An earlier draft appeared in Journal of Pyrotechnics, No. 5, 1997.

A Survey of Concussion Powders

K. L. and B. J. Kosanke
PyroLabs, Inc., 1775 Blair Road, Whitewater, CO 81527, USA

ABSTRACT It should be noted that the purpose of this

study was to measure the performance characteris-
A collection of six commercial concussion tics of the concussion powders. The purpose was
powders were test fired in concussion mortars to not to rate the performance of the concussion
determine internal mortar pressure, air blast powders. To rank the concussion powders, it
pressure, and the duration of air blast positive would first be necessary to establish a set of eval-
phase. The internal mortar pressures for various uation criteria; however, those criteria are likely to
powder types and load masses ranged from less be different for various users with different appli-
than 200 to nearly 100,000 psi (700 MPa). For cations. Hopefully, this study provides basic in-
the same powder loads, the air blast pressures at formation that both users and manufacturers will
a distance of approximately 70 in. (1.8 m), ranged find useful, or at least interesting.
from 0.07 to 1.7 psi (12 kPa). This corresponds to
sound pressure levels (peak—ultra fast—linear)
ranging from 148 to 175 dB, and relative loudness Background
values ranging from 1.0 to 6.8. For the same
powder loads, the durations of positive phase In its most common form, a concussion mortar
ranged from nearly 4 ms down to 0.7 ms. consists of a thick, cylindrical steel bar, welded to
a heavy base plate. The mortar contains a combus-
Keywords: concussion powder, blast wave, mor- tion chamber (barrel), produced by drilling a hole
tar pressure, sound pressure level, loudness on-axis into the top end of the steel bar. The mor-
tar used in this study was 2 in. (50 mm) in outer
diameter, with a 1-in. (25-mm) hole drilled to a
Introduction
depth of 4.5 in. (115 mm). The construction of the
In an earlier article,[1] the results of a fairly de- mortar is illustrated in Figure 1, which also shows
tailed study of one commercial concussion pow- it loaded with a charge of powder and an electric
der (Pyropak™) were reported. The present article match for ignition.
reports on a study of a collection of concussion Upon ignition, because of the confinement
powders from various manufacturers. Since it was provided within the combustion chamber, the
not practical to repeat the full study for each of the concussion powder burns explosively, see Figure
other concussion powders, it was decided to com- 2. The high internal pressure causes the combus-
pare the performance of the powders at only a few tion products (gases and solid particles) to be ac-
selected load masses. All suppliers of concussion celerated upward. As the gases exit the end
powder known to the authors were contacted; all (mouth) of the mortar, they expand to produce a
but one agreed to participate in the study and pro- shock wave that is heard and felt by the audience.
vided samples of their powders. As a result of the ejection of combustion products,
Since beginning this study, the catastrophic a downward recoil force is produced. In a previ-
failure of a concussion mortar (without injury) has ous study, it was demonstrated that the shape of
been reported.[2] This increased both the interest the recoil force curve follows the shape of the in-
and the relevance of this study, and provided the ternal pressure curve.[1a] Further, it was demon-
impetus for an early release of some of the re- strated that the magnitude of the recoil force de-
sults.[3] The present article is a more thorough pended on the rigidity of the surface under the
presentation and discussion of those results. mortar. Accordingly, the value of collecting recoil
data is diminished and that data was not generated
in this study.

Page 402 Selected Pyrotechnic Publications of K.L. and B.J. Kosanke

 ambient pressure. This portion of the blast wave is
referred to as the positive phase. Following the
positive phase, there usually is a negative phase,
during which pressure drops below ambient. In
Explosion essence, this is caused by over expansion of the
Chamber gases, wherein the outward rush of air continues
Concussion beyond that necessary to relieve the pressure pro-
Flash Electric
Match duced by the explosion. Thus, a partial vacuum
Powder
Leg Wires forms at the seat of the explosion, producing the
negative phase of the blast wave. It is less extreme
Concussion than the positive phase and lasts longer.
Mortar
Base Plate

Figure 1. An illustration of the construction and

setup of a concussion mortar.

Figure 3 illustrates a typical blast overpressure

profile. Before the arrival of the blast wave at the
pressure sensor, there is no indication (with re-
spect to pressure) that an explosion has taken
place or that the blast wave is approaching. When
the leading edge of the blast shock wave arrives, it Figure 3. An illustration of a typical overpressure
produces an essentially instantaneous rise in pres- profile (blast wave) produced by an explosion.
sure from ambient to some maximum value.
Thereafter, the pressure gradually decays back to
Sound pressure level (SPL, in decibels, dB) is
a physically measurable quantity and can be cal-
culated from blast overpressure measurements
using the relationship shown in equation 1.[4,5][a]
Expanding Exiting High Using the recognized standard reference level of
Gases Velocity 0.0002 dyn/cm2 and converting pressure to pounds
Gases Expand
per square inch (psi), this becomes equation 2.
P
High Pressure
Gases Move
SPL = 20 ⋅ log (1)
Po
Upward

