The B-2 Bomber: Those Who Forget History Are Doomed To Reverse Engineer It

The Drive had an interesting post recently, about someone noticed a procurement from the U. S. Air Force to reverse engineer the B-2 bomber’s Load Heat Exchanger (whatever that is). You’d think if the Air Force wanted to reverse engineer something, they’d be looking at another country’s aircraft. What can this mean?

Presumably, the original plans for the system have been lost, or maybe the company who made them is long gone and the tooling to create new ones along with it. Then again, maybe the assembly needs parts that you can no longer get. The Drive has another interesting speculation: perhaps the plans were so secret that were accidentally destroyed.

You don’t hear much about the B-2. There are only 20 left of the 21 built, at least that we know about. Original plans in the 1980s called for 132, but the end of the Cold War spelled the end for the stealth bomber. They get an overhaul every nine years. The Drive also speculates that this may be part of the Air Force’s desire to digitize spare parts and use 3D printing, but — honestly — it doesn’t sound that way to us. Especially since the fleet will retire no later than 2032, so whatever is replaced is only needed for a decade.

If you think you want to have a go, here’s the help wanted ad from the Air Force. If you read the text, it’s pretty clear they have some defective units that need replacement and it sounds like no one knows how to do it with existing materials. Not many of us get to design things that are still working nearly three decades later. Keeping a supply of parts and even know-how for something built in the 1990s isn’t trivial. Something to think about if you design something with a long service life.

The B-2 is a stealth bomber and while one did crash, it wasn’t shot down. The F-117A — the stealth fighter — was shot down against all odds, though. While the B-2 appears to be quite a plane, we prefer our bombers a little bit older. Still, you might enjoy the video below about the B-2’s chief engineer, although he doesn’t mention the Load Heat Exchanger.

82 thoughts on “The B-2 Bomber: Those Who Forget History Are Doomed To Reverse Engineer It

  1. “Presumably, the original plans for the system have been lost, or maybe the company who made them is long gone and the tooling to create new ones along with it.”

    Or maybe they no longer have a tape reader, or the person who understood EBCDIC is no longer with us? :-p

      1. > reader the same life-span as the media

        Are you referring to the drive or the human reader? You can argue that to read old technical documentation might require some technical cultural understanding that can only occur if the organization has a constant continuation of engineers with enough time to onboard the newbies over time.

    1. “Or maybe they no longer have a tape reader, or the person who understood EBCDIC is no longer with us?” I find both of those hard to believe, as lots of people (myself included) are still fine with EBCDIC, and I can still read the tapes I made at work in the 80s :-). I agree tapes from the 60’s might be a problem.

      I suspect that they either never had them (external company) or lost them…

      And the key to all this stuff is to use formats that will last. I can still read my emails from the late 70’s as they are in plain txt, and bmp files will be able to be read till the sun goes nova.. JPG and ZIP will be able to be read for at least my, if not my grand kids, lifetime – as the read (as distinct from compress) algorithm is straight forward in both cases, ie if every bit of software on the planet that did zip or jpg disappeared I could write something to process them from memory let alone if the spec standards were still available. The just aren’t that hard to decompress.

      It’s the proprietary formats that are the killers – and they should be avoided like the plague..

  2. “The Drive has another interesting speculation: perhaps the plans were so secret that were accidentally destroyed.”

    Or the other derp derp of highly classified compartmented information, they’re still on file somewhere, but nobody still serving, or alive has been read-in to that part of the project and is able to retrieve them. Or the people with blanket authority have retired or died, and therefore nobody knows where to look for them, because they don’t officially exist. Practically the same result, they may as well have been destroyed.

    1. Our bosses bought new phone systems, kept the manuals for themselves and sold the company.
      I did a search on the web and couldn’t find the manuals and no one knew the password.
      I found one manual but everyone throws stuff away and everyone else at the company suffers.
      We’re missing power supplies because the higher ups throw stuff away.

  3. I worked on the project. Hang on while I check my hard drives. Just kidding about some of that.
    It doesn’t surprise me. It was a paperless factory that printed thousands of pages per employee using laser printers fed with 1-ton rolls of paper.

