The optimization of piston tube length utilizing a floating head piston launcher, as recommended by these same authors in their NARAM-28 research and development report, was studied. The performance of the launcher was evaluated using an improved speed determination apparatus capable of monitoring launch velocity as a function of engine burn time from the moment of ignition. The launch mass of the test models represented a typical 1/2A streamer competition model. The models were flown with no piston (controls), a 9 inch piston tube and an 18 inch piston tube. Graphs of velocity versus time showed that at a 7 inch piston tube is required to achieve the optimum benefit of the floating head piston launcher with a nine inch piston tube being desirable to allow a margin of safety. No significant additional benefit was observed for piston tubes of longer length. Control models flown with no piston showed that thrust produced during the first 0.1 seconds of engine burn was insufficient to significantly accelerate a model and that the gases produced are essentially wasted. These same gases were effectively used by the floating head piston launcher to accelerate a model to velocities of approximately 26 feet per second in 0.04 seconds.
The apparatus developed to evaluate launcher performance
proved highly effective and it is recommended that it be
utilized for future research in optimizing other
floating-head piston launcher design parameters and
studying their relationship to other competition
Email: cbweiss at frontiernet dot net
Slingerlands, NY 12159
Email: jeffvincent at verizon dot net
At NARAM-28, the same members of this team presented research demonstrating a significant increase (34%) in performance of a "floating-head" piston launcher design over the standard zero-volume piston launcher. A novel technique utilizing the interruption of an infra-red beam by the fin of a model rocket at a fixed distance above its launch position was demonstrated to evaluate launch performance. Since the objective of the NARAM-28 study was to compare the performance of the "floating-head" piston and standard zero-volume piston launchers, design parameters such as piston tube length, and diameter were not necessarily optimized. The optimization of these parameters were suggested for future research.
The objective of the research presented here in this report was to empirically evaluate the parameter of piston tube length in relation to performance for the floating-head piston design, and to determine if an optimum length exists. In order to achieve this objective, a more sophisticated evaluation method based on principles similar to the NARAM-28 study was developed. The method allowed the visualization of model velocity as a function of engine burn time during the launch process, thus making it possible to compare performance in terms of impulse experienced by the model.
The principle of operation of the floating-head piston launcher was fully described by Weiss and Vincent in the Research and Development report "The Floating Head Piston Launcher", Odd Couple Team, T-085, NARAM-28 and therefore will not be expounded upon in this report. All research conducted in this study was performed using the floating-head piston design.
The basic concept behind the advantage produced by a piston launcher is based upon the principle that up to a critical point in engine burn time, the mechanism for producing acceleration in a piston launcher is more effective than the action-reaction mechanism of a rocket engine alone. The acceleration produced by the launcher is related to the rate at which gas is produced by the rocket engine, the weight contributed by the model and piston tube, frictional forces associated with the launch, and the potential leakage and rapid cooling of the engine exhaust gases in the piston tube. Since the length of the piston tube can effect, or is related to, several of these factors, it is reasonable to assume that an optimum piston tube length exists, beyond which no further acceleration advantage is experienced.
Thoelen, Bauer, and Porzio demonstrated that no increase in altitude performance was gained for piston tube lengths greater than twelve to fifteen inches using a standard zero-volume piston launcher. In order to better understand the force-generating mechanism occurring in the floating-head piston launcher, we sought a direct determination technique that would allow the evaluation of launcher performance during the critical period of engine burn. A valid direct determination method would also be valuable in optimizing other design parameters.
The speed measurement device is based upon a set of sixteen infra-red (IR) emitter/detector pairs which are mounted in the path of the fin. If the fin is of known length, the time it takes the model to traverse this distance can be used to calculate the speed of the model at that point. A hybrid Atari BASIC/6502 machine language program is used to sense the ignition of the engine, starting the timing loop. As the model's fin passes through a given sensor, the program records the time of the interruption and restoration of the IR beam. The difference of these times is used to calculate the average velocity over this distance (the fin chord). The timing device has shown a computed accuracy of +/- 0.80% at the model velocities encountered in this test. It is strongly recommended that anyone interested in further details of this device consult the Appendix.
