Author here. Did not expect to see this on HN at all. Just an engineering war story I shared.
I love hearing about these engineering challenges. Media loves to point to these design iterations as proof that the F-35 is over-hyped or inferior to existing jets. But what I see is innovation and trying new stuff. Sometimes failing, but in the end making an amazing jet.
I just kind of wish we lived in a world where we didn't NEED a new fighter jet and could instead invest this time and effort into peaceful pursuits.
Yeah it's amazing how it is currently the best jet in the world considering how reviled and criticized it was in the media. And said criticism had real consequences, here in Canada we basically got stuck buying 1970s trash just because the f-35 became a taboo and a meme due to projected costs, even if it means that we will pay even more for the alternatives for much much less capabilities.
To be honest, I think the F/A-18 is an excellent jet for Canada's needs. Also Canada currently has 88 F-35s on order and will get their first ones in 2026.
The ancient F/A-18s airframes Canada purchased however, were not.
Canada expects to receive its first four F-35As beginning in 2026, another six in 2027, and six more in 2028, with the full fleet to arrive in time to enable the phase out of the CF-18s by the end of 2032. But its CF-18 fleet, even bolstered by the purchase of 18 ex-Royal Australian Air Force F/A-18A-Bs, may not be able to effectively hang on until then.
They wasted a lot of time and money setting up supply chains and training pipelines for a fleet of near-end-of-life airframes that required constant maintenance, provide only the bare minimum capabilities and won't end up in service for very long. And they're having so many retention issues with their pilots that even that is wasted.
I don't mean that the airframes are any good, I just mean that Canada has a population similar to that of California and almost half the GDP. They neighbor their closest ally and are separated by ocean and arctic wastes from anyone who would want to invade them. They don't really have much expeditionary need and their defense would be backed by all of NATO.
So a couple old multi-role fighters are sort of okay for what they are doing. Mostly air to ground missions and demonstration flights. They need to up their defense spending a bit to meet NATO obligations, but not that much honestly. No one is going to kick out Canada, especially after the USA dragged them into Iraq.
I just mean that Canada has a population similar to that of California and almost half the GDP.
That makes the mistake in not buying the F-35 the first time around all the worse.
especially after the USA dragged them into Iraq
This is so not right, it's not even wrong.
F-35s are significantly cheaper than they were a few years ago, and as more are produced the prices continue to drop
The regime of steadily decreasing F-35 prices as new lots are contracted is at an end [0]. With inflation and the cost-overruns on the Tech Refresh 3 upgrade package, we are in a regime where prices will slowly increase.
[0] https://www.defensenews.com/industry/2022/11/18/f-35-costs-h...
TIL, although it does make sense, thanks!
F-35s are significantly cheaper than they were a few years ago, and as more are produced the prices continue to drop
I've heard this argument before about Canada's F-35 saga and it is now what it was then, massive copium overdose. By this logic Canada ought to wait until the F-35 is obsolete and other nations are selling airframes off for cheap.
That Canada did not have to use its air force in war during the Trudeau years does not mean that its 100% politically driven decision to shirk on the F-35 buy, then jump back on it again, was not a mistake.
i'm not saying canada not buying f35s to begin with was necessarily the correct choice, but the fact that they don't have a massive GDP and have now managed to purchase them for a lower price isn't "infinitely worse" than if they had bought them to begin with, it has lucked out in their favour
The world changed significantly since that initial order, with the threats greatly increasing (from China and Russia). F/A-18 jets might have made sense in a more peaceful world.
Now Canada has bigger problems and needs to better interoperate with NATO allies. Many countries switched their plans to F-35's after Russia invaded Ukraine.
The issue is really that they bought used, worn out airframes just because they weren't as old as ours. From another Airforce that basically deemed them to be too old and worthy of replacement by the f-35... the same f-35 that we chose to not buy instead! It's such a Canadian thing but we just basically swept the problem away for the couple of years that we can get from the airframes, but we had to basically go through all of the procurement again not even half a decade after that decision.
So there was no point at all, we could've just bought actual replacements (f-35 or not) that would last for an entire generation back in 2p15. But hey, problem solved for the current government so who cares about what happens in 10 years! But then the issue got too big and we ended up circling back to buying the exact same f-35 that we wanted to have for the past 20 years. Just with a worse deal and even more clapped out f18 and pilot accidents.
As a guy who knows the F-35 and the program pretty well, I think the best Canadian minds on the F-35 are Richard Shimooka with the Macdonald-Laurier Institute, and former CAF and F-35 test pilot Billie Flynn.
Shimooka has a number of works chronicling the Canadian F-35 decision making process, e.g. https://macdonaldlaurier.ca/assessing-damage-canadas-fighter...
Billie Flynn discussing the F-35 and the current state of the CAF on "The Merge" podcast: https://youtu.be/kibWNHr9hdg
the f-35 became a taboo and a meme
Who benefited? I assumed that while criticism is healthy, some calls for cancelling the aircraft were from adversaries. Easiest way to defeat the plane is to get Congress to kill it.
The cost of a single F35 could fund so much peace in the world. The only reason this isn't happening is because the people making sure the American people keep endlessly funding these programs have no intention whatsoever to make peace. They just don't have the intention to do so.
