Timing effects on fragmentation

P.D. Katsabanis, A. Tawadrous, C. Braun and C. Kennedy

Department of Mining Engineering
Queen’s University

Abstract
A series of small scale tests, simulating multi-hole blasts have been performed to establish the
effect of delays on blast fragmentation. The blasts were performed in high quality granodiorite
blocks, which were cut from stone prepared by dimensional stone quarry operations. The pattern
used was equilateral triangular, with a distance of 10.2cm between boreholes, which had a
diameter of 11mm, were loaded with detonating cord and the coupling medium was water. The
delays used were achieved using different lengths of detonating cord for the cases of delays
between 0 and 100 microseconds between holes and a sequential blasting machine firing seismic
detonators for larger delays up to 4ms. All fragments were collected and screened. The
experiments showed that the coarsest fragmentation was achieved with simultaneous initiation of
all charges. Fragmentation became finer with delay time between holes up to 1ms between holes.
The results show that in moderate to competent rock, similar to the granodiorite of the
experiments, fragmentation optimization requires delays of few milliseconds per meter of
burden. In view of the findings, which agree with previous published work, the current trend of
selecting fast firing times for fragmentation optimization is questioned.

Introduction
Timing of detonation has long been of interest to blasting. It has long been associated with muck-
pile profile and heave, and, since it became obvious that timing could be controlled accurately
using electronic detonators, it has been associated with fragmentation. It is in this part that the
most controversy exists. The reasons are the lack of clear understanding of how fragmentation is
created and the ambiguity in the relative importance of stress wave propagation and gas penetra-
tion of the rock. It is quite common today to use few ms between holes of the same row in large
scale open pit blasts and there is a trend to use “faster” timing between boreholes. The reported
results are often “encouraging” or “good”. However claims must be substantiated, especially
when, research conducted a few years ago, has suggested that very short times may not be
beneficial.
Stagg and Rholl (1987) using reduced scale experiments suggested that fragmentation improves
when delays exceed 3.3ms/m of burden. In the same study, with larger scale experiments, the
authors suggested that short (1ms/m of burden) and long (larger than 26ms/m of burden) delays
had adverse effects on fragmentation.Otterness, Stagg and Rholl (1991) have presented full frag-
mentation data for 29 reduced scale shots. They suggested that delays of 3.3-13ms/m (1-4ms/ft)

Copyright © 2006 International Society of Explosives Engineers

2006G Volume 2 - Timing Effects on Fragmentation 1 of 11
of burden had an improvement of 12-20% on fragmentation. Chung and Katsabanis (2001)
analysed the same data and found that there is a power relationship between average fragment
size and time. The relationship is given in Figure 1.

22

20

18

16

14

12

10

8
A

6
f

4
0 20 40 60 80 100 120 140

Time, ms
t i

Predicted
Measured

Figure 1. Average fragment time as a function of time (Chung and Katsabanis, 2001)

Rossmanith (2003), using Lagrangian diagrams has suggested that the interaction of the waves
and features, such as cracks or free faces, is of importance in rock blasting. This implies the use
of “short” delay times, since wave propagation is rather fast in typical geologic materials.
Rossmanith has suggested that mines throughout the world have taken advantage of this concept,
using electronic detonators.
Rosenstock (2004), has advocated the use of very small delays and has backed his claim with
industrial observations.
McKinstry (2004) has discussed timing selection at Barrick for their electronic detonator produc-
tion shots. He has suggested a time of 3ms between holes of the same row to create collisions
between the stress waves and fragmentation optimization.
There are several studies to support both points of view. Obviously one is concerned with the
discrepancy. Where was the mistake, if there was one, in the earlier work? Where are the results
of comparative studies to verify the late trends?

Copyright © 2006 International Society of Explosives Engineers

2006G Volume 2 - Timing Effects on Fragmentation 2 of 11
The purpose of the present paper is to shed some light into the debate. Experimental data will be
presented in this paper and their implications will be discussed, while further analysis will be
attempted later.