SPL = 170.8 + 20 log P (2)

Explosive where Po is the standard reference value of 0.0002
Burning of
Powder Load
dyn/cm2. As can be seen, there is a logarithmic
relationship between sound pressure level and
peak overpressure (P).
Loudness is a subjective measure of sound lev-
el, dependent on the processing of nerve impulses
by the brain. The loudness scale is linear, such
Downward
that a sound with a loudness value twice that of
Recoil Force
another sound will be perceived by a typical lis-
tener to be twice as loud. Loudness values (N, in
Figure 2. An illustration of the firing of a concus- sones) can be calculated from sound pressure lev-
sion mortar. els using the relationship shown in equation 3.[5]

Selected Pyrotechnic Publications of K.L. and B.J. Kosanke Page 403

 log N = 0.03 Ln − 1.2 (3) loudness and tonal quality, which, however, is in
the opposite direction as that reported for pure
N  10(
0.03 SPL −1.2 )
(4) tones.[6,7] Accordingly, for the purpose of this arti-
cle, it will only be assumed that tonal quality cor-
where Ln, loudness level, is equal to SPL in dB for relates with phase duration, at least for equally
a 1000 Hz tone. Thus, for this study equation 4 loud sounds.
was used to calculate approximate loudness.
Because loudness expressed in sones is not a Experimental Method[b]
unit of measure with which many readers are ac-
customed, in this article, loudness is reported as Table 1 lists information regarding the binary
relative loudness. The least loud average concus- concussion powders (so-called A–B mixes) used
sion mortar air blast was assigned a value of 1.0. in this study. The powders are listed in order of
Thus a concussion mortar blast reported as pro- the internal mortar pressures they produced, from
ducing sound with a relative loudness of 2.0 or 5.0 the lowest to highest pressure. Most of the infor-
will be one perceived by the average listener to be mation in the table was gathered from Material
approximately two or five times as loud, respec- Safety Data Sheets and user instructions provided
tively. by the manufacturers. One exception is the fuel to
In addition to the loudness of a concussion oxidizer ratios that were obtained by simply
mortar blast, the tonal quality of the sound may weighing the contents of the containers for the
also be of interest. That is to say, does the sound two components and rounding to the nearest 5%.
produced tend toward being a sharp crack or a Not included in Table 1 is information about par-
more mellow boom? The feature of a blast wave ticle size of the components; it was felt this would
that is thought to correlate with perceived tonal be proprietary information of the manufacturers. It
quality is the duration of the positive and negative should also be noted that some of these samples
phases. All else being equal, shorter phase dura- were provided several years ago; thus, it is possi-
tions are expected to be heard more nearly as ble that the manufacturers will have made changes
sharp cracks, and longer phases as more mellow in their formulations, which are not reflected in
booms. There are at least two reasons for being this study. Further, Astro Pyrotechnics has an-
somewhat cautious about proclaiming that tonal nounced that they have discontinued the general
quality correlates with phase duration. First is that sale of their concussion powder.
tonal quality is a subjective (mental) response to a The test concussion mortar had been modified,
physical stimulus (the blast pressure wave), and as shown in Figure 4, to allow attachment of a
the brain may not process this information as one quartz piezoelectric pressure transducer for re-
might expect that it might.[6] Second is that the cording internal pressure during its firing. Unfor-
authors are not aware of any comprehensive study tunately, in retrospect, this was not the ideal con-
of perceived tonal quality as functions of blast figuration. The length of passage to the pressure
wave phase duration.
It is also conjectured that the rate of rise of the
leading edge of the positive phase (dP/dt) may
affect tonal quality. It is expected that sharper ris-
es may be perceived as being sharper sounding.
However, at the short distance at which measure-
ments of the blast waves were made in this study,
all had a near instantaneous pressure rise.
While on the subject of tonal quality of blast
waves, it is appropriate to mention that the subject
is made more complicated because a complex re-
lationship has been demonstrated between per-
ceived loudness and tonal quality for pure tones.[6]
Further, the results for a brief study of spectator
responses to the sounds produced by fireworks Figure 4. An illustration showing the installation
salutes, suggests a strong correlation between of the pressure transducer.

Page 404 Selected Pyrotechnic Publications of K.L. and B.J. Kosanke

Table 1. Concussion Powder Information.

