里。襟翼不是刹车片,是配重片。
猎鹰 9 的遗产:几乎为零
多德问了个很多人想问的问题。如果当年造猎鹰 9 的时候,就有现在这套快速迭代、快速试错、敢扔原型的能力,会不会也用“先扔一个最小可用版本出去”的打法?
马斯克的回答先肯定后否定:
没有猎鹰 9 的积累,就造不出星舰。(马斯克)
但他紧接着补了一句,星舰和猎鹰 9 在共同的设计元素上几乎没有。多德追问,航电总该有共性吧。马斯克说,全是新的。
这句话解释了 SpaceX 内部的一种分裂感。外人看这家公司,觉得它在造两种完全不同的火箭。事实就是如此。猎鹰 9 用铝锂合金和煤油,星舰用不锈钢和甲烷;猎鹰 9 落在地上,星舰被夹在塔上;猎鹰 9 的软件栈和星舰的软件栈也没法互相移植。前者给后者提供的,主要是经验,不是零件。
二号发射塔:更高的塔,更短的臂
二号发射塔的部件已经在厂里造好了,正在往驳船上装,从佛罗里达方向运过来。
塔会更高,为了下一代星舰。但机械臂会重新设计,变短。
原因很物理。现在这双筷子非常长,转动惯量极大,加速慢,刹车也慢,动量全攒在那儿,还得承受巨大的应力。新的臂只比接住点长出几米,能快得多。多德问是不是就多出那么两三米,马斯克说差不多。
轨道发射架(OLM)完全重新设计,这回有真正的导流槽。马斯克说,现有的设计其实比他预期的扛造。第二次飞行起飞那一刻他以为要出问题,结果居然顶住了。
压紧臂也要换。现在那套有卡住火箭的风险,一旦卡住就是灾难性的。新设计在承力之后主动收回,避免夹死。
防护也升级了。混凝土掩体墙,大量钢结构包住塔架。多德提到,那堵墙是设计来在火箭整体爆炸时保住推进剂贮罐的。
马斯克补了一句,这次搞整体重设计,其实他心里打鼓,是团队特别想干。他说这话时语气是松的,像在承认自己被人推着走。
发射前 12 小时:工艺都写在隔热瓦上
多德注意到这次飞船的工艺明显精细,没有早期那种粗糙的焊接飞边。马斯克说,圆柱段现在基本是直的。早期版本像米其林轮胎人,一节一节鼓出来。
说到最早的 Mark 1 原型,马斯克笑着说了句:那个阶段,灯光很重要。需要浪漫的打光。
发射前 12 小时,马斯克对这次飞行的判断是这样的:这是有史以来最好的发射场和最好的火箭,所以成功概率最高。细节的用心程度、隔热瓦的贴合度,都上了一个台阶。
他说这次输不起的理由不只是技术。前三次飞行的数据已经证明,这套系统整体是能工作的,剩下的问题越来越局部,越来越具体。当火箭的失败点从“整个架构行不行”缩小到“某一块瓦的缝宽了几毫米”,就说明这条路已经走通了。
烧成骨架的襟翼,还在工作
发射之后几小时,两人重新坐下来。
再入段的画面里,能直接看到襟翼根部被烧穿。马斯克说,这不需要红外相机,肉眼就能看见它在发光。
钢的特性是,它会先发光很长一段时间,然后才失效。(马斯克)
他把这一飞最大的收获归给了材料:高温不锈钢合金的韧性惊人,它扛住了。
但随后他立刻踩了刹车,说不要急着下结论。多德提了电影《办公空间》(Office Space)里那个“跳进结论”的垫子,马斯克接了这个梗。他的意思是,第一直觉经常是错的,数据没分析完之前别下定论。
接下来要改的东西很清楚:把襟翼铰链区域做得极其结实。马斯克说这回要在质量上硬砸,暂时过度设计,宁可笨重。他的判断是,下一次飞行如果没能完整回来,原因不会是铰链。
其他措施包括:提高隔热瓦的安装精度、缩小瓦缝、改进填缝材料、把瓦加厚。哪一块的缝隙不对,就敲掉重贴。这活儿很烦,但必须做。
他还说了一句挺实在的话。如果当时觉得襟翼铰链那里有显而易见的问题,早就改了。当时看没问题,所以才没改。这跟事后诸葛亮是两件事。
右前襟翼:那块不幸运的瓦
出问题的是右前襟翼,而且是四块里最严重的。
马斯克说这里有容差波动,也有点随机性,像俄罗斯轮盘赌,那块瓦就是不走运。他猜右侧的瓦缝比左侧略宽,这还只是猜测。证据是飞行数据里,另外三块襟翼的展开量都正常,坏掉那一块必须拼命往外伸才有效果。
多德说那画面像电影《短路》(Short Circuit)里的机器人 Johnny 5,被砸得只剩骨架还在挣扎。
马斯克说,它在这种程度的损毁下表现得很体面。作动器还在工作,连杆没有烧掉,铰链还能转。亚音速段它几乎是全程满展开。
结果飞船保持了气动稳定,但没法精准转向,落点偏了约 6 公里,不过一直稳到落水。这个细节值得多琢磨一会儿:一块半烧毁的襟翼,靠剩下的一点点结构强度,仍然把一枚 100 多米长的飞行器稳稳带到了海面。这说明气动设计本身有很厚的余量。
飞船 30:前段整体重做
马斯克提到一个一直在推的改动:把前襟翼挪到背风面,不再放在 180 度的位置上。这样整枚火箭看起来会有点不对称,后襟翼依然在 180 度。
这个改动不会立刻出现,但飞船 30 会是第一个吃螃蟹的。它的整个前段都要重新设计,因为安装点变了,支撑结构、作动器底座全得跟着改。
团队并不反对这个方向,他们只是问了一句:这真的是现在最优先的事吗?马斯克承认,这个改动的工作量确实很大。
前段二版还会顺带改头部贮箱,整体更可靠、载荷更大、也更好造。最后一条在他眼里分量很重。他的原话大意是,容易制造本身就是了不起的改进。
32 台发动机的冗余:丢一台只掉 3% 推力
助推器这一飞几乎完美。
起飞时序里有一台发动机没点火。马斯克说那是主动关掉的,出于谨慎,可能是在启动序列里被淹了。
32 台发动机的容错空间很大。丢一台只损失大约 3% 的推力。换成别的飞行器,一台发动机可能就是 50% 甚至 100%。
着陆点火时又丢了一台,但还是精准溅落。助推器完整地漂在海面上,然后倾倒,从 70 米的高度砸下去。
多德问这是不是意味着下一次可以试着接住它。马斯克说会跟团队讨论,看有没有遗留的顾虑。只要助推返回能准确执行、着陆点火能稳定完成,他觉得应该让塔去接。
多德问要不要等二号塔建好。马斯克说不用。他们有机械臂的备件,甚至有一套更好的臂可以换上。现在那套是旧设计,太长,转不快。就算接的过程撞坏了,换一套臂的成本远低于重新造一枚助推器。
他还顺带讲了自己的管理方式。在 SpaceX,他很少直接下命令,通常是跟团队把问题过一遍,看共识在哪儿,然后往那个方向走。偶尔会强硬一次,但很少。
下一次会不会做猛禽在轨重新点火?马斯克说还没定。他认为点火能力已经验证过了,那三台转向发动机在经历极端加热之后依然点着了火。
还有个细节值得记下来。有两块被故意去掉隔热瓦的后襟翼位置,放了烧蚀材料做备份。一块是两层,一块是一层。两层的那块没有烧穿,一层的那块被侵蚀了。结论很直接:两层烧蚀可以作为兜底方案。
这条信息的分量在于,它给出了一个明确的兜底路径。如果将来某些位置的隔热瓦总是贴不好、总是容易掉落,那就干脆在下面铺两层烧蚀材料。牺牲一点质量,换一个不可能完全失效的底线。
30 天再来一次
从 SpaceX 的角度,如果不需要大改,30 天内就能再飞。要大改也就 60 天,但更可能接近 30 天。这还不算监管审批的时间。
多德提到一个更惊人的事实:这已经是连续第三次在发射窗口顶端准时起飞。第四次飞人类造过的最复杂最强大的火箭,没有拖上几天去找一个能飞的日子,没有反复排查。
马斯克说,地面有个推进剂贮罐的阀门需要调,把发射推后了一点。飞行器本身没有任何问题。
这个节奏里藏着一个不太被外界注意的能力:不是造火箭的能力,是快速检测、快速换件、快速重排发射计划的能力。一枚火箭在发射台上出问题,能在几小时内定位到某个地面阀门并且换掉,这比造火箭本身更难复制。
最后两人握了手。多德说这是在创造历史。马斯克说,占领火星。
这套打法的底层逻辑
把整期节目串起来看,星舰的每一个反直觉决定背后都是同一套算法:先问极限情况下最优解长什么样,再倒推现在该做什么。
落地回收最优解是把火箭搬上发射台的那套设备顺便接住它,所以不装着陆腿。再入姿态最优解是让飞船当跳伞运动员,所以襟翼的布局跟飞机差 90 度。迭代速度最优解是敢让原型炸,所以第四次飞行就敢在窗口顶端准点点火。
这套算法的代价也很清楚。它要求你在某些地方暂时过度设计,比如把襟翼铰链做得笨重到不可能坏。它要求你接受随机性,比如那块不走运的隔热瓦。它要求你把第一直觉当成假设而不是结论。
马斯克自己给这一飞定的调子是:钢会先发光很久才失效。这句话既是材料特性,也是这套打法的隐喻。先让它发光,看它能撑多久,然后再决定改哪儿。
还有一个反差值得记住。发射前 12 小时,他在发射台上说胜算在这边但爆炸概率不是零;发射后几小时,他在说不要急着下结论。一头一尾,两次都是在给判断留余地。这不是话术,是一个必须靠真实飞行数据做决策的人的基本纪律。
