174 Propellants, Explosives, Pyrotechnics 26, 174 179 (2001)
The Assessment of Hazard in the Manual Handling of Explosives
Initiator Devices
R. K. Wharton*
Health and Safety Laboratory, Harpur Hill, Buxton, Derbyshire, SK17 9JN (United Kingdom)
A. W. Train
Health and Safety Executive, Hazardous Installations Directorate, Merton House, Stanley Precinct, Bootle,
Merseyside, L20 3DL (United Kingdom)
P. F. Nolan and S. C. Campbell
Chemical Engineering Research Centre, South Bank University, 103 Borough Road, London, SE1 0AA (United Kingdom)
Summary The first stage of the work involved the development of an
apparatus for measuring the forces exerted in the manual
In this paper we review recent research undertaken with the aim of
handling of small cylindrical objects (dummy initiators).
developing a compression test apparatus that simulates the application
This apparatus was commissioned using laboratory staff
of mechanical forces to which initiator devices may be typically
and then used for on-site measurements in the explosives
subjected during handling. This resulted in the development of two
different apparatuses, one supplied pressure by means of a weighted industry. The results obtained(2) indicated that both the
beam and the other by using a pneumatic piston. The latter unit has
magnitude and rate of application of the handling forces
also been used to study the response of initiator devices to impact.
were dependent on the length and diameter of the cylindrical
Additionally, we discuss some of the key information that would be
objects being handled.
required in order to develop a safety-in-handling hazard index.
The study provided a definition of the parameters that
needed to be replicated in apparatuses constructed in the
second, laboratory stimulation stage of the project. Two units
1. Introduction
were built with an initial view of reproducing typical manual
handling forces(2); one used a pivoted beam(3 5) and the
A recent examination(1) of reported accidents in the UK
second employed a pneumatic piston(6). Extensive test pro-
explosives industry over the period 1976 1992 indicated that
grammes with each apparatus produced no ignitions for 13
123 had involved initiator devices (initiators) and that 13 of
different types of initiators: 3 percussion caps, an electric
these had involved the device being handled by an employee.
(conducting composition) cap, 4 stab detonators, a flash
Although it was possible to provide an explanation for 11 of
sensitive detonator, 2 electric detonators and 2 stab type
the incidents, the causes of the remaining two were not found
initiators.
and they were ascribed to the excessive application of
The units were then used in an attempt to determine the
handling forces by the process workers. Since 4 of the 11
force that was required to initiate each device i.e. by applying
cases had already been attributed to the same cause, this
progressively greater forces than those encountered in
suggested that roughly 50% of the reported occurrences with
normal manual handling. Even though the manual handling
initiating devices resulted from mis-handling.
thresholds were greatly exceeded it was still not possible to
Since there was no test method available to determine the
obtain initiation of any of the devices.
sensitiveness of initiators to the magnitudes and rates of
These results indicate that the recorded industrial ignitions
application of forces that may be exerted during manual
may not have resulted purely from compressive forces.
manipulation, the Health and Safety Executive placed a
Previous laboratory work(7) has shown that manufacturing
research contract with South Bank University to develop a
defects such as bowed pressure discs, voids in the composi-
simulation test method that could be used for screening
tion or composition trapped between the pressure disc and the
purposes.
crimped end of the detonator case could exert an influence on
the likelihood of initiation. Furthermore, Ye et al.(8) have
highlighted that a number of operator-related and environ-
*Corresponding author; e-mail: [e-mail]
mental factors could also be important.
# WILEY-VCH Verlag GmbH, D-69469 Weinheim, 2001 0721-3115/01/0410 0174 $17.50þ:50=0
Propellants, Explosives, Pyrotechnics 26, 174 179 (2001) Hazard in the Manual Handling of Explosives Initiator Devices 175
In the present paper we extend our previous work to sequence employed have been published elsewhere(6) and the
examine other results relating to safety in handling (those experiments yielded information on the highest piston gas
derived from impact tests) and discuss the type of data that pressure below which no ignitions occurred. The minimum
would be required to form the basis of a hazard index to impact stimulus was defined as the next highest pressure
reflect this aspect of industrial explosives processing. above this value at which an ignition resulted(1).