Experimental work
Small scale experiments were performed in blocks of granodiorite. The blocks were cut from
stone prepared by dimensional stone quarry operations for commercial use. The blocks had
length of 92cm, width of 36cm and height 21cm. The top and bottom surfaces were smooth,
while the surrounding vertical faces were rough. Boreholes were drilled in each block, creating
an equilateral triangular pattern with a distance of 10.2cm between borehole centres. A total of
23 boreholes were drilled per block. The pattern is shown in Figure 2. The holes were drilled in
identical geometry using a template. The boreholes were 11mm in diameter and had a length of
18cm. Each borehole was loaded with three strands of 5.3g/m (25 grain/ft) detonating cord. Thus
the powder factor used was 0.95kg of PETN per m3 of rock. If one uses a relative to ANFO
weight strength of 1.56, this can be translated in a powder factor of 1.5kg/m3 in terms of
commercial explosives. The same delay was used between holes of the same row and a delay
twice as long was used between the different rows of the blast. The delays used were achieved
using different lengths of detonating cord. Delays larger than 100μs were difficult to implement
using different lengths of detonating cord. The piece of detonating cord between successive holes
was 0.7m long, making connections prone to cutoff. A lot of care was taken to avoid
malfunctions, however it was impossible to use any longer lengths of detonating cord, given the
small size of the block used. Therefore, for delays equal to or larger than 1ms, a sequential
blasting machine (REO model BM-175-10ST-M) was used, which fired seismic detonators.
Blasting took place inside an 1m3 steel chamber, which had its interior covered by rubber mats,
to minimize secondary fragmentation due to impact of the primary fragments on the wall of the
chamber. All fragments were collected and screened.

Figure 2. Pattern used to blast small scale granodiorite blocks.

Copyright © 2006 International Society of Explosives Engineers

2006G Volume 2 - Timing Effects on Fragmentation 3 of 11
Results
The results of the tests are shown in Table 1.
Passing % for delay time (mus)
Size 0 10 10 20 40 80 80 100 1000 2000 4000
mm Repeat A B
254 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00
180 87.09 100.00 94.06 94.80 100.00 100.00 78.23 95.14 100.00 84.89 81.38
127 59.12 89.28 90.97 79.66 84.18 88.48 78.23 79.81 92.96 65.15 68.55
90 34.91 68.74 57.77 59.16 65.07 73.39 60.70 60.25 69.31 45.03 52.50
63 21.60 45.29 38.67 42.66 44.82 51.71 38.41 41.53 52.20 32.83 34.79
53 20.82 39.89 34.72 38.87 40.79 47.03 36.19 38.26 52.17 28.22 29.87
45 19.52 36.46 31.91 35.19 36.25 40.09 30.48 33.87 46.25 25.71 24.31
37.5 18.64 32.30 28.58 31.73 33.10 33.65 26.89 31.01 40.67 21.38 22.85
31.5 17.85 29.70 26.18 29.83 30.00 29.61 24.14 27.53 34.97 19.62 20.15
26.5 17.25 27.37 23.84 27.63 27.72 26.60 21.45 24.67 29.78 16.40 18.46
22.4 15.37 24.82 21.40 25.00 24.88 23.20 18.55 21.55 26.48 14.30 16.76
19 14.55 23.05 19.94 23.37 22.87 21.16 17.08 19.98 23.94 12.96 15.11
16 13.35 20.45 18.12 21.06 19.67 18.20 15.41 17.70 20.67 10.89 12.95
13.2 12.55 18.51 16.66 19.36 17.81 16.55 14.10 15.94 19.00 9.72 11.83
11.2 11.39 16.70 15.15 17.28 15.74 14.11 12.56 14.39 17.22 8.71 10.48
8 9.76 13.95 13.07 14.49 12.85 11.60 10.56 11.98 14.37 7.32 8.64
6.7 9.10 12.30 11.25 12.87 11.45 10.22 9.43 10.89 13.11 6.53 7.78
4.75 7.67 10.66 9.80 10.93 9.50 9.09 8.09 9.27 11.22 5.52 6.72
3.36 6.24 8.82 8.34 8.98 7.63 7.51 6.80 7.80 9.52 4.71 5.76
2.38 5.32 7.74 7.26 7.84 6.57 6.55 6.10 6.91 8.09 4.11 5.02
1.7 3.83 5.85 5.53 5.87 4.98 4.91 4.61 5.33 6.44 3.14 3.94
1.18 3.04 4.73 4.49 4.76 3.87 3.96 3.79 4.37 5.29 2.58 3.18
0.85 2.61 3.83 3.66 3.79 3.42 3.56 3.05 3.60 4.33 2.11 2.62

