how weird are the electronics actually used in nuclear weapons?

[dmills]

I know a few HOBBYISTS that are having a lot of fun playing with experimental designs for EBW, EF and Slapper detonators and associated pulse supplies, the tech is out there, in the public domain and within the reach of a decent engineers home workshop (Electronics and possibly detonators, not the main physics package). Last I heard they were experimenting with having the things made as flexi PCBs, with printed planar transformers to get the drive impedance somewhat reasonable, and there was talk of saturating reactors for pulse compression right on the flexi.

We got **REALLY** lucky that heavy isotope enrichment is such a pain in the arse, it means that special materials availability is a very effective barrier even to nation state level actors. If there was a viable chemical approach to say uranium isotope separation, and is U235 percentages were a bit higher we would have a real problem. We also got lucky with Pu in that Pu240 isotopic contamination makes the stuff significantly less useful for weapons if you cook it too long.

Electronically, I would be guessing that the pulse power stuff is where the interesting electronics design hides (But is probably not actually that different to what you would find in something like a pulsed gas laser or old school radar set), and the mechanical design to survive the acceleration probably make for interesting assembly issues.
 
Some of the interaction of the electronics and physics for the timing to give variable yield is probably black magic, and I would expect that electronically configured delays may well form part of the positive control interlocks (So that these need to be configured based on information that can only be decoded if the correct arming data is entered).

I don't see the electronics as a big deal, probably simpler then something like a Rb or Cs frequency standard physics package driver.

Interestingly the spherical implosion design (With its timing complexities) really only makes sense if you are critically short on material, there are fairly obvious geometries requiring as little as two detonators instead of dozens of the things in the early devices and that makes the timing problem much less of a pain at the cost of requiring slightly more Pu.   

All just speculation of course.

Regards, Dan.

[coppercone2]

i really intended for this thread to be around the logic level/power supply level design stuff not the actual physics package detonator electronics. I thought there might be something to learn.

The nutristor/neutron tube electronics are interesting in their own right though and are both in a different category then the nuclear control system (the goal of this thread) and the detonator system. I don't know what the categorize them as. Nuclear process control system maybe.

I wonder if its dangerous to be around a nuclear weapon incase the neutron generators kick in by accident. It would be really nasty to get a dose of neutron radiation. Worst kind of radiation IMO.  There must be a interesting control loop there.

The movie broken arrow had a cool bit about the control board. what a great movie i might watch that again.

I did see a switch though, made with something like 200 MOSFET somewhere, on a single substrate, with something like 'special military uses' in it. I think it was a giant cascode. I assume it might be for the detonator circuit.


but i gotta say i have been looking at some pictures and video and I see some seriously shady shit, like the bomb cart they use for the b61. what the fuck, its just all exposed out there, you can probably crash it into a vending machine and ruin the bomb and cause a radioactive disaster. they could put a plywood cover on it or something at least. probably costs like 1 million dollar to replace a bomb fin too. the carts also made like shit (its like impossible to clean and you can hide a tracker in the shoddy construction (no caps welded on).

who thought this was a good idea? seriously?


these people need a bigger budget.

and yes, i am pretty sure once in my life some where i saw a big blast door that had MUNITIONS stenciled on it with a vending machine in the same room.

[coppice]

I did see a switch though, made with something like 200 MOSFET somewhere, on a single substrate, with something like 'special military uses' in it. I think it was a giant cascode. I assume it might be for the detonator circuit.

That doesn't really make military circuits different. It just means a high BOM doesn't immediately make a military project a non-starter, as it would for almost anything civilian.

[coppercone2]

I did see a switch though, made with something like 200 MOSFET somewhere, on a single substrate, with something like 'special military uses' in it. I think it was a giant cascode. I assume it might be for the detonator circuit.

That doesn't really make military circuits different. It just means a high BOM doesn't immediately make a military project a non-starter, as it would for almost anything civilian.

i think a giant cascode is pretty weird. no one would make that discrete ever imo. i need to find it.