The energy imparted to initiators during pneumatic piston
impact tests can be estimated from either the kinetic energy
2. Impact Test Methods
of the plunger or the work done by the compressed gas on
expansion. Since the latter involves a number of gross
One factor to be considered when undertaking an assess-
assumptions (e.g. ideal gas behaviour, isothermal expansion,
ment of safety in the handling of initiators is their response to
no friction) the simple kinetic energy method will yield the
knocks or to being dropped on to a hard surface. Nabiullah
more accurate results, Table 1. Individual energies evaluated
and co-workers(9) investigated the reaction to impact of the
in this way are roughly a factor of 3 5 less than those derived
type of instantaneous electric bridgewire detonators used in
from considerations of the expansion of the compressed gas
mines. The study incorporated a review of earlier work,
in the pneumatic cylinder.
including that by Kimura et al.(10) and Beaker(11), and
The results obtained in the impact testing of a range of
concluded that these devices are susceptible to initiation if
initiators are summarized in Table 2.
impacted with sufficient force (or if heated enough). Ignition
occurred with impact energies in the range 7 10 J and the
detonators were found to be more sensitive to impact when
2.2 Drop Ball Impact Test
mounted horizontally, particularly for impacts on sections
containing the fusehead or delay element.
Drop weight impact tests have been widely used for a
number of years to assess the sensitiveness of explosives, and
2.1 Pneumatic Piston Test a selection of the most common units features in the Series 3
tests of the United Nations scheme for assessing safety in
Although designed initially as a compression test appara- transport(12).
tus, the pneumatic piston unit(6) was also used to examine the A similar test has also been employed to assess the
response of a range of initiators to dynamic (impact) forces. response of mechanically operated initiators to impact
Whereas manual compression forces are applied relatively stimuli. The method involves dropping a metal ball of
slowly and are to an extent distributed throughout an object, known mass from a range of heights on to a needle resting
impact forces are applied very rapidly and produce complex on top of the initiator. The purpose of the test is to determine
stress waves within the impacted object. Details of the testing the drop height that results in initiation but because of energy
Table 1. Range of Theoretical Impact Energies that can be Generated by the Pneumatic Piston Apparatus
Pneumatic cylinder pressure (MPa) 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Piston velocitya (m s 1) 0.77 0.96 1.13 1.28 1.41 1.52 1.61 1.69 1.74
Impact energyb (m J) 20.6 32.0 44.4 57.1 69.5 81.0 91.2 99.7 106
a
Derived from calibration graph of piston velocity against cylinder pressure obtained from measurements with a linear voltage
displacement transducer
b
Calculated as the kinetic energy of the piston
Table 2. Impact Test Results Obtained Using the Pneumatic Piston Apparatus
Initiator type Pneumatic cylinder Minimum piston velocity Minimum impact energy
pressure (MPa) for ignition (m s 1) for ignition (m J)
Boxer cap 0.30 1.13 44.4
Berdan cap, type 1 0.15 0.86 26.1
Berdan cap, type 2 0.15 0.86 26.1
Stab detonator (two discs),a type 1 0.15 0.86 26.1
Stab detonator (two discs), type 2:
Red disc impacted 0.15 0.86 26.1
Blue disc impacted 0.80 1.69 99.7
Stab detonator (one disc), type 1 0.20 0.96 32.0
Stab detonator (one disc), type 2 0.30 1.13 44.4
Flash sensitive detonator 0.85 1.72 103.2
Stab type igniter (two discs)b, type 1 0.35 1.20 50.8
Stab type igniter (two discs), type 2:
Small disc impacted 0.40 1.28 57.1
Large disc impacted 0.35 1.20 50.8
a
Red disc impacted, yellow disc gave no ignition
b
Brass disc impacted
176 R. K. Wharton, A. W. Train, P. F. Nolan, and S. C. Campbell Propellants, Explosives, Pyrotechnics 26, 174 179 (2001)
losses on recoil of the ball it is difficult to estimate the With estimated handling forces ranging from 37.2 55.9 N
minimum ignition energy and the method has become a and 23.5 33.3 N for male and female workers, respectively,
 go   no go means of monitoring consistent sensitivity in it is apparent that the test units are both easily capable of
production. reproducing those compression forces likely to be encoun-
More recently, these deficiencies have been examined by tered during manual handling. Similarly, the mean peak force
Woods et al.(13) and a modified unit has been produced with a application rates in Table 5 are significantly greater than the
more consistent performance and with instrumentation to estimated handling values in Table 4.
