Tests of Smokeless Powder in Simulated Rocket Recovery Ejection Systems
July 28, 2001

This is the relatively concise version
Click Here for the longer version with more photos, more detail, and unfortunately more narrative.

Construction of the test device
Series 1:  Constrained vs. unconstained powder charges
Series 2:  Finding the upper limit powder charge
Series 3:  Finding the lower limit powder charge
Series 4:  Tests using black powder
Series 5:  More constrained powder tests
Series 6:  Long-barrel tests
Discussion and conclusions

A topic arose on the Arocket discussion list a month or so back regarding alternatives to black powder as the ejection charge for recovery systems in amateur rockets.  Mr. Thomas Engelhardt indicated that a nitrocellulose paper was being used successfully in Germany.   I am not familiar with this paper and wondered if it is even available here, so I began to think about other possible sources of nitrocellulose, namely ping-pong balls, New Skin, and smokeless gunpowder.

The last of these seemed to hold the most promise, and I happened to have some.  So I ran a few quick-and-dirty tests using one of the larger Estes models, and found that a gram of Red Dot provided a satisfactory ejection of old underwear and a light plastic nose-cone.

Upon reporting this to the Arocket list, Mr. William Westfield directed me to a website by Mr. Scott Aleckson, who describes his experiments with smokeless powder ejection charges.  Mr. Aleckson reported some favorable outcomes, and suggested that powder should be contained in a case to ensure complete burning and thus good performance.  A number of Arocket list members echoed this opinion, and suggested a variety of containment vessels.

But my first tests had obtained a strong, smooth ejection without such containment.  Realizing that I already had everything needed, that assembling test devices could be done quickly and easily, and that I would get to light many fuses with little fear of injury or arrest, this adventure met all of my criteria.  I decided to do some testing.  Secondary benefit: it might yield useful information.

Construction of the testing device

Click here for more detail on construction

Ever the pack-rat, I found a 2-inch PVC tube in my scrap-heap with a male threaded adapter and a female screw-cap.  The total length including cap is 12 inches.  This seems to me a reasonable approximation of the recovery section of a modest-sized amateur rocket, and will hereafter be called the "tube."

A simulated ejection load is assembled by wadding up three double-page sheets of newspaper and ramming them down into the tube with an old hardwood chair-leg, one sheet-wad at a time.

A golf-ball is placed on the wads to simulate the mass of a nose-cone.  The top wad, a single-page sheet of newspaper, holds the ball in place.

Finally, a clean, white paper towel is placed on the powder-side of the load column to record the effects of exposure to the burning powder charge.  This will be called the "bottom wad."

Here is a load that has been packed, then disassembled:

All together the load weighs 113 grams.

Series 1:   Comparison of contained vs. unconstrained charges.

Click Here for more detail on Series 1

The charge:  1 gram of Red Dot smokeless shotgun powder


In tests 1, 3, and 5, the powder was contained only by one layer of two-ply paper toweling.  This will be called "unconstrained" even though it is indeed constrained by the ejection tube, the wadding, and the golf-ball.

In tests 2, 4, and 6, the powder was contained in a brass pistol or rifle cartridge, with a wooden plug securing the open end.  This will be called a "constrained" charge.

All tests were fired with a thin fuse.



Note that the ejection load is placed just beyond the fuse-port, about 2 inches from the end of the tube.  This creates a headspace of 6 to 8 cubic inches, restricting initial powder movement somewhat.

Female cap in place: 

The tube is fastened to a picnic-table with a band-clamp, muzzle about 30 inches off the ground.  It is to be fired horizontally.  The ground is not level, but rises gradually downrange, perhaps one foot in 30 feet.

Upon firing, I make a point of noticing where the golf-ball hits the ground.  It travels rather slowly by ballistic standards, so this isn't very difficult.  I can often find the mark it made in the sand.  I make note of the striking point, then record where the wads landed, the condition of the bottom-wad, and where the golf-ball ended up.

The wads were inspected for damage, particularly the formerly-white paper-towel bottom wad which was exposed most directly to the burning powder.