[LaserSteve]

  Long ago and far away (20 years now)  I worked for a surplus company that dismantled some prototype, training,  and manufacturing development modules made for a intermediate range missile. Lots of flexible circuit boards as interconnects.  Simple circuits and thru hole parts everywhere. Nearly all parts were 1% tolerance.  Massive amounts of gold plating and house marked parts.  Connectors were huge with big pins to withstand G.  Everything was conformal coated.  Wire bundles were used instead of cables and were dressed with flat nylon lacing tape.  The soldering was by wave machine and hand, with very tight specs on visual inspection.  Older "reliable" parts were used.  A high grade G10 everywhere, and massive hermetic sealing by using machined blocks and feedthroughs. A few custom packages, and most capacitors were in hermetic metal cans laid on their sides and held to the board by the conformal coating, A 4.7 uF tantalum would be a metal can part, not a plastic gumdrop as it would be on a civilian board.   

 We did get to set off one thermal battery that was used for taking pictures for the manual.. All the rest were dummies in "Inert" blue.

Mechanical stuff like the fin actuators had 10 turn or 360' sensing potentiometers made to mil spec and were all miniature wire wound in "servo" cans.  Everything was stock a gear motor, nothing exotic except the plating on the housings. Everything had baked on paint or corrosion resistant anodyne finishes. All labeling was black paint text stenciled on, except for the pcbs which had a white stenciled paint.
You got the feeling the overbuilt gear boxes for the fins would bend a car fender.  One memorable thing was solid tungsten balance weights installed where needed. And yes I'm sure it was tungsten.

FOR THE RECORD..
NO, I did not keep any parts for examples. My father in the plant across town was involved in building the things and I knew better... We recycled all the alloys in the proper ways.


Very "70s" and simple  inside. Wide spacing between components and wide, tinned, board traces. All wiring that carried current was white stranded Teflon  unless it was on a very heavy duty flex PCB made of something like Kapton. Wires soldered into forked terminals, like machined standoffs, never direct to the board.  Contrasted with early WIFI and barcode gear we were dismantling with massive amounts of SMD and VLSI, it looked simple, elegant, and tough.


Steve

[coppercone2]

why don't they use cables? for inspection requirements? or special weave for EMI/toughness?

i actually like the 70's look of gear, my own stuff ends up looking like that lol

[LaserSteve]

Making your own  cables from wire bundles is common in aerospace, more reliable, easier to qualify for testing, and more fireproof.  Until recently, whole jet airliners and things like space shuttles are wired this way. Differential signaling is used to mitigate the need for shielding.

Steve

 

[coppercone2]

it seems like you can slip a shield on easily.

how come the practice is going away now? whats replacing it?

[LaserSteve]

Power cables and a specialized  mil-spec optical fiber to local modules. Saves weight and is more reliable.

Uncle Sam likes SIMPLE and Tough. Getting the jacket off the wire requires strength and specialized tools, and an approved  mil-spec blade for your calibrated  wirestripper. (Hit the Ideal Custom Stripmaster web site and look at that expensive blade's price)

 One other thing, there was little or nothing in there directly useful if you found one laying on the ground un-used. That is by design.

Looks like this installed, scroll down a bit.

https://www.google.com/url?sa=i&rct=j&q=&esrc=s&source=images&cd=&cad=rja&uact=8&ved=2ahUKEwjm6Oya1o7fAhVOpIMKHekKDHEQjRx6BAgBEAU&url=http%3A%2F%2Fecolift.com%2Farticle.php%3Fid%3D1435874233&psig=AOvVaw2JQiHKIzJPGv-imaNCstBS&ust=1544304875968019

and this is the shielded:

https://www.google.com/search?q=mil+spec+aircraft+wiring&source=lnms&tbm=isch&sa=X&ved=0ahUKEwjpira01o7fAhWh6oMKHQ7jCSEQ_AUIDygC&biw=1366&bih=501#imgrc=q7fGsN7ztS4H2M:

This is installed, only for mil applications expect everything to be white (Even for Nato, Russian and Chinese Mil)  with numbers or identifiers custom  printed on the wire every few inches.
https://www.google.com/url?sa=i&rct=j&q=&esrc=s&source=images&cd=&cad=rja&uact=8&ved=2ahUKEwjdhf_f1o7fAhXq8YMKHTEkAWMQjRx6BAgBEAU&url=https%3A%2F%2Fwww.aviationtoday.com%2F2016%2F04%2F01%2Fwire-and-cable-aircraft-ewis-and-swap-in-focus%2F&psig=AOvVaw1K1Jo2piSTn2q6EPPgHPdz&ust=1544305041346658

Steve

[coppercone2]

how do they quantify the reliability of the fiber optic assembly compared to the cable (if you know?)

It seems ridiculously difficult to design a test.