They intend for there to be endless war, which is what these machines produce. It is the only thing they can be used for..
I abhor war. I believe the only way to secure peace is to be very good at war. That's why I participated in flight testing the F-35, and why I work on electronic warfare simulations now.
I wish I lived in a world where there's no need for any of this, but as far as I know, war is as old as the species.
You may have crossed paths with some of my bad code. I'm sorry.
I have crossed paths with a lot of bad code. They write it faster than I can fix it.
How do you fund peace? Defense is pretty straightforward to me, but even Toms trying to give away free shoes in Africa created a lot of conflict.
the American people keep endlessly funding these programs
Fighter planes are a significant export earner for the USA - it isn't just domestic demand.
I think if you want endless war you would increase the numbers of your armed forces by several orders of magnitude.
Not spend very large sums of money on things piloted by 1-2 people that can end wars relatively quickly.
Yes, America funded Russia's recent invasions. Yup. All America.
I would love to live in a world where I can know for a fact that war will never again happen. However, the path to that world is a very long one. In the meantime I want to know that the acquisitive psychopaths that run many of the countries in the world have a very good reason to not line me up in front of a wall and shoot me.
Should we be taking steps to a more peaceful world that we aren't right now? Yes, very much so. However, unless you want to imitate the path of Tibet or Ukraine, then you better spend some money on guns and fighter jets.
This is such nonsense. Could the USA have won WWII by paying off Hitler?
But what I see is innovation and trying new stuff.
I like your positive attitude. Though I think there were some engineering shortfalls that should have been avoided with common sense.
Eg. The original hook didn't work because the shoe was angled up too high to catch the wire. The engineers designed it based on a flawed simulation model. The guys field testing took one look and knew it wouldn't work. Heck, I showed this photo to my partner (non-engineer) and the first thing she said was "it's not pointing right".
https://the-engi-nerd.github.io/posts/welcome/images/clipboa...
You can see the original (blue) vs revised (red):
https://the-engi-nerd.github.io/posts/welcome/images/clipboa...
The guys field testing took one look and knew it wouldn't work.
You have a hearsay, hindsight story (no offense to the author) that one person thought it wouldn't work.
And now we have a hindsight HN comment that they would have known it all along. I'm guessing the people who worked on it weren't idiots, though people seem to delight in supposing they are smarter than all the dumb people whose plans don't work out perfectly.
No offense taken. The observation of an instrumentation technician and an engineer (me) definitely counted for not much at all in the grand scheme of things. And we could have just as easily been proven wrong.
Yes, that's what I'm thinking.
Since you're the author: can you remember any cases where the person with "common sense" thought "this crap ain't gonna work" but it worked anyway? Surely people only remember those cases when common sense won, and selectively forget those where it didn't?
can you remember any cases where the person with "common sense" thought "this crap ain't gonna work" but it worked anyway
I have one! Totally different field though. Cruise ships (and roro ferries) look sooo ungainly in water that regular people frequently ask how do they not just roll over. The Icon of the Seas goes 9 meter underwater and 20 stories over the water. It does not feel or look right. Yet it is right, and keeps upright :), because it does not have uniform density. The engines and machinery, and tanks at the bottom of it keeps the center of gravity low enough to make it stable.
The funny twist is that vehicle carrier ships also look unstable the same way and there the intuition is more correct. There have been multiple accidents where such ships capsized. But the intuition there is still not correct about the reasons why they flip over. (It is not that they don’t have enough draft, but due to free surface effects and the cargo destabilising).
For F-35 flight test specifically, nothing comes to mind. Perhaps I'm a victim of the forgetting you mention.
My first class in Calculus based physics, my professor did an interesting thing. We would be asked to intuit the answer to problems before we did the math to know for certain. Physics is simply not intuitive.
Now, with regards to the simulation, the thing I think they failed on wasn't a lack of common sense. I think what they should have done is reproduce the results in real life using a similar jet. They relied on the model a bit too much and "Tested in production."
However, as far as mistakes go, this is a pretty small one.
There isn't a similar carrier aircraft in the inventory that could have been used, as far as I know.
Looking at the image, and knowing designers assumed the cable would rebound before being cought by the hook, the original design of the hook itself makes sense: catch the cable in the air and make sure it doesn't slip down the hook.
Obviously it doesn't work to catch cable lying flat on the ground. Which was, again, not the initial design requirement.
In another thread about Boeing, the topic of good sources to learn about real engineering came up. Well, this is a great example. Just assume the engineers designing the initial hook were not complete clueless idiots.
It's true, and it often forgets that most other aircraft go through the same teething problems.
As the article skillfully shows, there's a lot of work that goes into seemingly simple things like a hook. Other elements can be really complex to work out. The F-35's integrated power pack[1] was the source of quite a few issues if I recall correctly. But it was developments like that which allowed the plane to keep weight under control such that we now have a supersonic STOVL jet in the F-35B.
It's a pet peeve of mine when commentators say "that's stupid, they should just do <this>!" . Well, if it were so easy...
I get your sentiment regarding the need for new fighter jets. At the very least, some of these engineering developments end up helping commercial applications as well. A good example is the C-5 Galaxy, which went through torturous development. But lead to the development of the TF-39 engine, which was revolutionary in concept. It then became the CF6, which then went on to power a long line of successful airliners.