    1. I concur, although the sub-contractors could have destroyed the fixturing and setup equipment required to produce the product and to make more, the cost of reverse engineering to something new is cheaper than the millions to invest in some machine shop fixtures. (same with apollo, data is there but the original build contracts included for fixtures and machine tools, very expensive and involved to reproduce)

        1. Its a money grab by contractors. There is no way that the plans were lost for a system that is not that old in military terms. It is also pretty bad since I would assume that such plans are probably classified. If they are missing that would constitute a pretty severe security breach. I could see this if the prime contractor for a project no longer existed but that is not the case here.

          1. There are explicit retention requirements for both classified data and government property that the contractors do follow. The retention requirements typically expire after ten years and the government decides whather they want to extend the retention date and pay the storage and maintenance cost.

            Even retaining and maintaining all the old equipment does not always mean that it can be used to make new parts. I’ve seen this similar problem with power supply transformers. The original manufacturer still had the fixtures to build the custom 1987 transformers, but they were no longer compatible with modern factory line equipment used to wind and pot the core.

    2. Hahahhahah… No it Milspec documentation really isn’t copious. They often don’t buy the drawings or CAD models, so if the OEM doesn’t want to help you fix it at a sane price you have to reverse engineer the part.

      1. You’d have to read the contract to know what level of data rights were purchased. If the government didn’t purchase unlimited data rights, they wouldn’t have enough information to make new units even if there was no data lost or unreadable.

    3. Even back then, it would be rare to buy the engineering for a component like this (whatever it is). It likely has a specification and an interface drawing for integration purposes. There was a push to save money in the last few years before I left in the mid-’90s, and one thing you had to work to justify was delivery of documents. I was involved in reverse engineering and qualifying a lot of components. Usually the supplier would do the reverse engineering themselves in the hopes of future business.

  4. I can still remember details of projects I worked on in the 90s. None of them are important, but if the documentation is lost there are two things in my favour. Firstly, I did it, so I know it is possible and I am capable. Secondly, if I had to do it again I have better tools, equipment, and materials available to me now.

    Of course, if I was the guy that worked on this plane and I had died or something then that wouldn’t help.

    1. You always show up with an unreadable name (unicode, I assume?) and give cryptic comments that intelligently lead only curious people to something plausibly related.

      You should write a book. Then never publish it. And hide 6 copies in obscure places for people to find.

      I have no idea why.

      1. Funny you should mention that, once I dropped a hint about a new technology on a forum that was frequented by well connected intellectual property thieves because I wanted them to act in a way, out of greed, that resulted in the enabling technologies for the development of something else that I actually wanted to progress to a commercial product.

      1. Heat exchange is part of stealth for a machine. Thermal, radio, acoustics, everything must be accounted for a minimized to avoid as many detection systems as possible. The stealth engineer will very likely understand heat exchange as well or better than an HVAC engineer.

  5. Actually this more common than you might expect. While there are detailed plans for things like the space shuttle, no one knows how they were actually made to work by the guys/gals who actually made them work. The notes about “filed this”, “tapped that”, “this goes first” were lost as the people aged and facilities closed.

    1. I’m still trying to find the “Space Shuttle Garage” documentary about the servicing they went through between flights. It was only broadcast one time, a few months before the loss of Columbia. Don’t recall if it was on History Channel or Discovery but it was one of those. I didn’t get to see all of it but expected it to be shown again many times like they do with most of their shows. But since it showed how the procedures and processes were frozen circa 1980-ish, including a crew failing to get either of the special wheel nut torque wrenches to work, it’s apparently been tossed down the memory hole.

      Just wouldn’t do to repeatedly show how outdated and problem prone the Shuttle servicing was after another Shuttle was lost in dramatic fashion.

  6. 3D scan the existing good/bad parts with Neutron imaging, and then start to workout what you would need to try and do to recreate the parts. The very first step is in understanding what you have.