[coppercone2]

thats a fun activity though, making your own cables

[mc172]

what do you mean by non socially popular?

Tell people at a dinner party you design nuclear or some other weapons system and see how well it goes down.

I did do that when I worked in the aerospace industry designing weapons. People tended to think I'm some kind of genius and were amazed rather than getting angry! I wasn't all that bothered about telling people what I did, but people would always say "tell Bob what you do!", "have you heard what mc172 does for a living?" etc.

[coppice]

what do you mean by non socially popular?

Tell people at a dinner party you design nuclear or some other weapons system and see how well it goes down.

I did do that when I worked in the aerospace industry designing weapons. People tended to think I'm some kind of genius and were amazed rather than getting angry! I wasn't all that bothered about telling people what I did, but people would always say "tell Bob what you do!", "have you heard what mc172 does for a living?" etc.

The thing is to emphasise that you work on propulsion systems, because that makes you a rocket scientist.

[coppercone2]

'i make sure a jagged bit of metal severs an artery with 50% chance at 100 meters from where the man with the xbox controller decided the explosion should be'

then if the response is positive, you might wonder if you are sitting at dinner with the mason family. or if red white and blue runs down the toilet after dinner!

[duak]

The book "Command and Control" by Eric Schlosser is a popular account of the development of the safety systems in US nuclear weapons interleaved with the Titan II missile explosion in its silo in 1980.  The number of near missies shows that the designers of the gadgets were barely ahead of the circumstances that tried to accidently set them off.  The nerd in me wants to know more about the nuts and bolts of the safety system designs and how, by sheer luck, none have actually been fully circumvented.  I'd love to have a more technical analysis about some of the incidents.

For a few years I designed parts of and worked on products that had hazards of one type or another.  It was a learning experience because they had to be fail-safe.  I don't think it's possible for everything, but I was at least confident that a single point failure wouldn't expose anyone to a hazard.  I kept asking our UL/CSA experts how toasters can get approval when their cases aren't even grounded?  I didn't get an answer but one allowed that he was happy that he didn't have to do it.

I'll bet the gadget's systems are electromechanical with serial and parallel connections for redundancy.  You can't trust semiconductors to not short out so they are backed up with good, old contacts.

[CatalinaWOW]

Do a google on FMECA, and then read MIL-HDBK-1512 and MS1316.  Then MIL-STD-882.  It will give a start to understanding how such things are approached.  Without even getting into the extra layers for nuclear.

The approach is methodical and enormous.  There are two big issues.  One problem is that it can only protect against the problems you can imagine.  Almost all of the near misses come from something no one imagined.  Some of them are obvious in hindsight, others less so.  The other problem is that because of the enormity of the documentation, it is easy to just throw paper at the wall with thinking much at all.  It is extremely hard to properly review thousands or tens of thousands of pages of analysis thoughtfully to find holes and omissions.  When time or money runs short the product quality is evaluated by using a scale.  More pounds of paper is better.

[Domagoj T]

My understanding of implosion type nukes is that you surround the fissile core with conventional high explosives and that the major problem is getting those conventional explosives to go bang as synchronized as possible.
My question is, wouldn't a single mechanical switch do just fine, with detonators connected in parallel, assuming you match the lengths of trigger wires that go to the conventional explosives? Yes, I understand that in order to get better sync you need high power, but the mechanical switch I have in my mind are two beefy blocks of copper that are pushed together by a small explosive charge (gun type).
However, the declassified data we have available suggests that this approach would not work, so what am I missing?

[T3sl4co1l]

Even explosively propelled switches close in microseconds (velocity of detonation ~ 10km/s, closing a gap of say 10mm is 1us).  There will of course be a spark along the way, but it is probably the case that the spark isn't efficient enough (it should still be sharp enough though).

The obvious solution to timing, implies that either the devices cannot practically be connected in parallel or series (perhaps as they come apart, the impedance changes too much, shunting power from others?), or that more precise timing is necessary, beyond what is reasonable with transmission lines (or that the transmission lines themselves are impractical -- they may be on the order of single ohm Zo, requiring a rather wide build and being quite lossy?).  Or that using transmission lines is prohibitively heavy (since this is typically a flown munition -- not that weight ever trumped strategic value with these things..).

There's also the oblate version, which only requires two detonators and some more precise explosive lenses.  That would presumably be even easier to fire with a power splitter and a single source (even if the impedances go wildly wonky, it should be practical to deliver enough excess to complete both?).