Fuel : Ox. Loading

Supplier and Product Name Fuel(a) Oxidizer(a) Ratio(b) Instructions(c)
Luna Tech (Pyropak®) Strontium 1 oz. (28 g)
Magnesium 50 : 50
Concussion Flash powder Nitrate (maximum)
Newco Products Potassium 14 g (1/2 oz.)
Magnesium 70 : 30
Fast Theatrical Flash Powder Perchlorate (maximum)
MP Associates Potassium
Magnesium 60 : 40 Not Specified
Super Flash Powder Perchlorate
Astro Pyrotechnics Potassium
Magnesium 50 : 50 1 cap full(d)
Sound Flash Powder Perchlorate
Theatre Effects Potassium 1/4 tsp.(e)
Aluminum 30 : 70
Sonic Chemical Perchlorate (typical)(f)
Precision Theatrical Potassium (g)
Aluminum 30 : 70
Concussion Perchlorate
(a) As specified on manufacturer’s Material Safety Data Sheet.
(b) Determined by weighing the contents of a single pair of bottles supplied for testing. The ratios are rounded to
the nearest 5%.
(c) As specified in the manufacturer’s instructions supplied with the product.
(d) One cap full is approximately 2.7 grams.
(e) One-quarter teaspoon full is approximately 1.7 grams and one heaping quarter teaspoon is approximately 2.5
grams.
(f) There is the additional instruction, “Increase ... slightly if ... not loud enough.”
(g) The supplier reports that the powder is supplied with a recommendation to use a mortar with a 1-inch (2.5-
cm) wall thickness.

gauge should have been shorter and more im- (Presumably, this related to having a long path
portantly it should not have had a 90º bend. None- with a 90º bend between the chamber and the
theless, this is the configuration that was available gauge). To develop a pressure correction curve, a
and that was used. The overall results of this com- series of twenty-four firings were performed using
parative study should not have been significantly various amounts of Pyropak concussion powder.
affected by using this configuration. These results were compared with data collected
previously[1b] in which the passageway had been
The pressure gauge was a PCB Piezotronics
filled with a lightweight silicone grease. The cor-
(Model 109A02) calibrated to 120,000 psi
rection factors produced are the average ratio of
(830 MPa). To protect the gauge and to keep con-
the values obtained with and without the sealant.
cussion powder out of the passageway—from the
These values were plotted in Figure 5 and a
combustion chamber to the pressure transducer—
smooth curve drawn through the data points.
the open space was filled with an opaque silicone
sealant (Permatex, High Temp RTV, #26B). The Air blast overpressures were measured using a
sealant was allowed to cure for at least a week PCB Piezotronics free-field blast gauge (Model
before use. Occasionally during the testing of 137A12), setup as shown in Figure 6. This geome-
concussion powders that produced the highest try was chosen to duplicate that used in an earlier
internal pressures, the silicone sealant loosened study,[1] which had been chosen for convenience
and was eroded. On those occasions, the sealant and because it seemed a reasonable choice. The
was removed, replaced and allowed to cure before pressure sensor was shielded from thermal radia-
testing continued. tion by a thin film of silicon grease that was cov-
ered tightly with a 0.001-in. (0.025-mm) film of
When measuring pressures below approxi-
aluminized Mylar.
mately 2000 psi (14 MPa), the silicone sealant
acts to attenuate the pressure sensed by the gauge.

Selected Pyrotechnic Publications of K.L. and B.J. Kosanke Page 405

 Permanent storage and plotting of the data was
2.4
Correction Factor
accomplished using a computer.
2.0 Typically each concussion powder was tested
with loads of 7, 14, 21, and 28 g. However, when
1.6
any concussion powder load mass produced an
internal pressure that approached 100,000 psi
1.2
(700 MPa) for any individual firing, or if a series
1.0
of firings produced pressures that averaged more
than 30,000 psi (200 Mpa), no greater loads of
0.8
0 500 1000 1500 2000 2500 3000 that powder were tested. When these values were
exceeded for light powder loads, additional tests
Observed Pressure (psi)
with various low mass loads were performed.
Figure 5. Graph of factors used to correct mortar
pressures for the presence of silicone sealant. Results
In this section, the results of the test firings are
Overpressure reported without comment. (The discussion of the
Signal results is deferred until the next section.) The re-
Free Field sults from the individual firings are reported in
Blast Gauge Tables 2 through 7. Typically, there were three
Concussion test firings for each load mass with each powder.
36"
Mortar However, in some cases additional firings were
conducted. Sometimes this was because of a fail-
60" ure to successfully capture both types of pressure
18"
data for reasons such as data being off-scale. Oth-
(ground) er times this was the result of some initial testing
being performed in a test chamber in which it was
felt that there was insufficient space above the
Figure 6. An illustration of the physical setup
concussion mortar to collect reliable air blast data.
used to collect concussion mortar air blast over-
[This was only a problem for those concussion
pressures. (For conversion of units, 1 in. = 25.4 mm)
powders with formulations that were particularly
fuel rich (i.e., Pyropak’s and Newco’s.)] This is
The electric matches used to ignite the concus- discussed further in the next section.
sion powders in this study were Daveyfire SA-
In Tables 2 through 7, the column headed “Pres.”
2000. The electric matches were installed near the
presents the data from internal mortar pressure
bottom of the powder charge (see Figure 1). The
measurements. The column “FWHM” is an ab-
procedure was to insert the match until it touched
breviation for Full Width at Half Maximum. This
the bottom of the mortar, and then withdraw it
is simply the width of the internal pressure peak,
approximately 1/8 in. (3 mm). Over the course of
measured at one half of its peak value. It is an in-
these and earlier tests, the diameter of the electric
dication of the width of the pressure peaks. When
match hole in the mortar had eroded quite large, to
the pressure curves have a simple shape (see Fig-
an irregular diameter of approximately 0.25 in.
ure 7B in the next section), FWHM is a reasona-
(0.5 cm). In addition, prior repeated use of the
bly good indicator of relative peak width. Howev-
mortar had also eroded the bore of the combustion
er, when the pressure peaks have an irregular or
chamber to approximately 1.05 in. (2.7 cm). It is
complex shape (see the other curves of Figure 7)
likely these erosions caused the measured internal
this is a less reliable indicator, but it is still some-
mortar pressures to be slightly less than might
what useful. The column titled “P. Imp.” presents
otherwise have been the case.
pressure impulse data, the area under the internal
Output from both PCB transducers (internal mortar pressure versus time curves. The column
mortar pressure and blast overpressure) were fed headed “Blast” reports peak air blast overpressure
to amplifying power supplies (PCB Model results. The column titled “Pos. Ph.” presents the
480D09) and recorded using digital oscilloscopes. duration of the positive phase portion of the air
blast wave. Finally, the column headed “B. Imp.”