外界经常把 SpaceX 的成功理解成敢冒险。看完这两段对话,你会觉得恰恰相反。它敢冒险的是硬件,是原型,是那些可以炸掉重来的东西。它不敢冒险的是结论,是任何一个还没被数据支撑的判断。
给读者的行动清单
- 下次看星舰发射直播,别只盯着点火和落水。再入段盯两个东西:前襟翼有没有多余动作,后襟翼的展开量是否四块一致。不一致就说明某一块出了问题。
- 想知道星舰什么时候再飞,直接看 SpaceX 官网的更新和马斯克本人在 X 上的发言。二手媒体的“据悉”往往会晚几天,而且经常传错。
- 想搞明白星舰为什么长这样,去读 SpaceX 官网的 Starship 用户指南。那是最原始的一手信息,比任何解说都直接。
- 看发射直播挑有工程背景的频道。差别在于,有人会告诉你“襟翼动了”,有人会告诉你“这个展开量说明铰链已经失效”。
- 同一件事的报道,至少对比三家来源的标题用词。星舰第四次飞行的报道里,有的标题写“爆炸”,有的写“成功溅落”,说的是同一件事。
- 如果你在做工程决策,把“暂时过度设计”当成一条正经策略。前二十次迭代别抠质量,先把不可能坏这件事做出来,再谈减重。
- 把“跳进结论”这件事当敌人。马斯克在飞完之后第一反应是夸材料韧性,第二反应就是提醒别急着下结论。顺序值得抄。
- 想判断一件新技术到底走到哪一步,看它的失败点有多局部。失败点从“整体架构不行”缩到“某块瓦的缝宽了几毫米”,就说明路已经通了。
0:00hi it's me Tim DOD the everyday astronaut welcome back to Star base this is part two of our tour with Elon Musk in part one we took you around starfactory we saw what's new A lot's changed in 2 years but in this one we're going to take you out to the Launchpad and we're actually going to see the entire full stack rocket with Elon before its fourth flight test then we're going to go talk to Elon after it flies see what went right what went wrong and what to expect on future launches so part two let's get started
0:45have you ever got a chance to see it from like the beach side sorry have you ever got a chance to see it from like the beach from the dunes yeah it's pretty insane I love the heat shield side is the coolest part by far yeah it's I mean it's a damn T rocket yeah so and it's going to get TOA it's just can we get even closer [Music]
1:20yeah I don't H even when you're underneath it you don't quite get the sense of scale yeah it's just insane Jee look at that he I got to get a wrap around with you right there that is so cool I mean the scale just doesn't make sense there's nothing else out here to like compare it to you know if yeah if this is a huge Tower next to a ginormous rocket in the middle of nowhere with no buildings no like no anything so it's
2:05it's just so hard to hard to put to scale what I mean those grid fins are the size of a car yeah and they're so far up there that nothing makes sense yeah it's gigantic rocket I mean it's 5,000 tons at left off it's the largest flying object of ever made of any kind and you're trying to catch it yeah catch it with me Zilla arms I was actually slightly inspired by Kong versus Godzilla yeah you know mechazilla yeah yeah yeah yeah that's the reason for this well the the reason yeah the reason for the like I said there's there's there's many ways to solve this problem obviously you could do it with Landing legs the problem with Landing legs is you if you well you got the weight of
2:54the landing legs and you got to protect them on Entry especially for the ship that's tricky yeah and and then you've got to get a giant rocket from where it landed back onto the P onto the launch stand yeah so I was trying to think of what is the in the limit the most efficient most rapidly reusable architecture would be that the thing that lifts the rocket onto the pad is what catches it what catches it yeah so how is there still plans like how will the ship actually get caught are there similar Nubs on the ship as there are there are currently Nubs but they will be yeah okay like deploy catch on the armpits you know yeah I like do you just grab it by the
3:34flaps or is there the the Nubs then some this is a matter of much debate but again there's like a lot of ways to solve this yeah it it kind of depends on what range of kind of Landing velocity is they going to deal with you know is it because if it if it if it comes in at like twice of it's you know if if if it comes in with with a lot of momentum mhm you you have to have much stronger catchers than if it doesn't right now the downside if if it if you if you try to reserve propellant such that it is a very precise Landing you're burning propellant to achieve that Precision so that's also so which one is better right that's also Mass yes yeah so do you want