[1] https://www.defenseadvancement.com/feature/3-aircraft-system...
It's a pet peeve of mine when commentators say "that's stupid, they should just do <this>!" . Well, if it were so easy...
Why don't they just put windows in the submarine...
It's good to remind ourselves and occasionally others that if the answer to a problem in a domain we don't have much knowledge on seems simple. Chances are the people with the knowledge are well aware of your answer and know why it won't work.
Making a new experimental jet design is great. It's the committing to buy and pay for thousands of them before you've even confirmed whether that design works that I object to.
Some would argue that arming oneself is the pursuit of peace.
The program decides to officially stop trying to chase the off-center arrrestments and wire only arrestments.
What does this mean? That the F-35C can only hook correctly when it lands very close to center? And what does "wire only" mean? Aren't all arrested landings on carriers "wire only"?
The whole purpose of this series of tests was to try to exercise the arresting gear in the most punishing ways. One way that's usually done is to try to arrest far off the centerline (where the arresting force will be applied far more intensely to one side) and also to try to have the arresting hook grab the wire while the jet is still wheels above deck (this slams the aircraft down, HARD)
After this incident it was determined that we had fulfilled the intent of the test plan.
Also, instrumented aircraft capable of doing arrestments were in short supply: the program only had two of them, and we pushed one to its very limit.
After this incident it was determined that we had fulfilled the intent of the test plan.
Ok, so it was considered good enough? (This quote made it seem like the testing had failed and they were giving up: "The program decides to officially stop trying to chase the off-center arrrestments and wire only arrestments.)
Also, I still don't understand what wire-only arrestments are. Aren't all arrestments wire only?
Thanks.
I assume wire-only means no reverse thrust and no brakes.
No, aircraft land on carriers while applying full forward thrust and (I am 99% sure) no wheel brakes. The idea is that if the wire fails to catch they "bolter", i.e., do a touch-and-go, so they can come around for another landing attempt. (If they stopped or reversed thrust and the wire didn't catch, they'd end up in the drink.)
Based on other comments (or re-reading the authors comment carefully), it turns out that "wire only" mean that the wire catches before the wheels touch the ground. (This puts additional strain on the wire and airframe.)
no wheel brakes
There are brakes on the wheels (that can slow a plane moving at flying speed)? That's a lot of force. I assumed the wheels merely prevent friction between the plane body and the deck, and the engines and control surfaces, and the wire, did the braking.
All planes have wheel brakes, including naval aircraft. But in an arrested landing the wire stops the plane, not the brakes.
"no wheel brakes" here means that the brakes aren't engaged, as stated so that if the aircraft misses the wires it can touch and go without drowning the pilot and destroying an $80m aircraft
You're correct, no wheel brakes and throttle to full as soon as the wheels touch.
If the cross-deck pendant snaps, the engines don't have time to throttle up before you go over the edge. And of course if you don't catch a wire you really don't want to be trying to stop.
No reverse thrust in carrier aircraft.
I think "wire only" means the hook catches while the wheels are still off the deck.
I suppose that hard landing might have, in some ways, replicated the hard slam-down this would produce. Author, is that the case? Was the hard landing judged to have been a decent proxy for the wire-only arrestment?
Seems unlikely. One is slamming due to a heavy glideslope. Two is slamming due to a serious yank on the rear section. The airframe stresses and flight dynamics will be different.
Yeah, dynamics will be different, though caveat I am not a structures/loads engineer.
I just don’t think anyone had the risk appetite to chance a test asset against a very difficult to achieve test point.
It is also fair to assume the decision to not do additional testing of wire-only arrests was well analyzed by the respective engineering teams.
Program management does not take decisions like this by themselves.
Absolutely. We had a whole carrier suitability team full of people who lived and breathed this stuff. It was just my responsibility to make sure the aircraft instrumentation system got them the data they needed, at a high enough quality, to empower their analyses and decision making process.
Based solely on the above description -- wire-only is when you don't have wheels on deck, also slowing the craft down.
Got it, thx!
Ok, so it was considered good enough? (This quote made it seem like the testing had failed and they were giving up: "The program decides to officially stop trying to chase the off-center arrrestments and wire only arrestments.)
Kind of both: it was too dangerous to test a wider range of parameters, and the testing was therefore "successful" because it was crystal clear that going beyond the point where they had the problem would not be safe. So in this case "giving up"/stopping and "determining the limits of the landing envelope, were reached at the same time.
(BTW, the twitter link on your blog is mistakenly going to twitter.com. I think you meant to link to your account: https://twitter.com/the_engi_nerd Cheers!)
Oh, thanks for the spot. Not sure why it's doing that...
Fixed
I thought there's also the net they can use if, say, an aircraft lost tailhook
Sure but that's emergency only.
An interesting side-note on the F-35C - when it was ordered/designed there was no aircraft that could deliver replacement engines (even when disassembled) to an aircraft carrier. They wouldn't fit into the C2 Greyhound. Kind of an odd oversight.
They can fit into the CMV-22B variant of the Osprey, which is grounded for now, and I believe the CH-53K King Stallion. But they those aircraft didn't exist until recently.
edit:
I meant to say that the C-2 couldn't carry F-35 engines in particular because they don't fit, not that they couldn't carry replacement engines in general.