Page 406 Selected Pyrotechnic Publications of K.L. and B.J. Kosanke

presents the blast impulse, the area under the posi-
tive phase portion of the air blast overpressure
curve.

Table 2. Results of Measurements Using Pyropak’s “Concussion Flash Powder”.

Powder Pres. FWHM P. Imp. Blast Pos. Ph. B. Imp.

Load (g) (psi) (ms) (psi⋅s) (psi) (ms) (psi⋅ms)
7 190 7.4 1.4 0.07 4.4 0.11
7 180 7.8 1.1 0.08 3.4 0.11
7 180 8.8 1.2 0.05 3.4 0.08
14 1000 2.0 2.0 — — —
14 1600 2.3 4.3 — — —
14 630 4.6 2.9 0.21 4.1 0.33
14 1200 2.3 3.1 0.21 3.8 0.46
14 1500 1.9 3.1 0.46 2.9 0.64
21 3500 1.8 6.4 — — —
21 2700 2.4 6.3 — — —
21 2800 1.6 4.4 0.88 2.1 0.75
21 1600 2.0 3.3 0.87 2.1 0.86
21 1300 1.8 2.3 0.72 2.1 0.81
28 3300 2.2 7.3 — — —
28 4000 1.7 7.5 — — —
28 2700 2.2 4.9 1.1 2.1 0.99
28 2400 1.9 4.7 1.3 2.3 1.1
28 2300 1.9 4.6 1.5 1.5 1.0
(For conversion of units: 1 psi = 6.89 kPa, and 28 g ≈ 1 oz.)

Table 3. Results of Measurements Using Newco Products “Fast Theatrical Flash Powder”.

Powder Pres. FWHM P. Imp. Blast Pos. Ph. B. Imp.

Load (g) (psi) (ms) (psi⋅s) (psi) (ms) (psi⋅ms)
7 1400 1.2 1.7 0.55 2.2 0.48
7 1500 1.0 1.6 0.56 1.5 0.55
7 1400 1.2 1.7 0.58 1.1 0.52
7 — — — 0.48 2.1 0.39
14 2800 1.3 3.5 — — —
14 2800 1.3 3.3 — — —
14 4800 0.84 3.6 — — —
14 2600 1.5 3.7 1.0 1.5 0.70
14 3100 1.2 3.7 1.3 1.5 0.68
14 2800 1.4 3.6 1.0 1.3 0.71
21 3200 1.4 4.4 1.6 1.3 0.85
21 4900 1.2 5.7 1.6 1.3 0.88
21 3600 1.8 5.7 1.5 1.6 0.93
28 5100 1.3 6.9 1.8 1.3 0.99
28 5800 1.1 7.1 1.9 1.3 1.1
28 4400 1.7 6.8 1.6 1.3 1.1
(For conversion of units: 1 psi = 6.89 kPa, and 28 g ≈ 1 ounce.)

Selected Pyrotechnic Publications of K.L. and B.J. Kosanke Page 407

Table 4. Results of Measurements Using MP Associates’ “Super Flash Powder”.
Powder Pres. FWHM P. Imp. Blast Pos. Ph. B. Imp.
Load (g) (psi) (ms) (psi⋅s) (psi) (ms) (psi⋅ms)
7 1600 1.4 2.0 0.54 1.8 0.44
7 1700 1.2 2.1 0.81 1.5 0.44
7 1700 1.2 2.1 0.71 1.4 0.45
14 7800 0.50 4.0 1.4 1.2 0.71
14 5700 0.70 3.7 1.5 1.1 0.75
14 5600 0.58 3.7 1.5 1.1 0.75
21 12000 0.48 5.8 1.9 1.2 0.94
21 9900 0.56 5.6 2.0 1.1 0.96
21 13000 0.42 5.6 1.8 1.1 0.96
28 16000 0.30 6.3 2.3 0.94 1.1
28 6700 1.0 7.1 2.1 1.0 1.1
28 15000 0.12 6.5 2.0 0.96 1.1
(For conversion of units: 1 psi = 6.89 kPa, and 28 g ≈ 1 ounce.)