4:23static mass or dynamic mass in the form of propellant you know there some combination of the two is needed in order to have the ship get caught by the arms and and not have the Ford flaps sheared off when it comes down right are the flaps there talks the flaps kind of having a little bit of a redesign too like maybe moving a little bit is that still there's actually
4:54if you understand what's trying to what the ship's trying to achieve it's actually bugged me for a long time the for at least for the the lowth orbit missions it's it's coming back with an empty cargo bay so the ship naturally wants to come in engines first yeah cuz that's the heavy part it's the center mass is is low y so now the rear flaps count counteract the center of mass the center of pressure so you're really balancing pressure and and mass it's like it's basically a seaoil yeah yeah so now Ford flaps that are not fully stowed are actually hurting you because Ford flaps are if you have got a maximally stowed position of the Ford flaps that means you still have some
5:39frontal area that is pushing the the rocket backwards and making it want to go engines first right right so whereas frontal mass in the rear is helpful you know basically frontal area in the rear yeah is helpful y frontal area in the front is counterproductive counterproductive so you you really ideally want the flaps to be in the stowed position in invisible to the flow so they're not trying to tip the rocket backwards but don't you I guess it's not like you're trying to maximize braking necessarily you're trying to achieve a certain glide ratio and things like that you've got to control it through Hypersonic Subs Supersonic transonic and subsonic yeah
6:21and at least until you get to subsonic the the rocket really is the the challenge is not having the engines go for not having the engines go go you know yeah basically you want you're trying to maintain an angle of attack that's cool around 65 70° like so it's if this is this so if this is the Earth it's it's it's doing roughly this angle yeah and so the flat the rear flaps the F flaps are out as much as possible to try to keep the engines from scooting forward yeah so you have the rear flaps if the rear flaps deploy that that's that that's that's that's how you know that that that increases the drag and pushes the the ship this way when the rear flaps deploy when when
7:09the when the front flaps deploy it actually pushes the ship back yeah so what the front flaps really are do they're really trim flaps right right they're not trying to change the the the pitch angle they're trying to maintain it almost well that you need to control the ship in your en roll mhm in addition to pitch yeah so the but
7:34the work of of of of you know essentially controlling pitch is really the the rear flaps mhm and and they're they're trying to move the centeral pressure such that the the the engines don't go don't go into the flow like so ship's gone like this re fls deploy it put it makes it want to tip forward which is good it counteracts the at the rear yep yep the the the flaps at the front are really there for trimming your role and Y