A special apparatus was required to hold the piston launcher, the ignition sensor, and the sixteen speed sensors in position. The guide for the piston tube was similar to a three-pole tower launcher. The piston support rod was centered between the poles by a center-drilled wooden cylinder that was held in place by a stop at the base of the tower poles. A masking tape stop on the support rod provided the appropriate positioning of the piston launcher relative to the ignition and speed sensors. The apparatus also had two adjustable fin guide rails constructed of 2 x 3/4" redwood. The rails were mounted parallel and adjacent to the piston guide poles. The rails were spaced far enough apart so that the fin of a model could just pass freely through them with little friction. The sixteen IR sensors were mounted in the rails 2.5" (+/- 0.05") apart. The inside of the rails were blackened to reduce stray IR reflections. The ignition sensor was held adjacent to the piston tube in a special holding device. The holder consisted of a 19/32" o.d. brass tube 1.5" long, with a perpendicular side tube. The side tube was constructed of 1/4" o.d. brass tubing and was located 5/8" (on center) below the top of the larger tube. The ignition sensor was fitted in the 1/4" brass side tube approximately 1/16" from the inside wall of the larger brass tube. The holder was held tightly in place between the three tower poles and could be easily adjusted to the desired test location.
The piston tubes were constructed of BT-5 tubing fitted with a translucent fiberglass collar that extended 3/4" above the top of the BT-5 tubing. The piston tube and collar could slide freely through the 19/32" brass tubing. In the test configuration the engine was fitted into the model so that 1/2" of engine protruded from the bottom of the model into the translucent tubing. The top of the piston head was even with the top of the BT-5 tubing. This created a 1/4" window adjacent to the ignition sensor and completely shielded the detector from ambient IR light.
When no piston was used (control flights), the ignition detector was placed approximately 1/4" directly beneath the base of the engine. A special holder was constructed of 3/4" RB-52 tubing with a 1/8" layer of epoxy in the top end of it. A short piece of Estes maxi-lug was mounted under the epoxy. The maxi-lug provided a location to mount the ignition detector, and two small holes were drilled in the epoxy for the ignitor. This holder was held between the tower poles in a similar fashion to the brass detector holder. Photographs 1 through 6 show several views of the speed measurement device, including the control ignition detector and holder.
The floating-head piston launcher was constructed in a similar fashion to the design described in the NARAM-28 report. The piston head was constructed of 17/32" o.d. brass tubing, fitted over a wooden core. The brass sleeve was turned down to 0.516" diameter on a Unimat modeler's lathe to fit inside the BT-5 piston tube. The piston head was 0.735" long and weighed 2.2 grams. A 9/32" diameter x 5/8" hole was drilled in the bottom of the piston head to center it on the piston rod with minimal friction. Two 1/64" diameter holes spaced 1/8" apart were drilled in the top of the piston head so that they aligned with the terminals in the end of the piston rod. Solar ignitor wires were threaded through the holes and into the terminals. The piston rod was constructed of 1/4" o.d. x 24" aluminum tubing. Each piston tube contained an Estes EB-5B engine block glued inside the bottom of the tube to act as a piston head stop.
Three models were used to conduct the test program. The models were constructed identically from Estes BT-5 tubing and BNC-5 nose cones. The fins were constructed from 1/32" plywood for minimal thickness and maximum durability. The fins were rectangular with a span of 2.25" and a chord of 1.25" (+/- 1.0%). One fin on each model was identified as the fin used for the speed measurement device.
Tests were conducted using nine and eighteen inch piston tubes. The weight of the 9" tube (including the fiberglass collar and engine block) was 3.5 grams, the weight of the 18" tube was 6.5 grams.
The experimental design employed was intended to minimize error due to differences between test models or preparation procedure. For each test, the models were rotated in a regular fashion. Data was collected on a series of twelve test flights. Some additional flights were run initially to test and debug the hardware and software of the device.
The test program was conducted by the following procedure. The engine was selected (all engines were Estes 1/2A3-2T, batch 7Rl2) and fitted in the model with 0.5" of casing protruding from the base of the model. The engine was shimmed with masking tape to obtain the desired friction fit. All engines were fitted by the same operator to achieve uniformity. The model was then weighed and the weight was adjusted to 16.5 grams (+/- 0.1 grams). This weight was chosen to exemplify a typical 1/2A streamer duration competition model. The pistons were equipped with Solar Ignitors of appropriate length to insure ignition while the 1/4" ignition detector window was maintained. The model was fitted to the appropriate piston tube and the existence of the ignition detector window was visually verified. The launcher and attached model was then loaded into the speed measurement device so that the designated fin rode in the fin guide rail and the piston was at the predetermined rest point. The rest point was chosen so that the trailing edge of the designated fin was exactly 2.5" below the first speed sensor, measured from the center of the sensor. This rest point was secured by the masking tape stop on the piston rod and the centering dowel at the base of the tower poles. The ignition detector holder was locked in the correct position for the proper alignment with the translucent window. The launcher was then checked to insure that the piston tube was properly aligned and that no binding occurred.
The continuity of the launch system was then checked. The computer program was then run. It first ran a test to check the seventeen detectors for ambient IR light. It then tested the continuity of the IR beam in the sixteen sensors. If these tests are successful the program then enters the machine language timing program and waits for ignition. The model was then launched and recovered. After the flight, the program saved the time and computed velocity data to disk and printed out the results.