Table 5. Results of Measurements Using Astro Pyrotechnics’ “Sound Flash Powder”.

Powder Pres. FWHM P. Imp. Blast Pos. Ph. B. Imp.
Load (g) (psi) (ms) (psi⋅s) (psi) (ms) (psi⋅ms)
7 3400 0.56 1.9 1.2 0.92 0.40
7 3600 0.58 2.0 1.2 0.76 0.39
7 3600 0.52 1.9 1.2 0.84 0.41
14 6600 0.51 3.0 1.9 0.86 0.79
14 18000 0.12 3.9 1.6 0.84 0.63
14 7400 0.44 3.8 1.7 0.83 0.61
21 34000 0.14 4.4 2.1 0.87 0.87
21 12000 0.31 4.4 2.1 0.88 0.86
21 47000 0.17 4.3 2.1 0.87 0.87
(For conversion of units: 1 psi = 6.89 kPa, and 28 g ≈ 1 ounce.)

Table 6. Results of Measurements Using Theatre Effects’ “Sonic Chemical” Powder.

Powder Pres. FWHM P. Imp. Blast Pos. Ph. B. Imp.
Load (g) (psi) (ms) (psi⋅s) (psi) (ms) (psi⋅ms)
3 3700 0.16 0.89 1.0 0.58 0.21
3 1900 0.30 0.71 0.81 0.47 0.19
3 2100 0.27 0.72 0.78 0.58 0.20
5 4100 0.24 1.2 1.1 0.60 0.32
5 4400 0.22 1.3 1.2 0.65 0.34
5 4500 0.22 1.3 1.2 0.65 0.34
7 5600 0.25 1.8 1.3 0.70 0.43
7 3700 0.21 1.9 1.4 0.70 0.43
7 18000 0.04 2.0 1.3 0.71 0.43
9 18000 0.10 1.9 1.5 0.78 0.52
9 13000 0.12 2.1 1.7 0.78 0.57
9 19000 0.11 2.3 1.7 0.78 0.60
11 44000 0.03 2.6 1.6 0.79 0.61
11 — — — 1.7 0.80 0.65
11 92000 0.05 4.6 1.7 0.81 0.66
(For conversion of units: 1 psi = 6.89 kPa, and 28 g ≈ 1 ounce.)

Page 408 Selected Pyrotechnic Publications of K.L. and B.J. Kosanke

Table 7. Results of Measurements Using Precision Theatrical’s “Concussion” Powder.

Powder Pres. FWHM P. Imp. Blast Pos. Ph. B. Imp.

Load (g) (psi) (ms) (psi⋅s) (psi) (ms) (psi⋅ms)
2 1300 0.28 0.32 0.62 0.56 0.18
2 4000 0.10 0.38 0.81 0.50 0.21
2 1400 0.20 0.33 0.70 0.50 0.20
2 2000 0.15 0.34 0.77 0.50 0.21
3 5400 0.10 0.61 1.0 0.58 0.27
3 4900 0.11 0.58 1.0 0.60 0.27
3 6300 0.11 0.58 1.1 0.66 0.27
4 9400 0.10 0.91 1.1 0.60 .033
4 9900 0.07 0.88 1.2 0.60 0.33
4 8000 0.08 0.90 1.4 0.63 0.35
5 9500 0.12 1.2 1.4 0.62 0.39
5 19000 0.09 1.2 1.5 0.68 0.43
5 5000 0.19 1.0 1.4 0.63 0.40
6 6000 0.18 1.2 1.3 0.70 0.42
6 11000 0.10 1.2 1.4 0.74 0.44
6 7200 0.17 1.2 1.2 0.71 0.40
7 11000 .012 1.7 1.4 0.74 0.47
7 17000 0.08 1.5 1.3 0.74 0.46
7 19000 0.06 1.8 1.3 0.74 0.47
9 14000 0.15 2.1 1.4 0.77 0.52
9 15000 0.09 2.1 1.4 0.77 0.52
9 86000 0.02 3.1 1.5 0.77 0.55
(For conversion of units: 1 psi = 6.89 kPa, and 28 g ≈ 1 ounce.)

At the end of this section, Table 8 presents av-

erages for the test results from Tables 2 through 7.
In addition, average sound pressure levels and
relative loudness values are reported in the col-
umns titled “SPL” and “Rel. Loud.”, respectively.
These were calculated from the peak overpressure
data using equations 2 and 4.
In Table 8, because of the small number of test
firings for each load mass, and because of the
large variations observed in the individual results,
it was felt to be inappropriate to report standard
deviations. One reason for including the results
from individual test firings (Tables 2 through 7) is
that a simple inspection can provide a rough esti-
mate of the variability of the results. Note that
relatively little of the variability is thought to be
the result of the measurement process, but rather it
is from actual differences in the combustion pro-
cesses from test to test. A consequence of the sig-
nificant variability of the results is that they are
only reported to two significant figures.