Replacing a whole engine at sea is something I wouldn't necessarily expect to be done often. Do they swap out wings while underway too? Having to stick a pile of spare engines in storage and replenish them when you dock doesn't seem like the worst limitation to have.
I don't know the details on how often engines get repaired, that's a good question.
The lack of ability to do it during a "Real War" means that if the engine has problems, that fighter is now sitting in the hangar useless, except for parts to cannibalize.
In terms of the F-35 engines they under-speced them, so they have to run twice as hot[0] to cool the electronics:
"The original program engine specification allocated 15 kW [kilowatts] of bleed air extraction to support system cooling requirements, and the F135 engine was designed, tested, and qualified to this specification with a level of margin available for future growth," Schmidt wrote. "During the final stages of initial aircraft development, air vehicle cooling requirements grew to exceed planned bleed air extraction."
"To provide the necessary bleed air, the engine is required to run hotter, and the program is realizing the effects of this through an increase in operating temperature, and a decrease in engine life, which is driving earlier depot inductions and an increase in lifecycle cost,"[1][2]
[0] https://www.defenseone.com/threats/2023/03/f-35s-need-more-p...
[1] https://www.twz.com/f-35-engine-running-too-hot-due-to-under...
[2]https://www.pogo.org/newsletters/the-bunker/the-bunker-hardw...
The lack of ability to do it during a "Real War" means that if the engine has problems, that fighter is now sitting in the hangar useless, except for parts to cannibalize.
If it has problems severe enough that the engine can't be repaired and must be replaced, and they don't have sufficient replacement stock, yes. But what I was trying to get at with my question about replacing the wing is that having an aircraft return to the carrier damaged severely enough that it's unflyable and unrepairable must be a somewhat routine/normal occurrence during "real war". (Of course the more you can repair while underway the better, all else being equal).
Oh you absolutely can change an engine at sea. Naval aircraft are very modular for good reason. Engines can be removed and tested (there's a whole massive test jig and setup on the fantail for it), and removed and replaced at sea.
A whole wing replacement, I'm actually not sure but it wouldn't surprise me at all.
You might not see a port for nine months or a year -- that's not an acceptable amount of time to just have a plane or multiple planes down waiting on parts that are available. There's no way to get them off the ship other than a crane pierside if they can't fly.
The entire reason for the switch from the C-2 to the MV-22 for Carrier resupply was to be able to bring an F35 engine replacement aboard.
Your statements are vague and incorrect.
First off, there are CTOL airplanes which can deliver replacement engines to aircraft, just not F-35 replacement engines (because of their large blade diameter). USN had previously used C-2 Greyhounds for these sorts of duties, but they have too small a fuselage, and were being decommissioned. There was talk of converting some decommissioned S-3B for COD, exchanging their fuselage for a wider one to accommodate the F135, but this was not pursued. https://archive.ph/20150209193642/http://www.defensenews.com...
Second, lots of helicopters can carry F-35 replacement engines, including the Boeing Sea Knight. https://en.wikipedia.org/wiki/Boeing_Vertol_CH-46_Sea_Knight I believe that the USN didn't want to depend on conventional helicopters because of their relatively short range.
I edited my comment to make it clear it was just F-35 engines that couldn't fit. But I think the rest stands true. There was no V-22 COD variant at the time and the main COD aircraft couldn't deliver replacement engines.
Other helicopters could deliver replacements to the carriers, as you said, but the carrier would have to be relatively close to shore.
Other side note: the V-22 isn't pressurized, so they have to stay at a relatively low altitude.
The CH-46 has been out of US Navy service for 20 years and Marine Corps service for almost 10, just as a program note.
I thought tailhooks predated the f35. Did they need a different design on this aircraft for some reason?
The author says the design was constrained by the space available for it when stowed in the airframe.
Yes, the tailhook bay is very, very small. We had a primary disconnect in the bay for the instrumentation wiring for our tailhook sensors. Any time we had to get at that disconnect without having the tailhook trestle removed, we would call it "proctology".
Tailhooks do predate the F-35C, but this particular airframe needs to maintain certain properties (low observability, aerobatic performance, weight, etc.). You can't simply enlarge the tail hook compartment and use the other aircraft's hook without compromising some of these properties.
Most machines don't have modular, swappable systems. For example you can't generally take the wheels of one model of car and just bolt them onto another (even if the bolt holes and centering ring line up) expecting it to go well. A tailhook is undeniably more complex than a car wheel--it's not a reasonable expectation to be able to just bolt one on from a different aircraft.
"Our existing service already has 'export to PDF' functionality. Why do we need to spend money building and testing 'export to PDF' functionality in our new service? Can't we just reuse the same code?"
Yes, the packaging geometry is pretty different on the F35C compared to other carrier-operated platforms like the F14, F/A-18, or E2C. Notably the platforms I just listed were designed from the ground up for CATOBAR operation. The F35C is just one variant of the platform, and must share certain geometries and constraints with its conventional and hovering sisters.
Since the new aircraft carriers have this new fancy electromagnetic catapults. Why don't they just use regenerative braking like the hybrid cars? They could save a lot of energy recharging those catapult accumulators.
