Shop OBEX P1 Docs P2 Docs Learn Events
Neutron Flux Density Increasing — Parallax Forums

Neutron Flux Density Increasing

CounterRotatingPropsCounterRotatingProps Posts: 1,132
edited 2009-10-10 20:27 in General Discussion
No, were not talking about Star Trek or Flux Capacitors - it's a real, measured phenomenon.

Found the article and especially the charts·interesting.

" Data from around the world shows that sea-level neutron flux rates are continuing to increase, which means that problems once confined to the heavens have come home to roost for system designers here on Earth. "

http://www.pldesignline.com/showArticle.jhtml?articleID=220301387

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔

Comments

  • HollyMinkowskiHollyMinkowski Posts: 1,398
    edited 2009-10-08 14:51
    That really introduces uncertainty into a critical system :-(

    A really bitter pill for FPGA guys to have to swallow.

    Redundancy and error correction would help but that
    makes everything more expensive, slower and harder
    to design.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    "Where am I? Where am I going? Why am I in a handbasket?"
  • CounterRotatingPropsCounterRotatingProps Posts: 1,132
    edited 2009-10-08 16:56
    If I read the article correctly, it seems like this could potentially affect every IC.

    What really drills this home is that these solar neutrons "can penetrate several feet of concrete".

    Forget the Tanning Salons - just sit in your Cold War era bunker!


    _______________________________________________________________________
    OK OK before the purists jump all over me - I know its U.V. not N. that tans - at least
    the radiation might give a free skin cancer removal treatment (or cause it) *yikes!*

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
  • sylvie369sylvie369 Posts: 1,622
    edited 2009-10-08 17:38
    Sounds to me like the best excuse ever.

    "Sorry, but the neutron flux density increase ate my homework".
  • CounterRotatingPropsCounterRotatingProps Posts: 1,132
    edited 2009-10-08 18:16
    LOL Paul,

    and the teacher's reply?

    "Yeah, I know what you mean - the influx of the flux seems to have had many side effects. Next to your name in my grade book the 'B' now seems to say 'D' !"

    hehehe

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
  • edited 2009-10-08 18:55
    I was discussing this on the net and I'm told that FPGA's are not affected because only a wide spectrum of an electro magnetic pulse are required·to do this.· They said neutrons don't carry electrical charges.

    (Edited for clarity)

    Post Edited (Chuckz) : 10/8/2009 7:32:56 PM GMT
  • sylvie369sylvie369 Posts: 1,622
    edited 2009-10-08 20:05
    Chuckz said...

    They said neutrons don't carry electrical charges.

    Okay, if any of my students try this as an excuse, that'll be my response. Thanks.
  • AlsowolfmanAlsowolfman Posts: 65
    edited 2009-10-08 21:02
    neutrons themselves do not carry charge, but free neutrons have a half life of 15 minutes, and when they decay they can produce a proton and an electron. so they lead to free charges.
  • edited 2009-10-08 22:33
    Alsowolfman said...
    neutrons themselves do not carry charge, but free neutrons have a half life of 15 minutes, and when they decay they can produce a proton and an electron. so they lead to free charges.
    "Free neutrons are produced in nuclear fission and fusion."-Wikipedia
    http://en.wikipedia.org/wiki/Neutron

    "To actually destroy or destabilize electronic devices like FPGA's you need a *intense*, wide spectrum EM-pulse. An exploding atom bomb can do this, and certain devices specifically designed for that task."

    http://www.natami.net/knowledge.php?b=6&note=11946

    The Natami is an advanced Amiga in FPGA.




    Post Edited (Chuckz) : 10/8/2009 10:38:55 PM GMT
  • AlsowolfmanAlsowolfman Posts: 65
    edited 2009-10-08 23:54
    The article is not referring to the destruction of the device, it is referring to a single bit of ram being flipped by a free neutron put off by solar fusion, or some other cosmic source. It is discussed in the soft errors section of Wikipedia (http://en.wikipedia.org/wiki/Soft_error#Cosmic_rays_creating_energetic_neutrons_and_protons). This effects all ram based devices, but it is worst fpgas as a single bit change in the ram changes the hardware, not just the software.
    I was wrong about the origin of the charge from free neutrons being decay, it is neutron capture.

    "Neutrons are uncharged and cannot disturb a circuit on their own, but undergo neutron capture by the nucleus of an atom in a chip. This process may result in the production of charged secondaries, such as alpha particles and oxygen nuclei, which can then cause soft errors."- Wikipedia.
  • David BDavid B Posts: 592
    edited 2009-10-09 00:04
    I also was wondering how a neutral particle would affect electrical charges, and after poking around the web, found a site that said that another damage mechanism is that because of the neutron's speed, it can ionize the atoms that it smashes into, creating plenty of charged particles in its wake.
  • edited 2009-10-09 00:29
    Alsowolfman said...
    The article is not referring to the destruction of the device, it is referring to a single bit of ram being flipped by a free neutron put off by solar fusion, or some other cosmic source. It is discussed in the soft errors section of Wikipedia (http://en.wikipedia.org/wiki/Soft_error#Cosmic_rays_creating_energetic_neutrons_and_protons). This effects all ram based devices, but it is worst fpgas as a single bit change in the ram changes the hardware, not just the software.
    I was wrong about the origin of the charge from free neutrons being decay, it is neutron capture.