8:11primarily they all and they all interact with each other too which is yeah exactly that's why it's like is technically the front flaps also do some Pitch action right but but but but but for for for setting your your your fundamental pitch angle it's the rear flaps right and you want your front flaps to be when they're when they're back to to be invisible to the flow yeah so that your rear flaps don't have to do extra work right right that makes sense yeah so but you also need some margin though too so having those be able to dip out and do a few things to maintain you got to control your en roll so you really kind of need like three control
8:51surfaces and we have four so slightly more than you need yeah but and and and in the high heating regime which is where this really matters the the the air is very stable you don't have wind really it's very air is homogeneous right right so you when when the if the ship is operating properly it it look like it's barely doing anything it's just the flaps are barely moving and it's just maintaining this angle yeah cuz it's not having to react all these turbulences and all this stuff there's no disturbances huh I never thought about that until it gets in the lower atmosphere obviously yeah and then the whole thing is there almost like a you know when you're going
9:33from Hypersonic and Supersonic and is there like do I obviously the control schemes almost Chang a little bit to a certain degree yeah yeah and it changes a lot to subsonic yeah big time so that that's probably pretty crazy to have to model that and like get some sense of where that's going now but at the end of the day you're still going to have to see how it happens when it actually gets to that regime yeah once you get to subsonic the ship is falling like a sky dier yes like we've seen yeah so you're the flaps are kind of like arms and legs of a sky diver so it's counterintuitive it's not at all how an airplane works right airplanes don't P
10:08yeah it's it's like 90° to what an airplane would do basically yeah yeah it's 90 Dees to what an airplane would do it's it's really analogous to a giant sky diver yeah I mean exactly cuz that's exactly how a sky PR D close maintains their orientation yeah that is just the craziest thing when I I got to ask so you know you have the the ability now as as with SpaceX being you know where you're at to be able to do iterative design like this if you could have done that on Falcon 9 had a total blank slate and literally just minimum viable product we're going to throw that thing out would you have done that on Falcon 9 because Falcon 9 people forget like you know they look at
10:48this and they go oh they don't even know what they're doing well you guys know how to build a rocket and launch It For the First Time Falcon 9 Falcon heavy I mean you've already done that in hindsight saying we don't know what we're doing and what have they done oh exactly that's that's seem to be an armchair Rockets engine yeah exactly but they they don't understand the you know the the test PL the prototyping the testing phase so if you could have done something like this with Falcon 9 in the early days would you have done that if you could have or we wouldn't we wouldn't be able to make Starship without the benefit of falcon9 right so it's sort of like
11:21saying if you had a time machine or something you get into like time machine logic Loops right so I mean this this is really there's nothing in common with like we learned a lot from falcon9 but there's nothing in common that there's no design element that sarop has in common with falcon9 nothing I mean I I take the back you got Gins in both cases for for the you know that maybe some common is avionics or something no old new avionics really yeah jeez that's insane to think about what AIC Sol the software well new and Tower 2 is