enable evaluation of the magnitude and rate of application of When examined against this background, the fact that no
forces by the dropped ball to the firing pin. ignitions were obtained in the test programmes involving
It is possible to calculate the  all-fire impact (i.e. poten- both test methods suggests that factors other than simple
tial) energies for certain initiators from knowledge of the ball compression may have been involved in the reported indus-
masses and drop heights. Table 3 summarizes the results and trial incidents. Other considerations such as the extensive-
also presents data for the ball velocity at impact evaluated ness of the test programme when applied to low probability
from the equation for terminal velocity. events, the general robustness of electrically initiated
detonators (which are designed to be electrically and not
mechanically initiated), and the rapid rate of force applica-
tion generally used to initiate stab sensitive initiators, could
3. Discussion
also have exerted an effect.
Although not part of the originally planned manual hand-
From our previous study of the ergonomic aspects of
ling programme, additional information obtained from the
handling small cylindrical objects(2) it is possible to estimate
impact test work has provided a means of assessing certain
the maximum compression forces and application rates that
could be applied by a typical operator to the 13 ini- hazards associated with the accidental functioning of initia-
tors. Thus the work has also provided data on peak sound
tiator devices examined in the compression studies(5,6).
pressure levels, the extent of fragment production on func-
Table 4 lists the values in units which can be directly
compared to the mean forces and mean peak force applica- tioning and the ability of some, but not all, detonators to be
functioned by an impact at either end(1).
tion rates for both the pivoted beam and pneumatic piston
apparatus, Table 5.
Table 3. Calculated Drop Ball Impact Energies
Initiator type Diameter of Mass All fire No fire All fire No fire Ball impact velocity,
striker tip (mm) of ball (g) height (mm) height (mm) energy (m J) energy (m J) all fire (m s 1)
Boxer cap 1.52 114 310 75 347 84 2.46
Berdan cap, type 1 2.9 57 203  114  1.99
Berdan cap, type 2 1.98 227 178 51 396 114 1.87
Stab detonator (1 disc) 0.08 0.20 28 254  70  2.23
Stab type initiator 0.08 0.20 28 152  42  1.73
(2 discs), type 1
Table 4. Maximum Compression Forces and Application Rates that could be Applied to a Range of Initiators, Estimated from Ergonomic
Study(2)
Initiator type Diameter Length Estimated maximum Estimated maximum
(mm) (mm) handling forces (N) application rates (N=ms)
Male Female Male Female
Berdan cap, type 1 4.54 2.23 37.2 23.5 0.088 0.059
Boxer cap 4.42 3.06 37.2 23.5 0.088 0.059
Berdan cap, type 2 5.49 2.91 45.1 27.4 0.098 0.069
Electric cap, conducting 11.25 6.55 55.9 33.3 0.127 0.078
composition
Stab detonator (2 discs), type 1 4.16 5.11 37.2 23.5 0.083 0.059
Stab detonator (2 discs), type 2 4.15 5.16 37.2 23.5 0.083 0.054
Stab detonator (1 disc), type 1 4.21 5.44 37.2 23.5 0.083 0.054
Stab detonator (1 disc), type 2 5.69 14.74 45.1 27.4 0.098 0.069
Flash sensitive detonator 4.20 5.44 37.2 23.5 0.083 0.054
IEBWa (instantaneous) 7.23 48.75 51.0 32.3 0.113 0.078
IEBWa (50 ms delay) 7.20 97.50 51.0 32.3 0.113 0.078
Stab igniter (2 discs), type 1 4.81 4.55 41.2 25.5 0.088 0.059
Stab igniter (2 discs), type 2 5.63 3.87 45.1 27.4 0.098 0.069
a
Instantaneous electric bridgewire detonator
Propellants, Explosives, Pyrotechnics 26, 174 179 (2001) Hazard in the Manual Handling of Explosives Initiator Devices 177
Table 5. Mean Forces and Mean Peak Force Application Rates Produced by the Two Compression Test Apparatuses
Pivoted beam apparatus Pneumatic piston apparatus
Mean Mean peak force Mean Mean peak force
force application rate (N=ms) force application rate (N=ms)
(N) (N)
RTa FPTb KETc RT FPT KET
13.4 0.39 0.63 0.56 18.4 0.94 0.61 0.56
32.9 0.89 0.97 1.11 33.3 1.29 0.92 0.92
52.5 1.29 1.29 1.41 47.7 1.56 1.24 1.25
69.4 1.95 1.76 1.79
81.7 1.80 1.62 1.64 85.9 1.23 1.11 1.11
141.9 3.15 2.15 2.44 150.6 1.6 1.75 1.70
216.8 4.02 3.05 2.92 213.1 2.22 2.22 2.27
286.8 5.24 3.50 3.56 275.9 2.77 2.78 2.88
a
Rubber-tipped tool
b
Firing pin tool
c
Knife-edge tool
The main findings, however, relate to the minimum impact significantly more energy to be supplied during compression
stimulus required for initiation: the lowest piston velocity to than during impact. The fact that no ignitions resulted in the
result in ignition was 0.86 m s 1 (equivalent to 0.15 MPa compression tests may indicate that initiation depends not
pressure), while the maximum velocity required was only on the magnitude of the energy but also on the rate at
1.72 m s 1 (a pressure of 0.85 MPa). which it is applied. Durand and co-workers(14) have pre-
For the electric cap containing conducting composition, a viously reached a similar conclusion.