It may well be that the engineers involved, took the lazy way out -- reduce stray inductance and put the capacitor and switch right on top of the detonator, rather than trying to transmit power any distance, into any combination of detonators (including just one).  Maybe the impedance change during firing is so dramatic that transmission lines above a rather meager length simply aren't practical, let alone wiring in series or parallel?

Not that they needed to be lazy -- they certainly had access to top minds in all fields.  Though I don't know anyone offhand in EE specifically.  It seems they are rather overshadowed by the more famous physicists, in popular telling of the project.  (Checking at a glance, it looks that they're not completely forgotten, and indeed there is collected history, at least in brief, for many (most? all?) persons that worked on the Manhattan project.  It would probably be an easy history paper to research, concerning specific professions such as EE, ME, CE and such, on the project.)

Tim

[rstofer]

what do you mean by non socially popular?

Tell people at a dinner party you design nuclear or some other weapons system and see how well it goes down.

Or work for a company that builds nuclear power plants!  That will be lively...  See what happens when you bring up waste disposal!

[Gary350z]

what do you mean by non socially popular?

Tell people at a dinner party you design nuclear or some other weapons system and see how well it goes down.

Not a problem in my experience. In fact people are quite impressed.

[coppercone2]

considering the chinese are stealing clearance lists and stuff, you guys are not concerned telling people you work on damn weapons?

i honestly expected a shadowy mismash of smoke and mirrors as responses

[CatalinaWOW]

considering the chinese are stealing clearance lists and stuff, you guys are not concerned telling people you work on damn weapons?

i honestly expected a shadowy mismash of smoke and mirrors as responses

You don't think the lists they stole are a better source of information than this blog?  Or the lists they compiled from photographing cars on the way into the plants?

So no, I am not concerned.  At most they might get some information on who is willing to be careless with information that shouldn't be easy to get to.   

Your statement "damn weapons" implies an ethical judgement.  While I totally agree that it would be better if no one in the world worked on weapons, there is quite a bit of history to suggest that that solution is not easily achievable, and perhaps impossible.  So the ethical decision is not cut and dried.  I thought through my choice before accepting my first job in the area, and found reasons that were sufficient for me.  Those reasons are also part of the reason I am not concerned.   Doesn't mean I am right or that my decision is correct for anyone else.  Anyone who does not have doubts about their choice, whichever way they go, has not thought seriously about the subject.

[coppercone2]

no by damn weapons i mean its hot shit that alot of people are interested in.

you know like your walking through a bad flava and you pull out a damn rolex. in my mind alot of strange characters might get interested in you for mentioning such a fact.

i like all that stuff its better then dying of dysintary in a castle siege in the year 1305, i like all that peace through atoms stuff. can't say i have any love for bio chemical weapons though. I don't know if its much better once you start dealing with a world like 'the road' but it might be a bit of a upgrade to what war used to be. I think it lets a society set a savagery limit thats like, if you get this bad we will just kill everyone because we would rather not function in those extremes anyway. I also think that alot of 'creative' methods of waging war are basically not utilized so you have less weird contingencies to worry about.

[james_s]

What difference does it make? Anybody can say anything they want, I could post here that I work on weapons, it's not actually true but who would know and what could they get out of it? In my last job I worked with a guy who had previously worked on cruise missile guidance systems at one of the big aerospace companies, he also wrote a significant portion of the operating system for one of the popular 80s home computer systems, neither was any big secret.

Foreign nations already have networks of spies and other sources for information, they're not going to get anything useful from forum posts like this. There's nothing being discussed here that isn't already present on various other websites. The general information on how nuclear weapons work has been freely available for decades, that still doesn't make it easy to build one, as is true with many things, the devil is in the details.

[coppercone2]

i think its dangerous to attribute all knowingness to government intelligence agencies, they can get what they can get. Something can go from dubious to plausible or plausible to confirmed based on different information sources and something can occur because of this escalation. Or they might not want to do something without additional information for a psychological profile or whatever it is they want.

I mean its like your neighbors, they might be a peeping tom from their kitchen window with alot of ideas about why a car shows up in your drive at night, coming up with all kinds of crazy shit, but when they start going through your garbage and hanging out by your patio fence suddenly they become more of a pita because they know how you are so much more. The more people know you the less anxious they are about doing stuff.. quite a bit of crime, or asshole behavior (i.e. stealing food at work) occurs with people you know because they figure out your 'chill' or something.