12:04starting to head this way from Florida it's looking like and that's going to go over that way is that right yeah and Tower 2 will be have a number of upgrades it'll be taller to allow for the next generation of of Starship y so we're going to add a few sections to that and and then the redesigned
12:23arms will be shorter so if you look at those arms up there the those arms have very high moment of inertia so that you know they're they're very long right so you want you actually when you have long arms it's hard to move them fast right and because you have to move them and break yeah and so it's just a lot of momentum yeah and a lot of Ding and yeah so the the new arms will be able to move much faster and they'll be a bit shorter than those arms basically right at the catch point right like there's pretty not too much beyond that yeah so they'll they'll be you know Fair bit short maybe get like a couple meters beyond the cash point well the is there
13:07any kind of redesign of the O olm and the the the booster of a day and all the all the the flame trench and stuff like that is yeah that's it's a total redesign which I don't know if we should do a total redesign but the guys really want to do it so I was like okay like it'll have like more of a flame trench this time yeah this this has turned out better than I expected though I had to say yeah on flight 2 when it took off and it's was like hey look it actually it actually worked yeah it blow up so but the like the ring and stuff is anything changing is it is it almost like a whole everything new or is it pretty much wow yeah this just crazy the the hold
13:49down arms also change well it's not the really the launch arms cuz the current ones have some potential to bind and grab the rocket which would be real bad yeah so we have a new design for the launch launch arms that take the load of the rocket and then retract so that they're not going to bind and potentially cause the whole thing to explode yeah that'd be bad especially at this stage it's getting a lot more a lot more legitimate and you know yeah we're trying to protect against the explosions like got basically bunker like concrete walls yeah a lot of Steel protecting the launch Tower so gosh it's it's designed to like that wall is designed to protect the
14:36propelling tanks against a full explosion of the rocket yeah so Tower two that's on its way like they're literally loading up barges and you have sections out at the at the factory and everything so Tower two is like already I mean it's just crazy to think how quickly it's all and that'll be you're going to be building that here all year probably and wouldn't be surprised if that thing ends up being used already by next year can't even keep up with this that is ridiculous wow tomorrow after this thing launches would you want to do a quick little like rund on on how you feel about it depends on how it went let's say it makes it through re-entry both of them sure okay
15:23well we better Shake on that then huh that'll be awesome I I mean tomorrow's going to be phenomenal I feel how how are you feeling about it now that we've now that you've had time to see it all and you know looking at the data I think the odds are on ourside but there's still you know some chance that the rocket blows up on lifto it's not zero right I this is the best launch site and the best rocket we've we've ever made on saw ship so it has the highest odds of success and you can even tell like the the amount of attention to detail on things on the you know on the ship and the the tiles and everything it's it's much more finesse yes yeah it's it's not