Table 1:  Tests 1-6
Test # Powder Charge Powder contained in: Tube Condition: Golf-Ball hit at: Ball stopped at __ feet: Wads scattered from __ to __ feet Bottom-wad paper towel condition:
1 1g 
Red Dot
paper towel clean 15 feet 31 feet 12-17 ft. slightly scorched, smoldering at one point
2 1g 
Red Dot
.44 case with plug cleaned with alcohol 11 feet 38 feet 11-18 ft. smoldering at one point, more scorch than (1)
3 1g 
Red Dot
paper towel not clean, sticky 16 feet 38 feet 21-26 ft. hardly scorched, not smoldering
4 1g 
Red dot
.44 case with plug not clean, very sticky 8 feet 24 feet 12-17 feet smoldering at one spot, not otherwise scorched
5 1g 
Red Dot
paper towel not clean, very sticky 12 feet 38 feet 12-23 feet slightly scorched
6 1g 
Red Dot
trimmed 30-06 case with plug not clean, very sticky 9 feet 21 feet 11-20 ft. smoldering at one spot

Discussion:  Ejection was effective in all cases.  It was slightly more vigorous and consistent with the unconstrained powder charges.  Also, the flammable wadding was not damaged as much with unconstrained charges.  I am concerned about my own variations in the constrained charges, and that there is some leakage in the system because of the necessity of having a hole for the fuse to exit.  The results might be different with an electric ignitor, particularly a very energetic one that would simulate a primer by igniting all smokeless powder grains at once.

(See long report for results of Test 7 - slower-burning powder)

Series 2:  Finding the upper limit.

Click Here for more detail on Series 2

Since I had seen little damage to the wadding in the first tests, I wondered how much powder would be too much.

Experiments were designed to test the upper range of the use unconstrained powder for ejection.  These were essentially repeats of tests 1, 3 and 5, but using increasing amounts of Red Dot.  Evidence of excessive pressure, as indicated by damage to the wadding, would indicate the upper range.


Table 2:  Tests 8-12
Test # Powder Charge Powder
Contained in:
Tube Condition Golf-Ball hit at  Ball stopped at: Wads scattered from __ to __ feet Paper Towel condition:
8 2g Red Dot paper towel not cleaned, load very tight 17 feet 62 feet 31-37 feet smoldering
9 3g Red Dot paper towel clean tube, new wads 20 feet hit tree at 37 feet and bounced back to 31 ft. 20-27 ft. discolored
10 3g Red Dot paper towel unclean tube 18 feet bounced off tin at 58 ft 26-35 ft. very slight discoloration
11 5g Red Dot paper Towel unclean, light coat of candle wax 19 feet 51 feet (I think it hit something) 27-39 feet slight scorch, slight discoloration
12 10g Red Dot paper towel uncleaned, light wax 28 feet hit tin backstop at 59 feet 31-48 feet slight discoloration, small hole

Discussion:  Test 12 shows some stress to the wadding (below).  In the other tests, wadding was not damaged to any noticeable degree.

Series 3:  Finding the low limit

Click Here for more detail on Series 3

Experiments are the same as for Series 2, but used decreasing amounts of smokeless powder.  The lower limit would be indicated by failure to eject.

Table 3:  Tests 13-15b
Test # Powder Charge Contained in: Tube Condition Golf-Ball hit at  Ball stopped at: Wads scattered from __ to __ feet Paper Towel condition:
13 .5g 
Red Dot
Paper towel uncleaned, waxed 13 feet 38 feet 17-23 feet discolored, not damaged
14 .3g 
Red Dot
paper towel uncleaned, waxed 10 feet 21 feet 15-20 ft. still in mouth of tube, discolored
15 .2g 
Red Dot
paper towel uncleaned, waxed did not eject did not eject did not eject not observed
15b .3g 
Red dot
paper towel uncleaned, waxed 10 feet 16 feet (may have hit tree) 20-21 feet (one at 3 feet) still in tube, discolored, tiny scorch-marks

Tests 14 and 15b, bottom wad remained in mouth of tube: 

Test 15, charge moved forward but did not eject: 

Discussion:  One-half gram of powder gave pretty good performance, not substantially less than the one-gram charges used in most previous tests.  Two-tenths gram was not enough to eject the load.  Three-tenths was effective in two tests but left some pieces still in the tube, thus I would consider 1/2 gram the minimum.