Sensor # Mean Time R.S.D. Mean Velocity R.S.D. 1 0.125 sec 4.37 % 11.2 ft/sec 9.37 % 2 0.137 sec 4.29 % 21.1 ft/sec 5.77 % 3 0.145 sec 4.19 % 30.4 ft/sec 4.68 % 4 0.151 sec 4.11 % 37.5 ft/sec 4.64 % 5 0.155 sec 3.89 % 44.3 ft/sec 5.37 % 6 0.158 sec 4.08 % 50.7 ft/sec 5.81 % 7 0.161 sec 4.60 % 55.8 ft/sea 3.70 % 8 0.168 sec 3.93 % 63.1 ft/sec 4.46 % 9 0.172 sec 3.85 % 68.3 ft/sec 3.64 % 10 0.174 sec 3.87 % 74.8 ft/sec 5.10 % 11 0.177 sec 3.84 % 81.4 ft/sec 2.63 % 12 0.180 sec 3.79 % 79.9 ft/sec 6.67 % 13 0.182 sec 3.76 % 93.4 ft/sec 4.37 % 14 0.185 sec 3.74 % 102.1 ft/sec 4.59 % Note: R.S.D. is relative standard deviation.Go to this data in graphic form
Sensor # Mean Time R.S.D. Mean Velocity R.S.D. 1 0.0219 sec 17.2 % 17.1 ft/sec 12.3 % 2 0.0316 sec 14.7 % 23.3 ft/sec 11.1 % 3 0.0402 sec 13.4 % 26.5 ft/sec 12.6 % 4 0.0490 sec 12.8 % 25.0 ft/sec 17.6 % 5 0.0572 sec 12.2 % 26.2 ft/sec 9.54 % 6 0.0653 sec 11.5 % 26.2 ft/sec 8.94 % 7 0.0732 sec 11.0 % 26.9 ft/sec 10.9 % 8 0.0811 sec 10.6 % 27.5 ft/sec 8.26 % 9 0.0888 sec 10.2 % 27.0 ft/sec 6.60 % 10 0.0960 sec 9.76 % 29.5 ft/sec 5.37 % 11 0.102 sec 9.18 % 31.5 ft/sec 7.04 % 12 0.109 sec 8.94 % 32.4 ft/sec 3.19 % 13 0.115 sec 8.54 % 37.1 ft/sec 2.34 % 14 0.121 sec 8.18 % 41.3 ft/sec 2.05 %Go to this data in graphic form
Sensor # Mean Time R.S.D. Mean Velocity R.S.D. 1 0.0230 sec 16.0 % 16.8 ft/sec 16.1 % 2 0.0330 sec 11.5 % 22.2 ft/sec 13.1 % 3 0.0418 sec 9.23 % 25.5 ft/sec 8.89 % 4 0.0503 sec 7.97 % 25.8 ft/sec 4.80 % 5 0.0583 sec 6.85 % 26.8 ft/sec 2.78 % 6 0.0661 sec 6.01 % 26.8 ft/sec 1.45 % 7 0.0735 sec 5.02 % 28.9 ft/sec 4.93 % 8 0.0805 sec 4.36 % 31.0 ft/sec 10.5 % 9 0.0872 sec 3.86 % 31.2 ft/sec 7.56 % 10 0.0936 sec 3.60 % 32.5 ft/sec 8.21 % 11 0.0997 sec 3.35 % 35.5 ft/sec 7.95 % 12 0.106 sec 3.30 % 34.9 ft/sec 8.02 % 13 0.111 sec 3.19 % 39.7 ft/sec 7.70 % 14 0.117 sec 3.37 % 37.6 ft/sec 19.7 % The mean velocity and mean time in the table above was computed for three test runs only. Visual observations during one test indicated that the model separated prematurely from the piston tube. The deviation in the data points for this test can be visualized in the graphs for the individual velocity/time plots shown in Figure 7 (p. 15). We believe the test results for this run are an outlier and they have been excluded in the computation of averages.Go to this data in graphic form
The Data tables list the means of the average velocities and times measured at each of the fourteen speed sensors for the control, 9", and 18" piston tubes, respectively. Data for sensors fifteen and sixteen is not listed due to their malfunction on some of the test runs. The means for the control and the 9" piston are computed from four test runs. For reasons previously noted, the mean for the 18" piston tube are computed from three test runs only. The relative standard deviation (R.S.D.) is given for each mean calculated.