maximum gas pressure of 0.9 MPa was used and this did not Generally, safety margins can be expressed as the differ-
produce an ignition. ence or the ratio of the maximum predicted value of a
The data summarized in Table 2 show clear differences in quantity and the minimum value of that quantity that will
the response of a range of initiation devices to an impact produce an event. For instance, for the handling of initiators
stimulus and this could provide useful input to a broader the ratio of the minimum handling stimulus required to
assessment of the hazards associated with the industrial initiate a device to the maximum stimulus that could be
handling of such items. exerted by a typical person handling the initiator would be a
Emphasis is often placed on determining the minimum useful measure of safety. However, since it was not possible
ignition stimulus for initiators in terms of the transmitted to establish the minimum stimulus required to initiate any of
energy with a view to comparing this with the maximum the devices that were tested, this ratio can be usefully
energy likely to be applied accidentally, thus gaining an modified to involve comparison of the maximum handling
estimate of the margin of safety. Unfortunately there is little stimulus applied to an initiator that does not result in ignition
knowledge relating to how much energy would actually be with the maximum stimulus that could be exerted by a typical
imparted to an initiator during manual handling and hence it person when handling the initiator.
is not possible to base a system of safety margins for the Although the experimental work was unable to define the
handling of detonators on minimum ignition energies even input stimulus in terms of energy, results were available for
though these data are often available. both the maximum force applied to an initiator during testing
Comparison of the all fire energies for drop ball tests and the peak rate at which that force was applied. It is
listed in Table 3 with the impact energies in Table 2 therefore possible to define a compression force safety
indicates that for percussion caps the latter were lower, margin factor (CFS) as the ratio of the maximum force
whereas for the stab igniters and stab detonators this applied to an initiator by a test apparatus that does not
situation is reversed. This can be ascribed to the differences produce an ignition to the maximum force that a typical
in the tip sizes of the firing pin in the drop ball test and the adult could apply to that initiator. Similarly, a compression
firing pin tool in the pneumatic piston apparatus. The firing rate safety margin factor (CRS) can be defined as the ratio of
pin tool was of similar diameter to the firing pins used in the the maximum rate at which a force is applied to an initiator by
drop ball tests on percussion caps but much larger in a test apparatus that does not produce an ignition to the
diameter than the pins used in drop ball tests with the maximum rate at which a typical adult could apply that force
stab igniters and detonators. to that initiator.
The minimum impact energies in Table 2 can also be Values of CFS and CRS for both males and females
compared with the maximum compression energies gener- handling detonators examined in the compression studies
ated using the firing pin tool in the pivoted beam test. were evaluated for both the pivoted beam and pneumatic
A typical example is the minimum impact energy of piston apparatus. The results are summarized in Tables 6 and
26.1 mJ for the stab detonator with two discs (red impacted) 7, respectively.