    "Neutrons are uncharged and cannot disturb a circuit on their own, but undergo neutron capture by the nucleus of an atom in a chip. This process may result in the production of charged secondaries, such as alpha particles and oxygen nuclei, which can then cause soft errors."- Wikipedia.
    It is solved by having data duplication or by multi bit memory states.· I remember typing in machine language programs from a·magazine and the last hexadecimal numbers would always be the checksum so if I made a mistake typing the program in, I wouldn't be able to proceed until I fixed the line of code I was on.

    ·
  • BeanBean Posts: 8,129
    edited 2009-10-09 01:44
    We did one project that had to be Rad-Hard.
    We used SECDED algorithm to account for changed bits.

    Bean.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
    Does that byte of memory hold "A", 65, $41 or %01000001 ?
    Yes it does...


    ·
  • VIRANDVIRAND Posts: 656
    edited 2009-10-10 06:58
    turn.gif BWAHAHAHA! LOL ROFLMAO!

    My Conclusion as a loyal BRAND-X user is that prices may go down on Brand-X and rise on ACTEL.

    So, I will buy a lot more Brand-X and Vendor1 and Vendor2 chips,
    and make nice rugged widgets in solid lead enclosures,
    or whatever RoHS materials have comparable radiation blocking to lead,
    and has slow neutron absorption characteristics.

    What fools believe ads that compare their product as absolutely superior to BRAND-X anyway?
    Is there a law that prevents Bayer from advertising the FACT that the competing brand whose
    generic chemical name is acetominophen is REALLY THE LEADING CAUSE OF LIVER FAILURE?
    (Second only to morbid addiction to high-proof alcohol and cigarettes.)

    The Commercial Product formerly known as "News" has been Refurbished as "MisGuidance for the Consumer Caste",
    as "Advertising" becomes as Ubiquitously Obviously False as common SPAM.

    Need to backup your semiconductors with pre-transistor technology? Maybe, but...

    ACTEL's infomercial Reminds me of when I tried to find out what kind of batteries Eveready Titanium were;
    Eveready had such a stupid explanation on their site that it looked like it was written as a
    kindergarten class assignment, complete with misspelled absolutely unscientific silliness.
    I looked for it but couldn't find it again, so I'm sure it was an embarrassment to the company and some idiot got fired.
  • edited 2009-10-10 12:00
    The obvious is that if chips are receiving that much radiation to not work then we wouldn't be alive either and that is something to point out.
  • doggiedocdoggiedoc Posts: 2,246
    edited 2009-10-10 13:41
    I've always blamed "stray electrons" for when things didn't go as planned. Perhaps I should blame the stray neutrons instead?

    =)
    Doc

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    There are 10 types of people in the world, those that understand binary and those that don't.
  • CounterRotatingPropsCounterRotatingProps Posts: 1,132
    edited 2009-10-10 19:17
    @VIRAND

    was that really an ACTEL's infomercial ??

    I was looking at the study from the university in Finland, but didn't see any ads or references to ACTEL on any of the sites - perhaps you saw one of those rotating ad banners? Or is Embedded.com *sponsored* soley by ACTEL ?? (if so, grrrrr)

    @All

    Even though I posted this in a somewhat humorous vein, the effect is real --- no not at the levels of distroying the chip, but enough to flip a few bits. If my jump instruction's location were hit hard enough to change the jump to address in a critical piece of code ... well then the dog ate my homework [noparse]:)[/noparse])

    It seems like what you guys have dug up from the web seems to indicate that the RAD doesn't have to be strong enough to kill us to have an effect on a chip. It just has to 'hit the right spot hard enough' Right?

    I guess the question is just how much of an effect would it have? Wouldn't a single neutron impact in the right spot in a CCD cause both quantum and electrical effects - because the impact transfers the momentum of the particle and could indirectly trigger a state change? It's not likely that a single N would do this as it would have to be very high energy, but how likely is it that there would be only a single N? They'd be in packets.

    @Bean:

    can you give us an idea of what kind of RAD was involved in the tests you did that would flip a bit?

    cheers
    - H
    _________________________________
    Blaming Anti-matter for all my woes [noparse]:)[/noparse]

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
  • waltcwaltc Posts: 158
    edited 2009-10-10 20:00
    Couple of sites that may be of interest.

    NASA Office of Logic Design A scientific study of the problems of digital
    engineering for space flight systems, with a view to
    their practical solution.

    http://klabs.org/index_klabs_dot_org.htm


    NASA/GSFC Radiation Effects & Analysis Home Page

    http://radhome.gsfc.nasa.gov/top.htm
  • VIRANDVIRAND Posts: 656
    edited 2009-10-10 20:23
    RE:ACTEL infomercial = OP Link...

    Last few lines of article Page 3:
    " Data from around the world shows that sea-level neutron flux rates are continuing to increase, which means that problems once confined to the heavens have come home to roost for system designers here on Earth.

    Mike Brogley, a 20 year veteran of the semiconductor industry, is a product marketing manager at Actel Corp.

    "

    Chart from article:
    500 x 262 - 35K
  • CounterRotatingPropsCounterRotatingProps Posts: 1,132
    edited 2009-10-10 20:27
    " Mike Brogley, a 20 year veteran of the semiconductor industry, is a product marketing manager at Actel Corp. "

    Well, then GGGRRRR! [noparse]:)[/noparse] --- thanks for pointing that out ... wasn't thinking about that aspect...

    At least the University research was *independent* - er, ah, was it really ???

    They probably don't have funding disclosures in Finland !

    - H

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Sign In or Register to comment.