Jeffrey LewisW76 LEP Back on Track

The W76 Life Extension Program is back on track — although NNSA officials won’t confirm that they had trouble producing FOGBANK. Frank Munger reports from Y-12:

Federal officials today confirmed that workers at the Y-12 nuclear weapons plant had completed the first set of refurbished parts for the W76 warhead.

The milestone is an indication that the life-extension project is back on track after unspecified technical problems delayed Y-12’s work on the warheads for more than a year. The Oak Ridge plant specializes in so-called secondaries – the second stage of thermonuclear warheads.

W76 warheads are deployed on Trident submarine missiles, and they are considered an essential part of the U.S. nuclear defense strategy in the post-Cold War era.

“My understanding is the first production unit (of the W76) has been diamond-stamped here within the past couple of week,” Bill Ostendorff, the principal deputy administrator of the National Nuclear Security Administration, said today during a visit to Y-12.


Ted Sherry, the federal manager at Y-12, confirmed the production milestone but would not discuss details of the W76 work.

“We had a technical issue that required a lot of support from other sites, as well as drawing on a lot of expertise here at Y-12,” Sherry said. “It involved reconstitution of an old process. It’s quite challenging to reconstitute something you haven’t done in awhile.”

Ostendorff added, “We are pleased that we’ve been able to resolve these issues. But I think one should not underestimate the challenge of stopping a production process and then many years later trying to resume it. That’s been tough.”

Funny, NNSA officials were so willing to say FOGBANK when they thought it might shake loose a little green for the RRW. Now, mums the word.

Am finishing an article on RRW as you read this …


  1. Magoo Nair (History)

    A decade ago we had a very interesting discussion with a Senior nuclear bomb scientist from the US. He was very skeptical about the ability to replicate production of nuclear warheads. He had this to say, “He was of the view that Testing was absolutely necessary where strategy was predicated on an assured nuclear ascendancy over the adversary. In his experience of designing and developing nuclear weapons he found that computed mathematical models varied from tested results. Tests based on mathematical calculations did not work. The system, thereafter, was to remodel the weapon based on multiplication of the existing mathematical model by an intutional figure and doing the test again. Often it has taken 3 to 5 repeats of the process before the test was successful. This final mathematical model is then used for production. Therefore, 3 to 5 nuclear tests were a minimum must to assure the scientist of the validation of the design. Thereafter, if a system with a differing configuration had to be produced, this mathematical model would have to be put through the same guesstimate and the product proven by testing. Past experience has shown that when drawing on computer data to design a weapon system there have been numerous failures, Codes and tests are not necessarily the same. … Even the best computer codes have to be extrapolated through intutional multiplication. He is of the view that even with the current data base and computer codes as the first principle, at their disposal – there is no substitute for a nuclear test. He is sufficiently confident that the time will come when the laboratories cannot provide the assurances required by DOE and the Administration will take the LAST RESORT AND TEST, under the escape clause of the CTBT.
    Nuclear scientists in all NWS, with whom he had interacted were of the same opinion.

    I wonder how the Y-12 lab will guarantee the efficacy of the product of the W76 Life Extension Program to assure the user?

  2. Ben Reilly

    Does it really matter for MAD? So long as other states have a reasonable assumption that the bomb will work, MAD holds.

  3. kme

    And furthermore, the argument only applies to new or significantly changed designs. Presumably they are not changing the design of the W76 in the LEP.

    The argument also fails to take into account the improvements in computer power (and therefore model accuracy) that have been made in the last 16 years since the most recent tests. Statements like “Even the best computer codes have to be extrapolated through intutional multiplication.“ are nothing more than unproven generalisations.

  4. kg (History)

    …and we don’t need MAD – who is is aimed at? NOBODY…

  5. Carey Sublette (History)

    I want to drive home kme’s point a bit more forcefully.

    The senior nuclear bomb scientist’s claim presented by Magoo Nair regarding the necessity of testing applies only to new or modified physics package designs. As such, it is an argument against the RRW program, not any of the life extension programs.

    I do notice that the claim itself must be based on unstated assumptions because it is clearly not universally valid for all thermonuclear designs. I make this assertion based on the fact that the first Soviet radiation implosion design, a megaton class air delivered device of moderate weight, developed without any significant computing resources, worked on the first test.*

    With the benefit of 50 years of scientific advancement since that achievement, 1000+ tests under the U.S. belt, and truly incredible computing power now available there must be classes of TN weapon designs that are sufficiently robust that they can be relied on to work without the sort of fudging described by the U.S. designer.

    [*I cite this device instead of Mike or the Castle tests to circumvent objections based on their massive weight, and because even by the standards of the early 1950s the Soviet computing resources were negligible.]