I'm joking, of course.
Well the ship gets a little nudge in the right direction for free.
Or in the wrong direction, depending on the wind
Always sail into the wind doing flight ops.
If those hybrid cars just had nuclear reactors they wouldn't need all that complex regeneration stuff, or an IC engine even :)
Engineers, saving your program time and money out of the sheer laziness of not wanting to make a new XML format for an instrumentation project. This is how progress is made in the world, I guess.
I've worked in healthcare, fintech, and ads and this is one thing I've done in all three fields. I swear i've written or debugged XML parsers in 20 different languages at this point just so I didn't have to get consensus on a new format.
We made our XMLs with, horror of horrors, a Visual Basic script that ran in Excel and digested several input documents to generate a map template that we could then tweak by hand and turn into an XML through another VB script.
Honestly, makes sense. This is how much of finance runs their models.
We weren’t allowed to have any other real programming tools, and the telemetry “maps” we were trying to make were/are major/minor frame oriented. This maps nicely to a grid of data: a spreadsheet.
IRIG 106, Chapter 4 PCM telemetry covers what we were doing in this process, along with Chapter 9.
I feel your pain. I've written entire applications in Visual Basic in Excel onboard the CVN before. It was the only programming language I could get access to.
Awesome read! I worked on IFLOLS as a new grad software engineer during this time.
Since leaving the government to work at various software startups, I miss real world engineering like this.
What do you mean by "real world engineering"? That is, how does it differ in your perception?
Writing software to control a dynamic light plane compared to writing springboot rest apis.
I went from doing things like I describe in the blog post, to verifying and validating the most complex electronic warfare simulation the DOD has ever done, to being a developer of one of the enabling technologies of that simulation.
I believe what I do now is important, but getting an issue past test and into the release that's sent to customers isn't nearly as satisfying as "I fixed the tailhook last night, which let today's flight test happen". I miss having an aircraft that I can touch.
Just of curiosity, I would have thought that things like this are considered classified? Did you need to seek clearance to publish this? Or is there certain things that the US military isn't as fussy about (like this)?
I assume that since the plane is public and there are pictures of it landing, information regarding the tailhook doesn't have to be classified. People who want to know, will.
My own engineering war stories are just that, stories. Any technical information I gave was taken from released sources only. I am extremely conscious of this
One thing that stood out to me - based on the narrative here the tail hook never could work in real world conditions. The blog mentions that the computer model used by the manufacturer was wrong.
Does that mean that manufacturers don’t field test the hardware? If so, that is scary.
This is the field test of the hardware. When you've got a single customer anyway, it makes sense for the customer to participate in or fully drive the field test.
That's what the flight testing was for. I am not aware of a way to all-up test something as dynamic as an arrestment without actually building a jet and trying to catch a wire.
Isn't this over engineered aircraft supposed to VTOL?
There's three versions. The F-35B is the one that can take off in a short distance and land vertically[0] and it has a big lift fan behind the cockpit. The F-35A and F-35C don't have the lift fan; the C has large wings and a reinforced tailhook[1] compared to the other versions.
[0] The F-35B _can_ take off vertically, but it can't do so with any reasonable weapons or fuel load.
[1] Many non-Navy planes have tailhooks to work with emergency arresting wire systems at Air Force bases, but those are for emergencies and are rarely used, whereas the Navy uses tailhooks all day every day
Only the B, and the B is only intended to do short takeoffs not vertical ones (it is possible to take off vertically but pointless, you can't do it with a combat load).
The C is meant to do carrier takeoffs and landings. Landing on a carrier the traditional way is more reliable than trying to land vertically every time.
Great read. Looking forward to more! I was once a Harrier mechanic, and I was told very often that I'd be learning to work on the F-35Bs during my 2002-2007 enlistment, which obviously didn't happen. So, as a former mech and current engineer, I am very interested in hearing more about it's development.
I have other threads on Twitter discussing the F-35
https://twitter.com/the_engi_nerd/status/1758633498464952414 Labeling everything I could see in the cockpit > https://twitter.com/the_engi_nerd/status/1757243336941871159 a discussion of my primary job in flight test, aircraft instrumentation. > https://twitter.com/the_engi_nerd/status/1747803565987381495 riffing along with chapter one of "F35: From Concept to Cockpit", a compilation of papers written by Lockheed-Martin employees at the conclusion of F-35 system design/development.
I think every engineer has been burned by faulty test equipment. and I think every senior engineer has been burned by not trusting test equipment that IS working properly!
that was a pleasant read
TIL where "galling" comes from.
Question for anyone still browsing this far back in the timeline - any recommendations for other good engineering war-story blogs? Ideally aerospace, robotics etc.
This is really cool, thanks for sharing. What's wild to me is that the program started in the late 90s and only now is the F35 fleet up to originally specified? operational capacity.
Since then I graduated high school, got a degree, got married etc etc. The time span is mind boggling. Would be interesting to see how continuity is maintained for so long. In software it feels like if a project is more than 6 months old, we throw it out and rewrite it.
You write shit down and you have career engineers that enforce continuity
It's trendy in software to complain about doing annoying work like writing reports and documenting things. But most hard tasks require writing reports and documenting things.