Series 4:  Black Powder

Click Here for more detail on Series 4

Tests 16 and 17 used small to moderate amounts of black powder for comparison with the smokeless powder tests.

Table 4: Tests 16 and 17
Test # Powder Charge Contained in: Tube condition Ball hit at: ball stopped at: Wads landed at: Bottom wad condition:
16 .5g commercial black powder paper towel uncleaned, waxed 20 feet  53 feet 22-34 feet smoldering heavily
17 2g homemade black powder paper towel uncleaned, waxed 19 feet not found 31-37 feet smoldering heavily

Discussion:  Ejection was vigorous in both cases, but much greater damage was done to the flame-sensitive bottom wad than in any of the smokeless powder tests.

Series 5:  More tests using constrained charges

Click Here for more detail on Series 5

I suspected that the cause of greater wadding damage in the pressurized tests above (2, 4, and 6) was due to the primer flash-hole being pointed straight at a corner of the base wad.  So I plugged the previous fuse hole with a screw, and drilled another one in the cap.  Now the containerized powder charge will sit with its flash-hole facing the rear. A "base wad" is added, so the primer jet points at one paper towel, the wooden plug at another.  Tests 18 and 19 were performed with this configuration, test 20 the same but with an unconstrained powder charge.

Table 5:  Tests 18-20
Test # Powder Charge Contained in: Tube Condition Golf-Ball hit at  Ball stopped at: Wads scattered from __ to __ feet Paper Towel condition: Base-Wad condition:
18 1g 
Red Dot
30-06 casing well cleaned rolled to 9 feet 9 feet 5-9 feet bottom-wad slightly discolored, base wad scorched hole Evenly discolored, scorched hole at fuse
19 1g 
Red Dot
30-06 casing uncleaned, waxed dropped at muzzle 7 inches top wad at -1 inch, lower wads at 18-24 feet bottom-wad smoldering at one point, base wad scorched hole Evenly discolored, scorched hole at fuse
20  1g 
Red Dot
paper towel unclean, waxed 16 feet not yet found 19-25 feet bottom-wad barely discolored, base wad discolored, tiny scorching at fuse-hole. Evenly discolored, tiny scorched hole at fuse

Discussion:  Performance with constrained charges was much lower than in Series 1, a bit of a puzzle to me.  But this did demonstrate that the primer flash-hole emits a jet that burns the wadding.  This might be eliminated with an electric ingitor which allows the hole to be sealed.  The last test using an unconstrained powder charge gave results consistent with tests in Series 1

Series 6:   Some superfluous tests with a longer PVC ejection tube

Click Here for more detail on Series 6

Test 21

I recalled seeing another 2 inch diameter tube in my PVC pile with a male threaded end.  It was 4 feet long, leading me to wonder how the smokeless powder ejection process would work with a longer "barrel."  So I loaded as in test 1 with 1 gram Red Dot contained in paper toweling.

When the charge fired, I barely heard a whoosh, and the load did not eject.  It had moved to within an inch of the end of the PVC tube and stopped there.  I examined the bottom-wad and base wads, which appeared similar to those in tests 18 and 19, no substantial damage was observed except for a burned hole where the fuse went through.  Perhaps the powder gasses rushing through the fuse-hole caused this.

Test 22

So I tried it again, using 2 grams of Red Dot.  The same load was pushed back, new bottom-wad and base-wad installed, and a charge of 2 grams Red Dot inserted.

This one ejected well.  The ball hit the ground at 15 feet.  It continued on to hit the sheet of tin at 58 feet.  Wads were scattered from 19 to 30 feet, and slightly splayed.  Base-wad scorched here and there but not badly damaged.  Bottom-wad discolored but not scorched or torn.

Table 6:  Tests 21 and 22
Test # Powder Charge Contained in: Tube condition Ball hit at: ball stopped at: Wads scattered from: Bottom wad condition:
21 1g Red Dot paper towel clean did not eject  n/a n/a discolored, 1 small scorch
22 2g Red Dot paper towel uncleaned 15 feet 58 feet 19-30 feet discolored, 1 small scorch

This test suggests that a single gram of Red Dot might not be adequate for reliable ejection in larger systems.

