Comment by bumby | Hacker News Reader

Comment by bumby | Hacker News Reader

Parent Full thread

>minimal refurbishment

Is there any information available on the activities they actually plan on doing in terms of quality checks and refurbishment? I'm a big fan of SpaceX but I also feel like people talk about this point without actually knowing what goes into aerospace refurbishment regarding the level of quality assurance typically required.

I remember back when they had a Falcon 9 failure attributed to a faulty material spec on a strut and the response was they would perform quality checks on future material.[1] As someone who worked in the industry, my initial question was "Why weren't they already doing this?" as it's a commonplace within industry on safety critical or flight hardware.

I've also been involved in rocket rebuilds and to say they need "minimal refurbishment" is a understatement. Even if none of the components actually need rebuilt, the quality inspections still take a lot of time and money. And there's always the chance you find an issue that warrants a legitimate rebuild.

So, while I'm pulling for SpaceX to succeed, my worry is that every failure or close-call like the strut issue will layer additional quality checks that gradually erode that "minimal refurbishment" claim.

[1]https://arstechnica.com/science/2015/07/spacex-says-faulty-s...

We don’t have to “refurbish” airplanes after every flight. Are the stresses of rocket engines / getting to orbit simply so great there’s no way to engineer away “refurbishment”? And will that always be the case, or could advances in materials change that?

Aircraft are on very strict maintenance schedules and need to be signed off for airworthiness in order to fly. As the plane ages, this certification is usually signed off by an certified A&P mechanic. It's regulated (in most countries) and acts as the form of quality control. It's not like a car where you can just decide to continue operating at your own risk. (This doesn't generally apply to experimental aircraft, though). So in that context, airplanes are being continuously refurbished. That's also why it's relatively inexpensive to buy a plane, but extremely expensive to own one.

Flight hardware typically has much higher quality standards. Sometimes they use more exotic materials that don't have the background testing to ensure that quality. If you want to use a cheaper new seal material, for example, there's probably not much data available for its compatibility with hypergolic fuels. It's not uncommon to use one-off components in a design, which means they may not have the same type of well-tried process control during manufacture. Then you have the costs of maintaining hardware pedigree (chain-of-custody, maintenance of material test data, lot material coupons for later testing, some of which may need to be maintained in a bonded, climate controlled room etc.).

New materials and manufacturing processes will definitely help and I think SpaceX is doing good work here. But the demands of spaceflight adds constraints. Steel and titanium are the most common non-fatiguing materials. Steel may be too heavy so you might be left with only using titanium, which is much more expensive. And you still haven't done the compatibility testing etc. So while it's a trivial example, you can see how it can become a tough problem to factor in cost, quality, and speed.

SpaceX's problem was they trusted the manufacturer who had certified the strut. So theoretically it was already being done by the vendor.

SpaceX is also pretty good about eliminating checks that have become redundant or otherwise redundant. For example they no longer do a static fire on the launch stand before every flight.

The big promise of Starship is the quantity of flights they envision. A thousand flights will give huge insight into reliability, insight that space flight has never had because most articles are never recovered and inspectes.

>So theoretically it was already being done by the vendor.

Organizations with good quality programs know how to assess their suppliers, including those who are ISO/ANSI certified. There’s a lot of information that goes into evaluating a supplier beyond just asking them if they certify their material or have a quality program, including doing site audits. Things can still fall through the cracks of course, but it reduces the probability. I bet if you asked them why they don’t evaluate each lot before the mishap, at least some would claim it’s redundant and unnecessary because the supplier already certified it.

SpaceX Falcon 9 overall has an incredibly good flight record. They very early had one supplier part they didn't correctly validate.

The other failure (non-mission) was when they discovered a real unkown-unkown about COPV tanks and helium tanks in near 0-kelvin temperatures.

So yeah, they should not have made that mistake, but its also the case that nobody has a perfect record. Considering the errors most rocket companies have made, theirs seem fairly reasonable.

I think they’re doing a good job but I don’t think that was a reasonable error. I give them slack in the areas where they’re pushing the into unknown boundaries of engineering. Regarding the COPV, I wish they would’ve added to the fundamental science by really investing in understanding and researching the failure. Unfortunately, they just changed the design and left the fundamental science to NASA, but I guess you can make the case that’s not SpaceX’s role. That strut case, however, is a well known risk within industry that they just didn’t manage very well.

Agreed, Boeing and the rest of old space use a ton more vendors than SpaceX does, and don't have problems with them the way SpaceX has.

I guess the difference is that not double checking your components is a rookie level mistake, properly vetting vendors is a sophomore level mistake.

Rocket reuse is spacex secret sauce, they don't talk about what they do or how much it costs AT_ALL it's kinda frustrating.

Reply on news.ycombinator.com