compared with the maximum imparted compression of Whereas stab detonators appear to have the greatest
192 mJ: apart from one isolated case, the trend was for margin of safety from the data presented in these tables,
178 R. K. Wharton, A. W. Train, P. F. Nolan, and S. C. Campbell Propellants, Explosives, Pyrotechnics 26, 174 179 (2001)
consideration should also be given to the fact that all the to an operator. The effect of blast, but not fragment hazard,
initiators were subjected to the same compression forces can be roughly ranked by comparing the overpressures
during testing and that none of the devices were initiated in derived from the peak sound pressure levels generated at a
the tests. The variations in the values of CFS and CRS can be fixed point from the detonators when they are functioned.
largely attributed to the differences in the estimated handling Table 8 presents the data derived from such measurements
forces for each initiator as a result of their different sizes. when ranked relative to the performance of one of the Berdan
Bailey and Thomson(15) have discussed the essential type caps. The results indicate that the stab detonators and
factors that need to be considered when developing a igniters are more powerful than the percussion caps.
hazard index for energetic materials (including explosives). Although it has not been possible to derive a safety-in-
In a later study(16), the range of different stimuli, were handling hazard index directly from the present study, data
assigned different classifications according to whether they for certain parameters that are important in assessing the
were very sensitive, sensitive or comparatively insensitive. safe handling of initiators have been obtained, and these
The results obtained correlated well with documented could form an important input to future hazard index
records of industrial accidents involving explosives. development.
Thus the CFS and CRS could be used in assessing the
sensitiveness of initiators to handling forces and form part of
a wider safety-in-handling index. 4. Conclusions
One of the additional factors that would need to be
incorporated into such an index is an assessment of the It has been shown that the sensitiveness of initiators is
consequences of an initiator exploding in close proximity affected by the rate at which energy is transferred to them.
Table 6. Compression Force and Compression Rate Safety Margin Factors Evaluated from Tests with the
Pivoted Beam Apparatus
Initiator type CFS CRS
Male Female Male Female
Electric cap, conducting composition 5.1 8.6 41.3 67.2
IEBW, instantaneous
5.6 8.9 46.4 67.2
IEBW, delay
Stab type igniter (two discs), type 1
6.4 10.5 53.5 75.9
Stab detonator (one disc)
Stab type igniter (two discs), type 2 7.0 11.2 59.5 88.8
Berdan cap, type 1 7.7 12.2 59.5 88.8
Stab detonator (two discs), type 1
Stab detonator (two discs), type 2
7.7 12.2 63.1 97.0
Stab detonator (one disc)
Flash sensitive detonator
Table 7. Compression Force and Compression Rate Safety Margin Factors Evaluated from Tests with the
Pneumatic Piston Apparatus
Initiator type CFS CRS
Male Female Male Female
Electric cap, conducting composition 4.9 8.3 21.8 35.5
IEBW, instantaneous
5.4 8.5 24.5 35.5
IEBW, delay
Stab type igniter (two discs), type 1
6.1 10.1 28.3 40.1
Stab detonator (one disc)
Stab type igniter (two discs), type 2 6.7 10.8 31.5 46.9
Berdan cap, type 1 7.4 11.7 31.5 46.9
Stab detonator (two discs), type 1
Stab detonator (two discs), type 2
7.4 11.7 33.4 51.3
Stab detonator (one disc)
Flash sensitive detonator
Propellants, Explosives, Pyrotechnics 26, 174 179 (2001) Hazard in the Manual Handling of Explosives Initiator Devices 179
Table 8. Relative Explosive Power of the Initiators Using Blast Overpressure Derived from Peak Sound
Pressure Levels
Initiator type NEQ (mg) Estimated relative
explosive power
Berdan cap, type 1 16 1.0
Berdan cap, type 2 32 1.4
Boxer cap 22 7.4
Stab type igniter (two discs), type 1 136 10.6
Stab detonator (two discs), type 1
(estimated) 130 25.5
Stab detonator (two discs) type 2
Stab type igniter (two discs), type 2 104 26.3
Stab detonator (one disc), type 1 150 30.2
Flash sensitive detonator 146 36.7
Stab detonator (one disc), type 2
195 (estimated) 193.0
Electric cap, conducting composition
(6) R. K. Wharton, A. W. Train, P. F. Nolan and S. C. Campbell,
However, the study was unable to determine whether or not
  The Development and Application of a Pneumatic Piston Test
the range of initiators examined was actually sensitive to
for Assessing Safety in the Handling of Initiator Devices  ,
manual handling forces.
Propellants, Explosives, Pyrotechnics 26, 112 117 (2001).