16:00this doesn't have like grony bits exactly oh man it's going to be exciting can't wait for it it's going to be I hope people pay attention to this because it's it's history in the making you know yeah we're watching and the cylinders are now pretty straight you know early versions had were like kind of Michelin Man situation especially like Mark one like you were talking about earlier like Mark was let's just say lighting was important Mark yeah that is awesome lighting was important oh man you need like Ro romantic lighting for Mark 1 well I really really appreciate you taking time you know T-minus oh God we're within like 12 hours now so T-minus 12 hours and hopefully we'll
16:47see this thing lift off tomorrow again yeah it's wild all right Co it's wild thank you so much Elon thank you 2 1 [Music]
17:18the next morning that Starship stack with booster 11 and ship 29 did indeed launch for starship's fourth full stack flight our everyday astronaut Cosmic perspective Crews managed to get some amazing slowmo shots but only for the first few seconds due to cloud cover we can't wait for the next launch so we hopefully get some incredible slowmo shots of the booster returning and hopefully a catch attempt as well that will be crazy and of course we'll be covering it live in 4k when it happens but the good news is this test was the biggest success of the program yet with both the booster and the Starship upper stage re-entering and softly splashing
17:55down which was great news for us because Elon kept his promise and followed up after this historic test so just a few hours after the launch I met back up with Elon to discuss what went right what went wrong and what to expect in the next launch but before we hear from him I wanted to point out touring Starbase has highlighted the challenges of accessing direct reliable information about space flight and the Aerospace industry without being overly sensationalized and today's sponsor embodies these principles perfectly ground news is an app and website that gathers news articles from more than 50,000 sources globally and puts them in one place so you can compare how
18:30different Outlets cover the same story fun fact it was founded by former NASA engineer Harlen K who worked on the James web Space Telescope and New Horizons let's take a look at how that Starship launch was covered in the media there were over 160 articles and we can see visually where the biases lean on the bias distribution chart with 26% leaning left 62% in the center and 12% on the right there's also the factuality chart and the ownership chart which shows about 50% of these articles were written WR by media conglomerates these are all things that can influence bias and framing then we can scroll down and see every article about flight 4 and
19:06compare the headlines I think it's fun to see what words are used in each headline which ones mention exploding or test or Elon Musk I personally enjoy scrolling through my blind spot feed which helps me stay informed and balanced on topics and gives me perspectives that I might have previously missed Ground news has been important to me and helps me enhance my knowledge and critical thinking about space science and World issues so subscribe through my Link at ground.