Discussion and conclusions:  All tests

1.  Criticism:  My test-system is leaky.  The expediency of a fuse compromises the containment vessel.  An audible whistling was often heard upon firing.  This loss of gas may have changed the burning characteristics of the powder, and certainly reduced the force of the ejection by at least a small amount.

Perhaps leakage is more a factor in the tests where a plugged cartridge constrained the charge, attempting to build some pressure.  Leakage through the flash-hole might have prevented pressures from rising to a useful level.

Defense:  In most tests, forceful ejection was obtained in spite of this leakage.  I plan to repeat some of these tests using electric ignitors, but do not expect the difference to be significant.

2.  Criticism:  In these tests I am burning large amounts of smokeless powder in a manner markedly different from that for which they are intended.  Smokeless powders are designed to be ignited at once from the flash of a primer, and to burn at a carefully determined rate under pressure to accelerate the load to high velocity.  My charges are ignited at one spot and burn at low pressure, the flame spreading from particle to particle.  I suspect that this is inefficient, extracting less energy than the powder is capable of producing.  If memory serves, a gram (15.43 grains) would be a moderate charge for a 12-gauge shotgun shell.  Two grams (of slower powder!) is about the amount used in a 30-06 rifle cartridge (reloaders, please correct me!)

Defense:  High velocity might be desirable for the the rocket as a whole, but a typical ejection system only needs to push a nose cone off and a parachute out.  High velocity can be detrimental to the integrity of these components, as in broken shock cords.  High internal pressures could also damage recovery system components before they exit the tube.  Given the energy of a shotgun or rifle shot, inefficient burning might be a good thing.  I believe that the ideal is to generate the minimum pressure that provides reliable ejection, and to build that pressure rather slowly to minimize shock to the ejection components.

The problem of unburned powder which Mr. Aleckson reports in his tests was not observed in mine.  Very little, if any, unburned powder remained in the ejection tube after any of these tests.  I suspect this is because the head-space allowed for the powder-charge is smaller in my tests:  with the "unconstrained" charges, the powder charge is fired in a cylinder 2 to 3 inches long by 2 inches diameter.  Thus the grains of powder cannot fly far from the flame, and all ignite.  Perhaps the best containment vessel is the ejection tube itself, plugged by the parachute and wadding.  I hope to verify this notion by repeating some tests but using a larger headspace for the powder, and a better way of finding any unburned powder grains.

3.  Criticism:  The golf-ball is not exactly a nose-cone, especially when it's set back 4 inches into the tube.  This may give the ball more momentum as it is exposed to the forces of acceleration for a longer period.  And newspaper wads aren't much like a parachute.  There is no shock cord, and I am not using fire-resistant wadding!

Defense:  none.  I need to do more testing, using real recovery-system components.

4. Criticism:  These are static tests: the ejection system is neither in motion nor at altitude upon firing.
I wonder how smokeless powder performance might change at high altitudes, expecting that it would generate higher relative pressure because of the lower ambient pressure.  This might be countered somewhat by air resistance if the rocket is not at apogee, and is either rising or falling.  Since the consequences of a failed ejection can be dire, I might use a little more powder than the minimum to ensure reliable ejection.  Fortunately, the range seems wide - I had successful ejection using charges as light as 0.3 gram and as heavy as 5 grams.

Defense:  None.  I hope to find a way to test this system as it falls through the air, with a real nose-cone and deploying a parachute  But I can't fire much of a rocket around here and need to be able to do it over and over.  Thought about using a big slingshot to get the short tube up a little ways.  :)  If only I had a hot-air balloon, or a microlight aircraft...

Conclusion:  I believe that the present results support the notion that smokeless powder can be an effective ejection charge, that limited headspace provides adequate constraint to ensure full ignition, and that smokeless powder offers some advantages over back powder.  But at least one more round of testing is needed before trusting it in a launch of any consequence.

Please feel free to share your comments, questions, and suggestions regarding this project.  I am sure many will have criticisms and ideas that have not occurred to me, and I look forward to hearing them all.

James Yawn