(7) A. J. Barratt, Health and Safety Laboratory, unpublished results.
It was not possible to evaluate the quantity of energy
(8) Y. Ye, R. Shen and S. Dai,   Statistical Analysis of Explosive
imparted to a detonator on handling, nor could the minimum
Accidents  , 18th International Pyrotechnics Seminar, Breck-
impact energies or drop ball energies be used to estimate the
enridge, Colorado, 13 17 July 1992, pp. 1019 1026.
minimum compression energy for initiation. Hence, useful
(9) M. Nabiullah, R. N. Gupta and B. Singh,   Impact Sensitivity and
Thermal Behaviour of Commercial Detonators  , 15th Interna-
safety margins based on energy input are unable to be derived
tional Pyrotechnics Seminar, Boulder, Colorado, 9 13 July 1990,
and safety margins are best estimated from either the forces
pp. 743 755.
applied to the initiators or the peak rate at which they were
(10) M. Kimura, N. Izawa and M. Goto,   Impact Sensitivity of
applied.
Electric Detonators  , J. Ind. Explos. Soc. Japan 38, 216 (1977).
(11) K. R. Beaker,   Input and Thermal Sensitivity of Commercial
Our recent work in this area, which is reviewed in this
Detonators  , US Bureau of Mines Report RI 8085 (1975).
paper, has provided data on the handling forces that males
(12)   Recommendations on the Transport of Dangerous Goods: Tests
and females can apply to small objects, the sensitiveness of
and Criteria  , 3rd rev. ed., ST=SG=AC.10=11=Rev. 3, United
a range of initiators when exposed to such forces, and an Nations, New York and Geneva, 1999.
(13) C. M. Woods, M. A. Robinson, C. W. Merten, V. E. Robbins and
appreciation of their explosive power. These parameters may
D. R. Begeal,   Instrumented Drop Ball Tester for Percussion
provide a useful input to the development of future safety-in-
Primers  , 16th International Pyrotechnics Seminar, Jönköping,
handling hazard indexes.
Sweden, 24 26 June 1991, pp. 902 914.
(14) N. A. Durand, R. R. Weinmaster and T. M. Massis,   Hermetic
G-16 Percussion Primer  , 13th International Pyrotchnics Semi-
5. References
nar, Grand Junction, Colorado, 11 15 July 1988, pp. 197 207.
(15) A. Bailey and B. J. Thomson,   Is a  Hazard Index for Explo-
(1) S. C. Campbell,   Development of an Apparatus to Help Assess sives and Energetic Materials Achievable?  , 14th International
the Sensitiveness of Explosive Initiating Devices to Manual Pyrotchnics Seminar, Jersey, Channel Islands, 18 22 September
Handling Forces  , Ph.D. thesis, South Bank University, 1999. 1989, pp. 347 354.
(2) S. C. Campbell, P. F. Nolan, R. K. Wharton and A. W. Train, (16) A. Bailey, D. Chapman, M. R. Williams and R. Wharton,   The
  Measurement of Forces Exerted in the Manual Handling of Handling and Processing of Explosives  , 18th International
Small Cylindrical Objects  , Int. J. Indust. Ergon. 25, 349 (2000). Pyrotechnics Seminar, Breckenridge, Colorado, 13 17 July 1992,
(3) S. C. Campbell, P. F. Nolan, R. K. Wharton and A. W. Train, pp. 33 49.
  Development of an Apparatus to Assess the Safety in Handling
of Initiating Devices  , 21st International Pyrotechnics Seminar,
Acknowledgements
Moscow, 11 15 September 1995, pp. 109.
The authors would like to thank the safety officers and production
(4) S. C. Campbell,   Research to Develop a Method of Assessing
staff of those companies in the UK explosives industry that provided
Safety in the Handling of Initiating Devices  , Explosives Engi-
assistance to the research reported in this paper. South Bank University
neering, March 1996, p. 12.
staff would like to thank the Health and Safety Executive for financial
(5) R. K. Wharton, A. W. Train, P. F. Nolan and S. C. Campbell,
support.
  The Development and Application of a Pivoted Beam Test for
Assessing Safety in the Handling of Initiator Devices  , Pro-
pellants, Explosives, Pyrotechnics 26, 84 90 (2001). (Received October 9, 2000; Ms 2000=044)