19:30news/ everyday astronaut for 40% off the ground news unlimited access Vantage plan this month that's the plan that I personally use and I highly recommend it to anyone who wants to stay well informed and gain a deeper understanding of the news there's also a link in the description too okay now let's hear from Elon about how he thinks the launch went that was come on I mean sick that was amazing [Music] saw it you saw the whole thing I did predict that would happen although F the question why don't we do more about it like we did do a lot oh my gosh that was
20:12that was insane yeah yeah do you feel do you feel pretty satisfied with the the results of yeah it's awesome this is a really I'd say best outcome day really it is I mean yeah did as you were watching the grid fins start to burn through or I not the grid fins with the flaps what I mean you had to have thought one of those was just going to rip off right yeah I mean the the the heating you can see the heating from the internal cameras yeah you don't and you don't need infrared cameras you can actually just see it Glow right so the thing about steel is like it's going to Glow for a long time before it fails I think the you know one of the biggest lessons to say is the
20:49incredible resilience of a high temp stainless steel alloy right because it it hung in there yeah and so you know you always want to be careful about jumping to conclusions right you know what like the we need to jump to conclusions Matt from you know offer space right right yeah exactly jump to conclusions so we have to be a little hesitant about reaching conclusions before we fully analyze the data because sometimes your first guess is they're wrong right but
21:24the we we obviously need to make the the hinge we we need to make the the flap hinge area far more robust so we we'll go we we'll really go crazy hardcore on that and that hopefully I don't eat my words but that that will not be the reason like if if if the next flight doesn't doesn't make it back fully intact that will not be the reason hinge flat will not be the reason we're we're going to just spend a bunch of mass right just make it super robust over overdo it for now yeah yeah I mean there just things you can do you know to incre increase the tile Precision increase decrease the Gap improve the Gap filler thicken the tiles if if
22:15any Gap is is not perfect then then then you got to crack the tile pull it off do another one it's it's kind of a pain in the end right right so like it's not like we thought that like if we thought there was something obviously we done about the the the hinch FL flap hinch area we would have done it right but like I said we'll just go to town you know probably you know thicker tiles
22:45thinner gaps better Gap filler maybe even tightening the tolerances there's a whole bunch of things that we we can and we'll do for the next flight i' mentioned yesterday that we have an improvement improved design of Starship coming down the pike not immediately but it's it's a it's something that I've I've actually been kind of harping on the the you know just saying like guys we need to do this for for for a while which is to move the Ford flaps Leeward so that they're they're not at the kind of 180 degrees they they're moved kind of kind of leward yeah so it it'll look a little bit asymmetric with the rear flaps the rear flaps are where they should be at
23:27180 degrees so I mean those those aren't improvements you'll see already on like ship 30 because ship 30 it's it's a it's a total redesign of the whole front section because you go redesign the mount points the reason like it's not like the team disagreed with me with about this they did agree but they were like is that really the priority right so because you've got to redesign all of the support structure you know the actuator Mount it's it's a a huge redesign of the front end right and there will be a number of other sort of improvements we do you know to the header tanks and the whole front section and it's kind of like front front section front section version two
24:08which would be much better yeah so more reliable more payload you know and easy also easier to manufacture which is a big deal do you know did the AFT flaps and the four like did all of them kind of have similar burn through do you know or it no the the the the right forward flab was the one that had the I I you know because there's very this this this tolerance variations there's a little bit of just stochastic you know luck right you know you know it's like there's a bit of Russian Roulette going on here you know which was The Unlucky tile but it it looks like there there was probably I'm just guessing here this speculative that we had probably the gaps on the the
25:00the tiles were a little wider on the right side than the left side because it was really the the the right front flap that by far had the biggest issues okay and we can see this in terms of when the when the vehicle is trying to control itself how much deployment does each does the each flap have right so the the other three flaps had more or less normal deployment but the the the broken flap had to really go all out and stick its like skeleton hand like really far out to have an effect it reminded me of like short circuit like Johnny 5 and part two I don't remember if you remember that movie short circuit Johnny 5 is alive like he's a long time
25:38ago long 80s movie and he gets like all beaten up and he's like just this like skeleton all of his you know and he's like still fighting along it just reminded me that like seeing this like claw of a of a flap I couldn't believe how little it was needing to move though and was still I was moving a lot at the end was it oh yeah yeah it was it was going to like full deployment especially subsong M so but it performed admirably for its sort of insane you know level of Destruction and I can't believe it still could move when all of it looked like it you know like you think I me me neither I'm like holy is that thing still moving but the actuator was still