Selected Pyrotechnic Publications of K.L. and B.J. Kosanke Page 409

Table 8. Average Results for the Various Sources of Concussion Powder.

Powder Pres. FWHM P. Imp. Blast Pos. Ph. B. Imp. SPL Rel.
Load (g)(a) (psi) (ms) (psi⋅s) (psi) (ms) (psi⋅ms) (dB) Loud.
LP 7 180 8.0 1.2 0.07 3.7 0.10 148 ≡1.0
LP 14 1200 2.6 3.1 0.29 3.6 0.47 160 2.3
LP 21 2500 1.9 4.7 0.82 2.1 0.81 169 4.4
LP 28 2900 2.0 5.8 1.3 2.0 1.0 173 5.8
N 7 1400 1.1 1.7 0.54 1.7 0.48 165 3.4
N 14 3300 1.3 3.6 1.1 1.4 0.70 172 5.2
N 21 3900 1.5 5.3 1.6 1.4 0.89 175 6.5
N 28 5100 1.4 6.9 1.8 1.3 1.1 176 7.0
MP 7 1700 1.3 2.1 0.69 1.6 0.44 168 4.0
MP 14 6400 0.59 3.8 1.5 1.1 0.74 174 6.3
MP 21 12000 0.49 5.6 1.9 1.1 0.95 176 7.3
MP 28 13000 0.47 6.6 2.1 0.97 1.1 177 7.7
A 7 3500 0.55 1.9 1.2 0.84 0.40 172 5.5
A 14 11000 0.36 3.6 1.7 0.84 0.68 175 6.8
A 21 31000 0.17 4.3 2.1 0.87 0.87 177 7.7
(b) (b) (b) (b) (b) (b) (b) (b)
A 28
TE 3 2600 0.24 0.77 0.86 0.54 0.20 170 4.5
TE 5 4300 0.23 1.3 1.2 0.63 0.33 172 5.5
TE 7 10000 0.17 1.9 1.3 0.70 0.43 173 5.7
TE 9 17000 0.11 2.1 1.6 0.78 0.56 175 6.5
TE 11 68000 0.04 3.6 1.7 0.80 0.64 175 6.8
(b) (b) (b) (b) (b) (b) (b) (b)
TE 13
PT 3 5500 0.11 0.59 1.0 0.61 0.27 171 4.9
PT 5 11000 0.13 1.1 1.4 0.64 0.41 174 6.0
PT 7 16000 0.09 1.7 1.3 0.74 0.47 173 5.7
PT 9 38000 0.09 2.4 1.4 0.77 0.53 174 6.0
(b) (b) (b) (b) (b) (b) (b) (b)
PT 11
(For conversion of units: 1 psi = 6.89 kPa, and 28 g ≈ 1 ounce.)
(a) LP = Luna Tech / Pyropak; N = Newco Products; MP = MP Associates; A = Astro Pyrotechnics; TE = Thea-
tre Effects; PT = Precision Theatrical.
(b) Pressure limit criterion was exceeded for the next lower load mass; no test was performed for this load mass.

In this study, as in an earlier study,[1] a large

Discussion of Results degree of variability was observed for internal
mortar pressure pulse shapes. Examples of these
For test firings of the most fuel-rich concus- shapes are shown in Figure 7. It is likely that the
sion powders, it was found that credible data peak shapes are real and reflect differences in the
could not be collected in a blast chamber with on- combustion process. This is because the type of
ly an 8-foot (2.4 m) ceiling. The air blast peak pressure pulse shape observed tends to be predict-
shapes observed were seriously distorted com- able, based on load mass and powder type. Fur-
pared with those collected in a larger chamber and ther, it has not been possible to postulate a simple
for less fuel-rich powders. This observation may model for how a problem with the instrumentation
be consistent with the most fuel-rich powders could produce such widely varying shapes with
producing a portion of their blast wave from a consistent pressure impulses (peak areas). Another
fuel-air explosion above the mortars. reason to believe the various peak shapes are real

Page 410 Selected Pyrotechnic Publications of K.L. and B.J. Kosanke

is the recoil forces, measured independently in the
earlier study, tended to mirror the internal pres- 100
sure peaks.[1c] LP - 7 g

Pressure (psi)
80
The curve in Figure 7A is typical of that ob- (A)
60
served for light loads of the low pressure Pyropak
powder. There is a cluster of peaks spanning 40
about 12 ms in time, with a maximum pressure of
nearly 100 psi[c] (0.7 MPa). When this type of 20
cluster of peaks is observed, there seems to be
0
little consistency in the number of peaks in the 0 5 10 15 20
cluster, their relative amplitudes, or the spacing
between peaks. 3000
The curve in Figure 7B is typical of heavier N - 14 g