And this isn't limited to aerospace. My wife has spent a career in pharma (drug save & pharmacovigilance specifically) and it's the same way there. People complain about rigidity and sluggishness in these industries but there absolutely is an ingrained attitude of documentation and process compliance that pervades. At one point -- and this was just last year -- my wife took over running a monthly safety report that involves manipulating a bunch of data in Excel. Even that has a 9 page instruction guide, and since she now owns the output she also owns maintaining the manual.
Too often in the land of software we underestimate the potential negative impact the traditional "move fast and break things" approach to product development can have when it comes to real world use in mission critical systems.
On the other side this unwillingness and mental non-acceptance of those reports/manuals/etc. as a wasteful activity frequently comes from the understanding that there are more efficient ways of doing things, and that drives the "software eating the world" effect. While I naturally don't know the details of the case you mention and pharma is far from the domains I've been in, yet in many business/enterprise situations the software approach is to code the many-page guide into business logic, including ETL-ing the data instead of manual import, etc.
Move fast and break things brings you to the Moon in a decade using primitive tech, where is total process compliance can't do that even in 50 years using much more advanced tech.
So, an amusing anecdote related to your second paragraph - one reason it's taking so long the second time around is everything has to be repeated. They lost the knowledge of how to make rocket stages and engines of that size, and had to re-learn those lessons.
It's also quite important to remember how many lives were lost (or nearly lost) because of "breaking things" in the Apollo program. Something that's not nearly as acceptable today than it was at the height of the cold war. Something that directly implies moving more slowly and being more sure that everything works the first time, every time.
Seconded. People burned alive until we learned. Surely there is a middle ground that will let us speed up while staying fairly safe, but it's important to remember that outside of software, many rules are written in blood.
I don’t hold a strong opinion either way - in terms of process and documentation versus freestyling it - but that fire was predicted, and I think the concerns were documented.
It can and did happen again, twice, on the shuttle project. Both the O rings and the ice damage were documented.
Ultimately, any process (or lack of process) can be subverted by a bad culture. And unreasonably excessive process - as perceived by the participants - can damage culture as much as not enough.
The problem is that culture is ineffable, so we try to nail it to the ground with whatever we can think of.
a lot of people died in germany, the ussr, and the us making those rockets work. and in exchange for that we planted a flag there and have a handful of rocks in a glass viewing box.
move fast and break things worked real well for the folks who got literally creamed while they were viewing the titanic.
Heck, even maintaining my computers at home requires documenting things! I have lost count of the number of hours I’ve lost trying to rediscover how or why I set things up the way I did.
“The Phoenix pay system is a payroll processing system for Canadian federal government employees, provided by IBM in June 2011 using PeopleSoft software, and run by Public Services and Procurement Canada… By July 2018, Phoenix has caused pay problems to close to 80 percent of the federal government's 290,000 public servants through underpayments, over-payments, and non-payments.“
https://en.wikipedia.org/wiki/Phoenix_pay_system
That situation was (is?) absolutely mind-boggling. I personally know government employees that were being underpaid with no recourse for months on end because the software wasn't working and the government apparently had no alternate way to pay. Some people weren't getting paid at all. And as you quoted, it affected 80% of the workforce, hundreds of thousands of people.
Ironically, many of the agile development practices which are widely used today were pioneered in the Chrysler Comprehensive Compensation (C3) payroll application. It was never able to produce an accurate payroll for Chrysler and couldn't replace the legacy system, although the project was considered at least a partial success in other ways.
https://wiki.c2.com/?ChryslerComprehensiveCompensation
Someone should get fired for choosing IBM in this case
I work on software that has a lifetime once installed of about 30 years, and if a safety critical error is found during that time, ideally it needs to fixed with a minimal patch, so we have to maintain the capability to do so.
I guess the ethos is quite different to top tech company. We don't get the pay or perks that you would get in Silicon Valley, but we are unionised, and it's a viable option to spend your entire career just on one project so it's very stable.
Partly it depends on documentation, but also on thinking long term. There are certain people who are the technical authority for a particular area, and they know that about 5 years before they retire or move on they need to find someone who can take on their role for at least the next decade, to keep their knowledge rolling forward.
That's fascinating. Is it possible for you to share more details? Industry? Tech stack?
If your project started 30 years ago, that means DOS, or Network or maybe one of the IBM behemoths?
Then the maintenance includes pacing OS updates and dependency changes?
Not OP, but I work on medical devices. One product at my last job had an expected service life of 20 years. FDA requires that the manufacturer maintain the ability to support and service a medical device for, IIRC, 5 years after market exit.
In the 10 years after release that I was on that project, we went through multiple OS upgrades from Windows NT to XP Embedded, to Windows Embedded Industry (replacement for XP Embedded) and a number of replacement x86 CPU boards had to be qualified as one manufacturer after another exited the market. Since the device is validated as a complete system, we often had to buy a year or two stockpile of existing product to give us time to start the Validation process for replacement hardware.
You usually have plenty of warning from a supplier that a product (Windows or a CPU board) is going EOL at a certain point, so you need to start validating whatever the next replacement will be well ahead of time.