    1. You would only get the dimensions of the parts out of the limited samples, but not the allowable tolerances in the design. Too tight a tolerance and you are paying too much while too loose you might get into trouble for edge cases. It is a good starting point, but more work needed.

      Sometimes even with the CAD files, there might not be a way to manfacture them as the machines or the process might no longer be available.

  7. The B-2 bomber logistics are all contained in Oracle Database tables, CAD, and a Web GUI. The internal systems were all installed 30+ years ago from plans 40 years ago. Heat Exchangers are not classified, but the specific design capabilities is probably not published. Basically, you just want to get rid of the avionics and engine bay heat. I’m sure the iron based components (stainless steel, etc) are cracking, rusting, and blowing apart as the plane nears EOL. Whatever you replace it with, has to fit, and has to be made remove and replace accessible by human technicians.

    1. There’s maybe a bit of special sauce in this instance, as it would have been designed to spread heat over a larger than normal area, possibly blending it with extra airflow, in order that a stealthy IR signature be maintained with no hot spots.

      1. I don’t think it has anything to do with stealth. “The B‐2 Load Heat Exchanger (NSN 1660‐01‐350‐8209FW) uses air and Ethylene Glycol Water (EGW) liquid to produce cold air for the cooling system.” The col air probably cools the avionics.

  8. “There are only 20 left of the 21 built..”

    Is it just me or does this sound weird ?

    All planes of this type that were ever build are still there except for one.
    Only 95% left..

  9. Chances are the sub contractor who built the original unit is no longer in business. I know for a fact that several tier 2 and 3 sub contractors are history.

    Even if they had the plans, the know how and tooling doesn’t exist anymore, That means they have to find a contractor who will then hire the tooling specialists(which are rare thanks to 39 years f outsourcing all our manufacture to slave labor state like China) to design and build the jigs. Then hire and train technicians to build the replacements. However since the plans are gone, that means mechanical engineers and others who specialize in stealth aircraft will have to start from reverse engineered plans, then build prototypes to work the bugs out and then have tooling build the jigs for the production run.

    Expect it to take 2-3 years minimum.

    NG is lucky that Lockheed-Boeing still do Stealth craft manufacturing so there is still active pool of experts in the field.

    This the major drawback of privatizing defense work. Once the work ends, all the manufacturing know-how is lost. This is why we cannot duplicate the Shuttle or Apollo. Both were the end point of 30 plus years of constant work by hundreds of aerospace and other contractors who by the time those programs were started had amassed a vast amount of manufacturing and design expertise that cannot be replicated today. We simply lack the industrial base to pull it off.

    One of the worst things that happened was a spate of aerospace mergers in the 90’s(which were illegal as hell) that wiped out a lot of the smaller players from existence and put their workers into the unemployment lines.

    1. Supposedly all the F-22 tooling has been preserved, but a company contracting to set up a production line would have to figure it all out, then there would be updates and upgrades.

      At least it didn’t suffer the fate of the C-17 tooling. About 2 weeks after end of production, it was split up into multiple lots and sold on eBay. IIRC as tooling, fixtures etc for a “large cargo aircraft”. Dunno why the company would have been prohibited from stating it was for the C-17. If I’d had E. Musk amounts of money I’d have bought it all and set up a factory to build a civilian version.

    2. The United States Constitution expressly prohibits funding the military for more than 2 years at a time, thus long term contracts are always subject to cancellation.

        1. Article 1, Section 8, Clause 12:
          (Congress shall have the power) To raise and support Armies, but no Appropriation of Money to that Use shall be for a longer Term than two Years;

  10. Sounds like the tooling and process has been lost. Click the link to the job ad, and there’s a lot of interesting information. For example, they’ll send 2 heat exchangers to whomever takes the job.

    “This engineering effort is to reverse engineer the core of the B‐2 Load Heat Exchangers, develop disassembly process to remove defective cores, develop a stacking, vacuum brazing and welding process to manufacture new heat exchanger cores and to develop a welding process to install the new cores on existing B‐2 Load Heat Exchangers. The requirement includes reverse engineering the re‐core process for the B‐2 Load Heat Exchangers. The B‐2 Load Heat Exchanger (NSN 1660‐01‐350‐8209FW) uses air and Ethylene Glycol Water (EGW) liquid to produce cold air for the cooling system.”

