Large site logo

To Open The Sky

The Front Pages of Christopher P. Winter

Nuclear Thermal Rocket Tests


The NERVA (Nuclear Engines for Rocket Vehicle Applications) program conducted as many as twenty-three engine tests, achieving core power densities of 2kW per cm3 and coolant outlet temperatures of 2,500° C at pressures of 560 psia. Specific-impulse values of around 760 were achieved — twice that of the best chemical engines. Steady progress was made in engine efficiency and controllability, and in lowering the release of radioactivity. The XE-prime, the last engine in the series, was tested under simulated space conditions. It operated for a total of three hours 48 minutes (11 minutes at full power), starting and stopping 28 times. The table below, based on a page at the Web site of the Federation of American Scientists, provides sketchy information on these tests.

The NERVA program struggled with two major problems. The first was cracking of the ceramic fuel elements due to vibration. This was solved by redesigned engines that reduced vibration levels. The second was release of uranium into the exhaust stream caused by erosion of the fuel-element coatings and diffusion of uranium through them. At program termination, this contamination problem had been reduced but not eliminated.

Some Information on NERVA Tests
TESTED DATE (seconds) (kNt) LEVEL TEMP Time Time Notes
Source: (Removed; see the Wayback Machine 2006-2008.)
KIWI A July 1959 ? ? 70MWt 2683°K 8 Min. ? UO2 plates; Core damage
KIWI A' July 1959? ? ? 85MWt ? 6 Min. ? UO2 cylinders; Core damage
KIWI A3 Oct. 1959 ? ? 100MWt ? 5 Min. ? Improved cladding; some core damage
KIWI B1A Dec. 1961 ? ? 300MWt ? 0.5 Min. ? Fire due to H2 leak
KIWI B1B Sep. 1962 ? ? 900MWt ? <10 Sec ? Fuel elements ejected
KIWI B2 N/A N/A N/A N/A N/A N/A N/A Design not tested.
KIWI B3 N/A N/A N/A N/A N/A N/A N/A Design not tested.
KIWI B4A Nov. 1962 ? ? ? ? ? ? Core vibration damage
KIWI B4B N/A N/A N/A N/A N/A N/A N/A Design not tested.
KIWI B4C N/A N/A N/A N/A N/A N/A N/A Design not tested.
KIWI B4D May 1964 ? ? ? ? 1 Min. 1 Min. Nozzle cooling tube rupture
KIWI B4E ? ? ? ? ? 12 Min. 6 Min. First use of coated UC fuel
NRX-A1/2 Sep. 1964 760 ? 1100MWt ? ? Min. 5 Min. Duration limited by H2 supply.
KIWI TNT 1/12/1965 ? ? ~107 MWt ? ~1 mSec ? Intentional coolant loss; reactor destroyed
NRX-A3 Apr. 1965 ? ? ? ? 8 Min. 3.5 Min. Terminated by spurious turbine overspeed signal.
NRX-A3 May 1965 ? ? ? ? 45 Min. 16.5 Min. Success
Phoebus 1A June 1965 ? ? 1090MWt 2370°K >10 Min. 10 Min. New design; test successful
NRX-EST Feb. 1966 ? ? 1200MWt ? 110 Min. 28 Min. Total of tests over five days.
NRX-A5 June 1966 ? ? ? ? ? Min. 30 Min. Test successful; duration limited by H2 supply.
Phoebus 1B Feb. 1967 ? ? 1500MWt ? 45 Min. 30 Min. Test successful
NRX-A6 Dec. 1967 ? ? 1100MWt ? ? Min. 60 Min. Test successful; NERVA design goal surpassed.
XECF/XE' Mar. 1968 ? ? 1100MWt ? 115 Min. ? Min. Total of 28 starts; downward-firing
Phoebus 2A June 1968 ? ? 4000MWt 2310°K ? Min. 12.5 Min. Reactor restart demonstrated.
PEWEE ? 845 ? 503MWt 2550°K ? ? ZrC cladding; 5.2GWt/m3 peak
NF-1 1972 ? ? 44MWt 2500°K ? ? Multiple tests aimed at matching Tce
NF-2 N/A N/A N/A N/A N/A N/A N/A Design not tested; program terminated.


  • There's a dramatic picture of that KIWI TNT test — the one that blew up the reactor on purpose — at the FAS link cited above. Man, it blowed up real good! The FAS page also provides a good deal of textual information.
  • See also Space Nuclear Power at the FAS site for links to a series of reports and other sources.
  • I found a Microsoft Word document about the KIWI TNT test and cleaned it up. (A few words and phrases were illegible.) Produced in January 1968 by staff at Los Alamos National Laboratory, it covers the explosion and radiation release in considerable technical detail. Most noteworthy is its brief mention of the extensive changes (such as speeding up the control-rod actuators by a factor of 100) required to enable the engine to explode.
Valid CSS! Valid HTML 4.01 Strict To contact Chris Winter, send email to this address.
Copyright © 2003-2015 Christopher P. Winter. All rights reserved.
This page was last modified on 24 October 2015.