26:14working act the the ver you know the actuator linkages were did not burn away the hinges were were still working it's like damn okay that's insane insane so I couldn't believe that yeah yeah totally so now it it was able to maintain Aeros stability subsonically but it was it but because of the of the Ragged a flap it was not able to steer itself to the exact right position so it was about 6 km off positionally okay but it did manage to maintain stability to all the way to the ground does this make you like I we talked a lot about ship but booster obviously did phenomenally well too was that on target that it actually the booster was was almost perfect I mean
27:04you saw obviously there was one engine that that didn't didn't light or or that we it didn't explode but that that that we thought was we should not turn on basically it I think it'll bought it during the start sequence out of concern but 32 engines is fine it's it's resilient to multiple engines being lost and still completing its Mission yeah like engine lost is only 3% of thrust so exactly it's pretty minuscule compared to another vehicle that could be 50% or 100% even but and then on on Landing burn it it seemed to lose another engine as it was lighting it for landing burn but it still looked like it did a pretty precise touchdown it looks
27:43like and yeah yeah no it landed in the water yeah intact on the water it was literally bobbing in the water and then obviously it tips over and you know when when it tips over it it's it smacks down pretty hard cuz it's like pulling from 70 M like over 200 ft yeah so does that I know you mentioned on X that you were planning to actually try the boot you feel like you might want to try do the doing the booster catch next is that yeah we'll talk it over with the team see if there's any lingering concerns but at the point of which we can do an accurate boost back and we can execute the The Landing bone that which we' we've shown then I think we
28:22should try to have it have the tower catch it wow and you don't feel like you need to have Tower two ready ready before you do that you no not necessarily because we do have replacement parts for the for the for the arms MH and and there're actually a better set of arms we could put on there right those are kind of old design I mentioned that they're those arms are actually too long so the of no are too high so it's just you want actually shorter arms that can move faster yeah so my inclination you know but I mean the way the things work at at at SpaceX is not I I don't it's it's rare for me to sort of drop an imperial decree once in a while I will but it's pretty rare yeah
29:11usually it's like we'll to it over the team see see what the general consensuses and and move in that direction yeah that's going to be wild so so hopefully next though we'll are we going to kind of see more or less a repeat you think again and and of this Mission profile maybe bring the booster back maybe do it would you may do a raptor relight this time to kind of prove out orbital capabilities in Coast phase and yeah that's a good question I haven't we haven't decided are we going to do like clearly rapor relight does work because we had the three three relight yeah the three the three steering engines which would be the ones that would relight relit even after
29:47all that drama coming through crazy heating yeah and of the two tiles that we put in the rear where we we took the tiles off we actually put some backup ablative there one was two layers of ablative one was one layer of ablative the two layer of ablative did not burn through the one layer of a blade of did but I'm not sure if it actually Burt through the steel too but it did erode the single layer of a blade of but it just says that if you have two layers of our blade it does not actually burn through so that could be missing time backup that you're kind of talking about that's so so worst case scenario next time which will probably knowing you guys be in a couple months
30:26already yeah we're depending on how many on what changes we we make I think from a SpaceX standpoint I think we're we're ready within a 30 days yeah if if if we need to make a bunch of changes I'd say from SpaceX standpoint not counting out regulatory we're ready probably in less than 30 days if there are bigger changes I'd say not Les we're ready no later than 60 days but but probably closer to 30 and that's that's where people are you know going to be shocked that you're seeing these vehicles just continue to increase Cadence I mean this is the third flight in a row where you lifted off the most you know complicated vehicle ever made right at the at the top of the window
31:11you know not scrubbing for days and days trying to figure out how to launch just like the countdown was so smooth hit t0 we did have a a a ground valve propelling tank valve on the ground that needed to be tweaked but that's what p pushed it a little bit but the vehicle itself had no issues that's still incredible I mean on like the fourth flight ever of the largest most powerful rocket ever built I mean that's insane and it's just going to keep going so yeah we're really excited I know you guys have a lot to keep going here so I I really appreciate that you did follow up with us this is abely yeah yeah that's huge occupy Mars man yeah you're making progress yeah
31:51congratulations again thank you so much for your time and you can't wait for the next one all right cool awesome thanks Elon thanks again so much Elon and everyone here at SpaceX who made this tour possible huge thank you this is just incredible we're so excited for what's going on we're going to be out here for the next flight you better believe it so stay tuned there's a lot more excitement to come and I also owe a huge thank you to our patreon supporters if you want to help support the work we do here on everyday astronut head on over to patreon.com Everyday astronaut and another quick shout out to our sponsor ground news check them out at ground.
32:20news/ everyday astronaut for 40% off the ground news unlimited access Vantage plan this month and while you're online be sure and check out our incredible merch store for shirts like this our orbital shirt we've got our heat shield tile collection Falcon 9 model rockets lots of other incredible stuff over at everyday astronaut. comom shop thanks everybody that's going to do it for me I'm Tim Dodd the everyday astronaut bringing space down to earth for everyday people [Music]
资源与出处
- 视频在 YouTube 收看
- 档案silicon-valley.tech 档案页本页视频与文稿数据来自 Elon Musk Archive
- 精读本篇精读 Markdown 版