    1. With some experience in military contracting, it appears to me that they are looking for procedures and processes to REPAIR defective cores. That would be a valid requirement. They also appear to be looking for an improvement that might help with a recurring failure issue. That is not the same as “losing the plans”. There is a big difference between producing the original heat exchanger and a process for re-manufacturing them.

      1. Yes – that’s how the article looked to me. That they need a good way to repair existing heat exchangers without scrapping the whole unit. It’s possible that there never was a plan for how to replace the core without scrapping the whole unit. The description states that the core is welded in and not easily serviced, and now they need a way to service it.

        1. Sounds to me that something like an endoscopic version of a locomotive boiler tube tool is wanted. So you pop it out the remote end, and expand a bit that cuts off the flange of the tube, vacuuming out the chips, then you retract that, pull the tube, insert replacement, then it expands a flange roller at the remote end, and “rivets” it in like that… then you have to pull a high vacuum on the whole thing and braze around the flange.

  11. There’s a huge difference between the plans and specifications of an assembly and the procedures for manufacturing, disassembling, repairing and testing the same assembly. When the original equipment manufacturer runs out of spare assemblies to sell you, and when you run out of spare assemblies yourself, you must have either bought training and maintained staff to support it, or bought the training material from the OEM, or you’re in the position the Air Force is in now.

    It’s usually no problem to purchase an extended support contract of some sort: look at Microsoft, Windows XP, and the US Military, or the Navy’s CASS program. If you wait too long, though, what you want cannot be had at any price, and you have to rediscover what is lost. Unless the process is very subtle, you will end up recapitulating large portions of the development cycle: look at how the United States bungled around with the chamber pressures on the V2 rocket when they figured they’d “improve” performance.

  12. If the original plans called for 132, and they built 21, where do you think the funding went that was allocated for the other ones? I wonder what political pocket that went into, or, what black ops program it funded? Generally they award the contracts to the companies and fund the program at that point, right? The companies build test planes for them to submit, and then gear up after gaining the contract if I remember. Of course that takes a back seat to no-bid contracts, those are all shits, grins, and hookers….

    1. The way the USA does these contracts, the manufacturer(s) get the full bid price, no matter how many of the thing they build. Politicians against a thing will do everything they can to get the number ordered, cut, and cut, and cut some more so they can bitch about how expensive each one is as they try to get the thing cancelled completely.

      Some of the time they’re successful at killing a procurement, like the Comanche helicopter (here we are some years later and the US Army is again wanting a chopper with capabilities the Comanche was built to fulfill) and since it didn’t get past the prototype stage the company didn’t get the $$$ for the production contract. But then there are the almost but not quite killed programs like the Zumwalt class ships. That got cut down to three, with capabilities so severely reduced, they’ve become expensive “white elephants”.

      So if the government does a 50 million dollar contract for 50 of a thing but Congress cuts the order to 10, the manufacturer is still getting 50 million, and that can go even higher if there are problems. So those million dollar gizmos end up costing 6 or 7 million each.

      What ought to happen is when the contract is finalized, that’s the end of changes. The government initially wanted 750 F-22 Raptor fighters, by damn that’s what they should have had to take. Then the cost per plane would have been considerably less.

      1. At least if they got stuck with 130ish and could only run a couple of dozen, they could send a tech sergeant out to the desert boneyard with a sawzall to get parts.

      2. Complete rubbish. Congress doesn’t modify the order. They appropriate money for it or not. There might be a plan for 50, but if congress cuts it to 10 then only 10 are ordered. The contractor gets money for 10.

        Where prices appear to start going up is when the government pays for the design cost and it is reported in the unit cost. So using the B-2 example, the design costs were ~$23 Billion spread across 21 aircraft. Building each aircraft cost ~$750 Million. But the way the costs are reported in newspapers, the $23 Billion is spread over the 21 aircraft and it becomes a $2 Billion aircraft.