Pressure (psi)
loads of the Pyropak powder, all loads of the 2000
Newco powder, all but the heaviest loads of the (B)
MP Associates powder, and the lighter loads of 1000
the Astro powder. There is always a single peak,
but it is not always symmetric. The peak shown in
Figure 7B spans only about 2 ms in time and has a 0
maximum pressure approaching 3000 psi (20
0 1 2 3 4 5
MPa).
The curve in Figure 7C is somewhat typical of
the heavier loads of the MP Associates and Astro
powders, and the lightest loads of the Theatre Ef- 10000 PT - 7 g
Pressure (psi)

fects and Precision Theatrical powders. These

pressure pulses have one or more narrow high- (C)
pressure peaks superimposed on a wider, lower 5000
more modest pressure peak. On some occasions
the pressure pulses have fully developed oscillato-
ry features as seen in the curve in Figure 7C,
0
which only spans about 0.5 ms in time, with a
maximum pressure exceeding 10,000 psi (70 0.0 0.5 1.0 1.5 2.0
MPa).
The curve in Figure 7D is typical for the higher
90000 PT - 9 g
loads (but still only about 10 grams) of the Thea-
Pressure (psi)

tre Effects and Precision Theatrical powders. The

prominent portion of the pressure peak spans less 60000
(D)
than 0.1 ms, and the maximum pressure has risen
to well over 50,000 psi (350 MPa). 30000

It may be interesting to note for the powder

types and load masses tested, even though the 0
peak mortar pressures increased by a factor of 0.0 0.5 1.0 1.5 2.0
approximately 500, the pressure impulse only in- Time (ms)
creased by a factor of 3. Thus, the primary differ-
ence is time span over which the pressure pulse is Figure 7. Examples of a variety peak shapes
produced, which decreases by a factor of approx- seen in internal mortar pressure data (note the
imately 200. substantial differences in the time scales.)
Although the internal mortar pressure curves
generally trend as discussed above, still major
variations occur for apparently identical loads of
the same powder. For example, see Table 5 for

Selected Pyrotechnic Publications of K.L. and B.J. Kosanke Page 411

21-gram loads of the Astro powder. Here the max-
imum pressures were 34, 12, and 47 kpsi (230, 80, 1.5
and 320 MPa, respectively). Despite the large var- LT - 28 g

Pressure (psi)
iation in peak internal pressure, pressure impulse 1.0
and air blast overpressures are much more con-
sistent. The pressure impulses were 4.4, 4.4, and 0.5
4.3 psi·s (30, 30, and 29 kPa·s), respectively, and
the peak air blast overpressures were all 2.1 psi 0.0
(14 kPa). In this case, the total energy being re-
leased and the sound pressure levels are quite con- -0.5
sistent. Independent of concussion powder type, 0 1 2 3 4 5
this was found to be generally true; air blast pres-
sures correlate better with pressure impulse than 1.5
with peak internal mortar pressure. PT - 5 g

Pressure (psi)
1.0
If it can be assumed that the tonal quality
(sharper crack versus mellower boom) is a func- 0.5
tion of the durations of the positive and negative
phases of the blast wave, then there may be no-
0.0
ticeable differences in the tonal quality of sounds
produced by the various powders. (Testing with
-0.5
human subjects is planned to investigate this.)
0 1 2 3 4 5
Table 9 lists the average positive phase durations
for the various powder types, each with load
Time (ms)
masses that produced approximately equal peak Figure 8. Air blast pressure curves illustrating
air blast overpressures (loudness). Figure 8 shows the approximate range of differences in positive
the air blast waves for the two extremes of the phase durations.
various cases. It may be of interest to note that the
order of powder types in Table 9, by decreasing
duration of positive phase, is the same as that in A review of the data in Table 8 reveals that the
Table 8, where they were listed in order of in- durations of positive phase of the air blasts for the
creasing internal pressures for the same load mass. various powders is generally also a function of
load mass. However, the functional relationship is
different for the various powders. Note that for the
Table 9. Average Positive Phase Durations for three lowest pressure-producing powders (Lu-
Powder Loads Producing Air Blast Pressures naTech/Pyropak, Newco and MP Associates) the
of Approximately 1.5 psi (10 kPa). durations of positive phase decreases for increas-
ing load mass. Note further that the opposite trend
Powder and Blast Pos. Phase holds for the two highest-pressure producing
Load (g) (a) (psi) (ms)
powders (Theatre Effects and Precision Theatri-
LP 28 1.3 2.0 cal). Finally, for Astro’s powder, note that the du-
N 21 1.6 1.4 ration of positive phase is essentially independent
MP 14 1.5 1.1 of powder load mass. These relationships are il-
A 14 1.7 0.84 lustrated in Figure 9, which is a graph of positive
TE 9 1.6 0.78 phase duration versus load mass (each normalized
PT 5 1.4 0.64 to the values for the smallest load mass) for the
(For conversion of units: 1 psi = 6.89 kPa) different powder types.
(a) LP = Luna Tech / Pyropak; N = Newco Prod-
ucts; MP = MP Associates; A = Astro Pyro-
technics; TE = Theatre Effects; and PT = Pre-
cision Theatrical.