TABLE 1. Fragmentation analysis of the experiments

The B column for the 80μs delay corresponds to the second part of the block, which was blasted
after the main event, due to a cut-off. Due to the irregular shape of the block, the result of this
test is not representative of the testing parameters for the block.
The fragmentation results were used to statistically fit the Gaudin-Schumann and Rosin-
Rammler equations, which are commonly used to describe fragmentation. The parameters of the
fits are shown in Table 2. Figure 2 shows the comparison between the fitted equations and the
experimental data for the case of Rosin-Rammler distribution, while Figure 3 shows the same
curves for the Gaudin Schumann distribution. Apparently Gaudin-Schumann fits the bottom part
of the curve (fines) at the expense of the top part. Rosin Rammler on the other hand tries to fit
the entire range, penalizing both parts. To identify changes in the small (10%) and large (80%)
passing sizes, both fits were deemed unsatisfactory. It was decided to generate separate, different
fits for the finer and coarser parts of the distributions. A careful look at the distributions
suggested a change in slope around the 50mm particle size. Whether this indicates a change of
the mechanism of fragmentation for particles above and below this size needs to be investigated.
Fit of the Gaudin-Schumann equation was used for the finer part, while fit of the Rosin-Rammler
equation was used for the coarser part.

Copyright © 2006 International Society of Explosives Engineers

2006G Volume 2 - Timing Effects on Fragmentation 4 of 11
Copyright © 2006 International Society of Explosives Engineers
2006G Volume 2 - Timing Effects on Fragmentation 5 of 11
Copyright © 2006 International Society of Explosives Engineers
2006G Volume 2 - Timing Effects on Fragmentation 6 of 11
 TABLE 2. Gaudin-Schumann and Rosin-Rammler parameters resulting from the fit of the experimental
results.

0l 10 10 20 40 80A 80B 100 1000 2000 4000

Gaudin-Schumann
xm, mm 1513 279 390.4 313 234 165 294 274 146 319 420
m 0.45 0.55 0.53 0.53 0.6 0.71 0.63 0.60 0.67 0.71 0.61
2
R 0.99 1 1 1 1 0.99 0.99 1 0.99 0.99 0.99
Rosin-Rammler
xc,mm 131 83 91 94 87 77 102 94 77 120 114
n 2.13 1.89 2.11 1.57 1.71 1.62 1.26 1.61 1.74 1.60 1.41
2
R 1 1 1 0.99 1 0.93 1 0.99 0.99 0.99

180

160

140

120

100

80

60
P ti l Si

40

20

0 1000 2000 3000 4000 5000

Delay Time, μs
-10%
-50%
-80%

Figure 5. Effect of delay on 10%, 50% and 80% passing sizes

On the basis of the above, the various passing fractions were evaluated as a function of delay
time, to examine if there is any effect of timing on fragmentation in the experimental results.
Figure 5 shows the results for the 10%, 50% and 80% passing sizes. The results show similar
trends in all passing sizes. However the changes in the 10% passing fractions appear to be very
small. In fact this is expected, as fines would originate from close to the hole and would be

Copyright © 2006 International Society of Explosives Engineers

2006G Volume 2 - Timing Effects on Fragmentation 7 of 11
affected only by the detonation in each hole. In the 80% passing fraction, significant changes
were observed; however the results indicate a large amount of scatter, as plotted. This will be
analysed in the following.