    2. Why the conspiracy theory? The funding went poof. They ordered the last 4 in 1993 (see the tail numbers), and although they took 4 years to get delivered, it was all allocated around then, and the massive reduction is directly traceable to 1991, when the USSR disappeared.

      The MFGs don’t get full bid price when the contract gets cancelled underneath them. They got a certain amount of cash for R&D/setting up the production line, and a certain chunk of cash per plane. The first chunk is sunk once production begins. The second and succeeding chunks require the USG to buy the things. If it’s $31 billion for the R&D, and production setup, and each plane costs $500 million per plane, then they cancel it at 21 planes, the program cost per plane is $2 billion dollar, going up every time they reduce the number of planes, because the first chunk is sunk already. If they had built all 132 planes, then each would have cost ~$700million once the R&D cost was amortized across the fleet.

  13. A plane I used to work on, had an interesting Heat Exchanger design, which is probably common in modern aircraft, in that the plane has to operate from 0 feet to 50 kft temperature-humidity ranges. A common technique is to use the fuel tanks for cooling the radiators. Hot EGW (anti-freeze and water) pumped through the radiators is absorbed by the fuel. The problem comes when you are reaching a low fuel state, and the radiators become uncovered. Then-to precautions of pressurizing the EGW with a static nitrogen pressure to make sure fuel doesn’t leak back into it. So, basically you cool all the equipment with EGW which absorbes the heat , and then cools (exchanges) it in the radiators. In most cases you can’t really power-up systems until you reach 20kft+ as outside air starts getting cool enough to turn on the mission radar, and other high heat avionics. The B-2 has outside cold air ducting, so that I’m sure part of the heat exchanger is conducted outside the fuel tanks (if not all) at altitude, much like the Model-T Ford.

      1. By the by, it seems to have been for “marketing purposes” like the F-105 and F-111, not for overseas sales, but to “sell” it to the higher echelon of USAF pilots who all want to fly “fighters”.

        1. At least they tucked an A on the end. Like, we do that too, so don’t throw it around like fighter jock stuff.
          Fighters seem to be what gets the most love, promotions and funding.
          The continued need for the A-10 keeps embarrassing them. But ______ (substitute latest flavour here) can do it too! (then why haven’t any of them…)

  14. “This engineering effort is to reverse engineer the core of the B‐2 Load Heat Exchangers, develop disassembly process to remove defective cores, develop a stacking, vacuum brazing, and welding process to manufacture new heat exchanger cores and to develop a welding process to install the new cores on existing B‐2 Load Heat Exchangers.”

    It almost sounds like what they need isn’t a complete new part, but a reliable way to re-core the heat exchanger without replacing the complete assembly. Could be that the problem is that they still have the blueprints for a complete assembly, but want a cheaper way to extend the service life of the ones already in the field.

  15. At a former employer, in a time long ago and in another reality, I designed the ATE for stuff that formed the CASS for the F-14 program; that is, the test set that tested the test set. When the project was terminated, my employer literally begged the navy to take and preserve and archive the associated ATE and documentation, as we were not able to provide long-term storage. During the F-18 startup, the DoD demanded that my (now former) employer release all drawings and code hardware for the F-14 CASS because they wanted to use it to jump-start changes to the new CASS that supports the F-18. My former boss sent sent the old message traffic and meeting notes to the people wearing stars at NAVAIR to shut them up.

    Moral of the story is that all governmental bodies are managed by corrupt and/or incompetent and/or uncaring people.

  16. Or they have an interesting problem like the Apollo heat shield
    They have the exact method to build them, you just take “Stuff A” and “Stuff B” and combine them in X method.
    The problem? Stuff A and Stuff B were both proprietary materials that USED to be commonly available, and no longer are, and in fact, the companies that made both A and B have been out of business for decades

    It doesn’t even take high tech stuff. Let’s talk cotton twine. There was one called Dreadnaught cord – loved by folks who do knotting and others. Old rolls go for big bucks
    We know where the machines are – they are in the museum at Mystic CT. They are even basically setup. No one can make the cords – they don’t know the details of the setup and process

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.