I think that would be a bad way to operate, but what's worse is what we _actually_ do, which is write the project like it's gonna be replaced in 6 months and instead keep that poorly-documented untested duct-tape contraption around for a decade as the central load-bearing component of critical infrastructure.
A decade is infancy in that scenario. The world's economies are running on stuff way, way older, for example.
The F-35 contract was awarded on October 26, 2001. I was in my freshman year of undergrad, 18 years old.
I started on the program in August, 2010. I was 26 years old.
The program has just completed its Initial Operational Test & Evaluation, including its runs for score in the Joint Simulation Environment. I am 40 years old.
Why exactly did they redesign the tail hook? Surely they could have just used one off any number of other aircraft with some modification?
Or are all of those tail hooks bespoke designs because reasons?
It could be related to the fact that they didn't have much space for a normal size tailhook, as stated in the article.
I mean more the design of the hook itself, though, I don't know if that design is even atypical to be honest.
The article goes into that.
The model provided by -the manufacturer- correction NAVAIR (thanks OP!), stated that the cable will bounce up after having been hit by the landing gear. Thus the hook design made sense. The cable jumps up and over the hook. Plane arrested.
Instead, again as the article states, the cable is actually being pressed tightly against the flight deck and the elevated hook nose makes the entire hook get thrown up in the air when drawn over the tight cable, back towards the plane and would even destroy some parts of the monitoring mechanisms, so violently did that happen.
They also provide the new design, which is basically the old design and that is also why the techs that saw the new hook for the very first time (and know about the cable I presume) instantly said "That ain't gonna work!".
It's all in there.
Minor correction, the wire dynamics model was provided by Naval Air Systems Command (NAVAIR), the Navy engineering organization in control of research, development, test, evaluation, and sustainment of Navy aircraft.
Thanks! Corrected my comment.
I would actually love to know if someone on the hook design team questioned the model. I guess we won't know but I also it doesn't hurt to ask.
Like did someone go: odd, why would that cable go up and not tighten when waves are sent through it towards the outward attachments? But was inevitably shut down and didn't have "access to the customer" to ask/verify.
Like one of the first things to ask for when having to design this that comes to my mind is: I want high speed camera footage of current arrestor in action at the customer site!
Even if two different aircraft have the same space constraints for the hook (which is a pretty big if), they have different mass and deceleration characteristics (i.e. minimum and maximum approach velocity) during landing- changing the force exerted on the hook. Designing a lighter hook for the lower loaded aircraft is VERY desirable for high tech fighter jets- every ounce saved is better range, better agility, etc.
As far as the little lip at the very tip of the hook- it looks to me like the initial design was trying to minimize any risk of digging into the flight deck and causing damage- this is just a guess though.
“After the LSO finished what he had to say and left the ready room my B/N allowed that he might fly with me again. Me, I was still shaking inside.
The next morning I went up on the flight deck before flight ops started and walked to the aft edge of the deck. I was looking for something and found it.
About one foot from the end, there was a single, shiny, brand new, solitary hook imprint in the deck.”
https://thelexicans.wordpress.com/2013/09/10/one-foot/
Due to the planes and to the rest of the tailhook (the shank, etc.), they could hit at different angles, speeds, etc. That's just a guess, however.
Each plane costs ~$100 million and the entire program will cost over $1 trillion when it's done. Performance needs are extreme: They need to land in all sorts of adverse, imperfect conditions - damage to the plane, the carrier, the wire, the personnel; bad weather; bullets and missiles flying around. It seems worthwhile to design the highest-performing tailhook for this plane, rather than to save a few bucks.
Also, IME people doing something this sophisticated don't miss those really simple, obvious issues that we happen to be able to observe and grasp from the outside.
Have you seen an f14 in person? An f35?
Wildly different. For one the f14 is massive! And it's tail hook is like the size of a medium man
So yeah tail hooks vary wildy
They designed for the F-35B as the "baseline" with carrier requirements secondary. Also, the engineers knew but, "their concerns would have just as likely been ignored." This reference was 2012, when they knew it was a problem but before OP was fixing it.
https://www.f-16.net/f-35-news-article4494.html
Was there ever any consideration given to building a "testing harness" to physically simulate the F35 landing? Something like the "dead load" testing that the EMALS undergoes. Just in reverse. Anyway, that was great read.
There was a lot of static load testing done, and things like a drop test [0] of a full scale article. But to my knowledge, the only way to test the dynamics of a carrier arrestment is to actually do an arrestment. We do them on land; NAS Patuxent River and NAS Lakehurst (among others) have a full set of Mark 7 arresting gear like you would find on a Nimitz class. Lakehurst also has the advanced arresting gear present on the Ford class.
[0] https://www.youtube.com/watch?v=lGPseVNfZO0
How much of a difference is there between dry land arresting and carrier arresting? I would guess some since the carrier represents a somewhat dynamic surface, and flight conditions might likewise vary. Is there enough that a second round of carrier based testing is required that might trigger significant changes?
I don’t know if it’s significant but on the carrier the arresting system is going 25-30 mph. The ship is moving.
Again, maybe not enough to really matter, but enough to at least take into consideration.
I imagine it means that the arresting system is being tested to greater limits on the ground than at sea, all else being equalish?
Since the relative speed of the aircraft to the ship will be reduced.