Page 412 Selected Pyrotechnic Publications of K.L. and B.J. Kosanke

 edged that D. Pier provided the electric matches
1.6
used in this study, and T. DeWille supplied the
1.4 TE / PT concussion mortar.
Relative Positive
Phase Duration 1.2
The authors are also grateful for the technical
and editorial suggestions on this article provided
1.0 A by M. Grubelich, L. Weinman, M. Williams, and
J. Bergman. The authors are also grateful for the
0.8 LP / N / MP assistance of R. Greenberg.
0.6

0.4 Notes
1.0 2.0 3.0 4.0
Relative Load Mass [a] Commercial sound pressure level (SPL)
measuring instruments, operated in “peak”
Figure 9. Graph of positive phase duration ver- mode, typically have a time constant of 50 μs.
sus load mass (each normalized to the values for While this is fast for most sounds, it is still
the smallest load mass) for the different powder fairly slow for an air blast (shock) wave,
types. which has an essentially instantaneous pres-
sure rise, followed by a much slower (but still
fast) decay to ambient pressure. Therefore,
Conclusion such an instrument will underestimate actual
SPL’s of the sounds of explosions. This is the
It was not the intention of this study to rate the same type instrument, with the same time
performance of the concussion powders tested. In constant, that is used to establish acceptable
addition, further studies are needed before truly SPL’s for human exposure. On the other hand,
meaningful conclusions can be drawn. Planned the instrument used in this study to measure
studies include: the effect of distance on peak air blast overpressure, from which sound pres-
sound pressure levels and the duration of positive sure levels were calculated, has a rise time of
phase, and the loudness and tonal quality of the only 4 μs. As a result, the instrument used in
impulse sounds as perceived by human subjects. this study generates higher SPL’s for the
Accordingly, essentially no conclusions are pre- sounds of explosions than typical instruments
sented in this paper. would. This can be important, if the results of
Obviously, the various powders tested have this study are compared with results using in-
significantly different performance characteristics, struments with slower response or are com-
thus offering the user a wider range of perfor- pared with SPL regulations for acceptable
mance choices than might have been expected. It human exposure. For the durations of positive
is hoped that the information in this article proves phase seen in this study, the SPL’s reported
to be useful to consumers in selecting concussion will range from about 1 to 2 dB higher than
powders that: (1) fit their needs, and (2) are com- would have been measured using typical SPL
patible with the burst strength of their concussion instruments.
mortars. Also, hopefully the participating manu- [b] It is fairly common in the authors’ laboratory
facturers and other researchers find these results to work using a mixture of SI and English
of general interest. units. In this paper, for accuracy of reporting
in the text, the actual units of measurement
Acknowledgments are given first, followed by their SI or English
equivalent, with the same number of signifi-
The authors gratefully acknowledge D. Hyman cant figures. In tables, generally only the ac-
of Astro Pyrotechnics, T. DeWille of Luna Tech, tual units of measurement are reported, and
D. Pier of MP Associates, A. Rozzi of Newco conversion factors are appended to the tables.
Products, R. Nickel of Precision Theatrical, and The authors apologize for any inconvenience
N. Kahn of Theatre Effects, for participating in this causes.
this study by supplying samples of their concus-
[c] Note that none of the pressure data presented
sion powders. In addition, it must be acknowl-
in Figure 7 have been corrected for the pres-

Selected Pyrotechnic Publications of K.L. and B.J. Kosanke Page 413

 ence of the silicone sealant. This only affects 3) K. L. and B. J. Kosanke, “Brief Summary of
pressures less than about 2000 psi. For exam- Concussion Powder Test Results”, Fireworks
ple, if corrected, the peak pressure in Fig- Business, No. 153 (1996) p 2.
ure 7A would actually be nearly 200 psi and
4) Encyclopedia of Physics, 2nd ed., VCH Pub-
not 100 psi as shown.
lishers (1991) p 22.
5) Van Nostrand’s Scientific Encyclopedia, 5th
References ed., Van Nostrand Reinhold (1976) p 25.
1) K. L. and B. J. Kosanke and T. DeWille, 6) F. A. Geldard, The Human Senses, 2nd ed.,
“Concussion Mortar Internal Pressure, Recoil, John Wiley & Sons (1972) pp 196–205.
and Overpressure as Functions of Powder
Mass”, Journal of Pyrotechnics, No. 1 (1995) 7) Data collected by K. L. and B. J. Kosanke and
[a] pp 30–31; [b] pp 29–30; [c] p 31. M. Carter at the 1997 Western Winter Blast
(annual convention) of the Western Pyrotech-
2) S. W. Wilson, “Important Industry Warning— nic Association to be reported in a future arti-
Potential Failure of Concussion Mortars”, cle.
Fireworks Business, No. 151 (1996) p 2.

Page 414 Selected Pyrotechnic Publications of K.L. and B.J. Kosanke