Discussion
There is no doubt that the largest fragments were produced by the instantaneous initiation. Frac-
tures joining the boreholes were produced, resulting in large fragments. These disappeared as
delay increased. At very long delays, fragmentation became coarser again. Whether there is an
optimum delay, it is difficult to assess, with the number of data points available. Delays as short
as 10μs appear to have produced little difference from delays as long as 1ms.
The results of the 2ms and 4ms delays are close to each other. These delays are very long for
stress wave interaction. At the same time cracks from the detonation of each charge are fully
extended and open at delays as short as 140μs between holes. The P wave velocity of the
granodiorite, as measured in previous work was 3900m/s, so it is reasonable to assume a crack
velocity of 1/6 of this (650m/s or 0.65mm/μs), thus radial cracks would reach the free face of the
blast in about 140μs after initiation. If the fractures are open, any delay above 140μs would be
equivalent to blasting each hole independently, not changing the fragmentation produced. If the
fractures are closed, stress waves can still increase damage.
The first point worthy of examination is related to the small fragment size associated with the
1ms delay. It is worth noting that Winzer, Anderson and Ritter (1983) observed a continuous
reduction of the 80% passing size at delays between 1.3 and 4.9 ms/m of burden in block tests
and a somewhat unchanged fragmentation size between 10 and 30ms/m of burden in small scale
bench tests. The delay of 1ms in the present work, given that the burden is 8.9cm, corresponds
to a delay of 11ms/m of burden, which is in the range of the previous work of Winzer, Anderson
and Ritter, but beyond optimum. Since this delay is produced using a sequential blasting machine
operating at shorter than usual delay times, it was initially viewed as inaccurate. Examination of
the delays produced by the sequential blasting machine and the seismic detonators used revealed
very little scatter. The time difference between actual and desired timing of the detonators had a
mean of -0.223ms and a standard deviation of 0.41ms. Thus, it appears that damage accumulates
in the rock after the stress waves have passed. This is compatible with the findings of Stagg and
Rholl (1987) who observed that optimum fragmentation in their experiments occurred when the
next hole in a row fires so that it interacts with the late failure processes around an earlier
borehole. These late failure processes have been attributed to the action of gases. Stagg and Rholl
measured gas velocities as low as 6% of the P-wave velocity. In the present experiments such
velocity would be equal to 230m/s. This suggests that pressurization of the radial cracks before
venting could occur for almost 400μs. Considering that the boreholes are unstemmed, gas
velocities could even be lower. This explains the long delay window in which fragmentation is
close to optimum.
The longer delay experiments were performed to examine fragmentation with independent hole
detonation. It is comforting to observe that almost identical results with the 2ms and 4m delays
were obtained, suggesting that there can be no effect of timing on fragmentation at such long
delay times. Scaling the event up, suggests no influence of delay time above delays 22ms/m of
burden. The cutoff is probably smaller, however experimental proof is not available yet.

Copyright © 2006 International Society of Explosives Engineers

2006G Volume 2 - Timing Effects on Fragmentation 8 of 11
At the small delays, there is influence of delay on fragmentation. Figure 6 plots the results of the
small delays. The trend is very similar to the one indicated by the power equation suggested by
Chung and Katsabanis (1999). In fact the exponent of the power law is -2.37 which is close to the
-2.31 found from the fit of the USBM data if the delay is expressed as ms/m of burden. Under-
standably, finding the limits of the relationship is of importance.

180

160

140

120

100
Si

80

60

40
0 20 40 60 80 100 120

Delay time, μs
50% passing
80% passing

Figure 6. Effect of delay on 50% and 80% passing sizes

It appears that there is influence between delay and fragmentation when short delays are used.
There is little doubt that instantaneous initiation does not benefit fragmentation. From 0.11ms/m
of burden to about 1.1ms/m of burden there is little change of both average fragment size and
80% passing size. The experimental results suggest that fragmentation does not change much
even at a delay of 1ms of the experiments, corresponding to 11ms/m of burden. While one can
easily accept no change or improvement up to a delay of 1.6ms/m, which corresponds to the time
the cracks find the closest free face of the blast, the lack of significant change, or even the
improvement of fragmentation at 11ms/m of burden is rather unexpected. It suggests that damage
accumulates due to the action of gases and the combination of this with the energy provided by
later detonating charges is beneficial.
According to the previous analysis fragment size would increase at some point, when no energy
from a later detonating charge can be transmitted to the fragments produced by and earlier deto-
nating hole. Indeed, at delays above 22ms/m, fragmentation is coarser and does not change with
delay time.
Of interest are the observations when the strength of the fragments is examined. A very easy test
to examine this, is the point load test. Figure 7 shows the results of the measurements as a

Copyright © 2006 International Society of Explosives Engineers

2006G Volume 2 - Timing Effects on Fragmentation 9 of 11
function of the delay time. It appears that the index decreases rapidly with delay time. The lowest
index was recorded at a delay of 100μs. After this the point load index appears not to be affected
by the delay time. This is compatible with the previous analysis for fragmentation suggesting that
damage continues accumulating until open cracks isolate each fragment from the influence of
incoming waves. At very long delays, crack development and opening of cracks before the
detonation of the next charge is expected.