That is all correct. I have flown gliders and studied physics :)
The Navy developmental test community does carrier suitability testing of every new airframe, and there's a whole program of nominal and off-nominal arrestments they have to test in order to prove the jet can recover in all expected scenarios.
All of this was done as a work up to a carrier deployment. In software terms, trying the arrestments on land is deploying to test, doing them on a carrier is production. There were three separate developmental test deployments to carriers for the F-35C. Each deployment sought to expand the understood envelope and and handling procedures. The hook redesign happened before the first deployment. The hard landing story in the post happened during the work up to the third and final deployment.
Hey, would you mind adding an RSS feed to your blog?
I am but a grunt who mostly programs radar models, I didn't know Quarto blogs could do that until just now. Yeah, sure, I'll add it.
Done!
Thank you!
Quarto made that really easy. Very cool.
Sorry, it’s been so long that I’m afraid to ask. What will you do now that it has an RSS feed?
Uh, put it in my feed reader? What else is there to do with RSS feeds?
If you're Googs, you deprecate them
Did nobody with practical experience with arrested landings look at the arresting hook design prior to this? Obviously computer models can and do predict extremely novel solutions to existing problems, but it's worth double-checking the model when someone with practical experience says "it will never work"
In this case, it seems like a simple slow-motion video of an arresting wire going under the wheels of an F-18 would have been enough to debunk the model.
Random thought: this is a case where someone's intuition matched what actually happened, making us think "why don't they listen to people with common sense?".
But what about the many other cases where someone with "common sense" said "this fucker ain't gonna work" but the thing worked as predicted by simulations? Surely they must have happened too.
My point is that when models predict counterintuitive results (which they often correctly do; See e.g. Eurisco in the Traveller TCS championship, or the shape of the F-117 compared to contemporary stealth aircraft), it's worth double-checking.
Unfortunately, my decade plus as a military aircraft tech has taught me that no, practical knowledge does not make it through the system nearly as fast as engineering "expertise".
Same, but different industry. Lots of re-engineering of the wheel after the original designers give their warnings and recommendations, but everyone's too smart to try the simple thing that already works first.
I mean... it's very likely that the answer is no. The last new carrier aircraft made was the Super Hornet - and that design was basically done by 1995 (the F-35 tests in question were in 2011/2012). That expertise would also be at McDonald Douglas/Boeing. Northrop Grumman has a long history of carrier aircraft development, but it would have been long dormant by that point.
I'm sure there's all sorts of reasons the model's inaccuracy wasn't caught before hand, but sometimes... if you're given a model that's someone says that's been V&V'd, and it produces a result that's only a little weird, you just go with it. There are only so many things you can add extra testing onto in a project. Sometimes you choose wrong.
Anyhow, consider that the model results were probably exactly what they were expecting. Remember that the designers would be honing in on the shorter tailhook. You can imagine their mental model going - "ok on legacy aircraft, we have flatter tailhooks because there's enough time for the cable to settle". And then going "ok, with a shorter tailhook, there won't be enough time to settle". And then their model comes out and say "ya, with the shorter tailhook, it won't have enough time to settle - it'll be UP IN THE AIR". Whereas reality is "ya, with the shorter tailhook, it won't have enough time to settle - it'll still be displaced DOWN".
The name "ham-peas" might be familiar. If so, how've you been keeping lately? Been a minute!
I have not forgotten your user name here, ha. Yep, been a minute. I should hit you up via email.
I wouldn't mind that at all. If you don't still happen to have the address, the one in my profile here's good too.
I actually still have your personal email address. Expect something from me soon. It took a few years but I work as a developer now (C++, electronic warfare simulator)
That's great to hear! I'll keep an eye out, and hoping all's well in the meantime.
Are you actually euler_angles, or are you really tait_bryan_angles
Deep down I am indeed tait_bryan_angles
I appreciate your candor (and your article)
Instead of a lot of modeling and testing, wasn't Northrop just allowed to inspect an F18 and measure?
The F-18 tailhook geometry is far different than the F-35 tailhook geometry. F-18 hooks are much farther back from the main landing gear, and are also much longer.
Isn’t there a normally a mechanism that lifts the wire after the landing gear has crossed it?
Yes, there are pendants that are supposed to keep the wire above the deck, but the short space between the F-35C main landing gear and the tail hook point means that there's not enough time for the pendants to raise the wire above the deck in the manner that the original (erroneous) wire dynamics model would have suggested.
Hope you have proof that the diagrams aren't export controlled. Otherwise you're going to receive some unexpected visitors soon.
They all came out of a book that anyone can buy called "F-35: From Concept to Cockpit". That book is a compilation of papers presented at an AIAA conference in 2018.
Glad you did.
First of all, thank you for the super interesting read!
Now, as a Ukrainian I do have a philosophical question of sorts. What we have seen here in a real full-scale combat is that some of the modern machines are way too delicate for actual operations on the ground. For example, I have heard some feedback about the Abrams tank: way too finicky for real use, not durable, not reliable. The same goes about many other western items. (Some hardware demonstrated exceptional reliability, like Bradleys and HIMARS)
My question is about modern fighter jets like F-35.
Does that level of engineering and the amount of delicate electronics somewhat limit the durability and reliability of the airplane compared to much simpler designs?