10

0
0 1000 2000 3000 4000 5000

Delay Time, μs

Figure 7. Point load index Vs. delay time

The above suggest that fragmentation is affected by the delay time. Very long, or very short
delay times result in coarse fragmentation, while there is a time window in which fragmentation
is fine. For granodiorite this time window is between 0.11ms/m of burden to 11ms/m of burden.
The experimental evidence and the preceding analysis does not support that there is a single
delay that produces good fragmentation. It appears that there are several choices to select from.
Given that a good blast does not only produce good fragmentation, but acceptable vibration
levels, it is clear that opportunities exist to have excellent blasting results.
It is clear that the present findings are in agreement with the work of Stagg and Rholl (1987). It is
our opinion that stress wave superposition is not sufficient to explain the findings of Stagg and
Rholl and the findings of the present investigation. While there is no doubt that better fragmenta-
tion has been obtained when electronic detonators are used, it appears that the benefits are associ-
ated more with the precision of the detonators rather than the ability to detonate holes at very
short intervals.

Copyright © 2006 International Society of Explosives Engineers

2006G Volume 2 - Timing Effects on Fragmentation 10 of 11
Conclusion
In a series of small scale experiments in granodiorite it was shown that fragment size decreases
with delay time, from a maximum size, during simultaneous initiation of all charges, to an
approximately constant size, for delays up to 11ms/m of burden. When large delays are used,
fragmentation suffers, since charges detonate at such delay times that fragments are separated by
open cracks. In view of the findings, which agree with previous published work, the current trend
of selecting fast firing times for fragmentation optimization is questioned.

References
Chung, S. and Katsabanis, P.D. (2001): “An Integrated Approach for Estimation of Fragmenta-
tion”. Proc. of the 27th ISEE Annual Conference on Explosives and Blasting Technique.
Orlando, Florida, Jan. 28-31, Vol. 1, pp. 247-256.
McKinstry, R., Bolles, T. and Rantapaa (2004): “Implementation of Electronic Detonators at
Barrick Goldstike Mines, Inc.”. Proc. of the 30th ISEE Annual Conference on Explosives and
Blasting Technique. New Orleans, LA, Feb. 1-4, Vol. 1, pp. 349-361.
Otterness, R.E., Stagg, M.S. and Rholl, S.A.(1991): “Correlation of Shock Design Parameters to
Fragmentation”. Proc. of the 7th ISEE Conference on Explosives and Blasting Research, Las
Vegas, Nevada, Feb. 3-7, pp. 179-181.
Rosenstock, W. (2004): “Advanced Electronic Blasting Technology (AEBT). Breaking
3,205,000 tons of Ore within a Millisecond”. Proc. of the 30th ISEE Annual Conference on
Explosives and Blasting Technique. New Orleans, Louisiana, Feb. 1-4, pp. 61-68.
Rossmanith, H.P. (2003): “The Mechanics and Physics or Rock Blasting”. Proc. of the 29th ISEE
Annual Conference on Explosives and Blasting Technique. Nashville, Tennessee. Feb. 2-5, Vol.
1, pp. 83-101.
Stagg, M.S. and Rholl, S.A. (1987): “Effects of accurate delays on fragmentation for single-row
blasting in a 6.7m (22ft) bench”. Proc. Second International Symposium on Rock Fragmentation
by Blasting, Keystone, Colorado, Aug. 23-26, pp. 210-230.
Winzer, S.R., Anderson, D.A. and Ritter, A.P. (1983): “Rock Fragmentation by Explosives”.
Proc. First International Symposium on Rock Fragmentation by Blasting, Lulea, Sweden, Aug.
23-26, pp. 225-249.

Copyright © 2006 International Society of Explosives Engineers

2006G Volume 2 - Timing Effects on Fragmentation 11 of 11