The graphs in Figures 1, 4, and 7 show the average velocities as a function of time measured at each speed sensor for each of the individual test runs (data not listed). Figures 2, 5, and 8 show the mean of the average velocities as a function of the mean of the average times at each speed sensor, for the combined test runs, as given in Tables 1, 2, and 3. Figures 3, 6, and 9 show the mean altitude achieved by the model as a function of the mean time measured at the sensors for the combined test runs. Figure 10 is a composite illustration of Figures 2, 5, and 8 and Figure 11 is a composite illustration of Figures 3, 6, and 9.
The graphic representation of the data (Figures 10 and 11) shows that the control models (no piston) do not achieve significant forward velocity until after approximately 0.1 seconds into the engine burn. As expected, the initial thrust of the engine is essentially wasted. After 0.1 seconds, the control models accelerate rapidly. Both the 9" and 18" piston tubes show significant acceleration quickly after engine ignition, up to approximately 26 feet per second at 0.04 seconds. This corresponds to an altitude of approximately 6-7" for both piston tubes. At this point an inflection is noted in both curves. Little, if any, increase in velocity is noted for the 9" piston tube until approximately 0.09 seconds into the engine burn, corresponding to an altitude of about 21-23". At this point a gradual increase in velocity occurs, reaching approximately 32 ft/sec at 0.1 seconds, with a corresponding altitude of about 28". No increase in velocity is noted for the 18" piston tube up to approximately 0.07 seconds and an altitude of approximately 15". At this point the model has not yet separated from the piston tube. A gradual acceleration is then noted, achieving a velocity of approximately 35 ft/sec and a corresponding altitude of about 32" at 0.1 seconds. The altitudes achieved by the 9" and 18" tubes at 0.1 seconds are in the vicinity of sensors 11 and 13. Our visual observations indicate that the models for both the 9" and 18" piston runs were well separated from the piston tube at this time in the engine burn. The mean velocity for the 18" piston tube is approximately 9% greater than that for the 9" piston tube measured at 0.1 seconds. Given the standard deviation in the average velocities for the individual test runs, it is doubtful that a significant difference in performance has been demonstrated between the 9" and 18" pistons. The curves show, however, that at least a seven inch piston is desirable, with perhaps a nine inch piston insuring an adequate safety margin. It is important to note that interpretations of this data should be restricted to the design parameters of the floating-head piston, the engine impulse, and the launch mass represented by this experiment.
The objective of this study was to empirically evaluate the parameter of piston tube length in relation to performance for the floating-head piston design and to determine if an optimum piston length exists. In addition, we wished to develop a direct evaluation technique capable of monitoring performance in the form of velocity during the critical period of engine burn time. The results show that for the conditions employed in this experiment (a typical 1/2A streamer duration model), at least a seven inch piston tube is required to gain the full benefit of the floating-head piston launcher. A nine inch piston tube provides a margin of safety and no significant benefit is gained from piston tubes of greater length. In addition, the speed measurement device developed proved capable of monitoring piston performance in the form of velocity versus engine burn time, as desired.
The improved version of the speed measurement device demonstrated in this report can be useful in future piston launcher research (as well as in other researches). Other floating-head piston design parameters such as piston tube length and diameter and piston head length and weight may be evaluated and their relationship to launch mass determined. Further refinements to the speed measurement device, making it more precise and easier to operate, may also be devised.
The total expenditure on this project was approximately $120.00, broken down as follows:
* IR sensors $35.00 * Miscellaneous electronic items 25.00 * Photos 12.00 * Seven packs of Estes 1/2A3-2T 15.00 * Parts for piston/fin guidance apparatus 27.00 * Copying/typing costs 5.00
All other items cited on the Equipment List, such as the computer equipment, television, measurement devices, and lathe were items available at hand.
Atari 400/800 Hardware Manual. Sunnyvale, CA: Atari, Inc., 1980.
Chadwick, Ian. Mapping The Atari. Greensboro, NC: COMPUTE! Publications, Inc., 1983.
Scanlon, Leo J. 6502 Software Design. Indianapolis, IN: Howard W. Sams & Co., Inc., 1980.
Thoelen, Bauer, & Porzio. "Optimization of the Zero Volume Piston Launcher", NAR Technical Review, Vol. 2, No. 1, Fall 1974, 1974.
Weiss, C. and Vincent, J., "The Floating Head Piston Launcher", unpublished, NARAM-28, 1986.
09/14/98 - Sorry, but due to space considerations, the large
versions of these photos are no longer kept online
(hey, you had nine months to grab them... :)
Send an email to me (jeffvincent at verizon dot net)
requesting FHP2 pictures and I'll send back a 610 K
packet with the photos.
Photo 1 - Top view of piston guide tubes and fin guide rails.
Photo 2 - Overall view of detector side of the speed measurement device.
Photo 3 - Overall view of emitter side of the speed measurement device.
Photo 4 - View of the emitter relays and alkaline D cell.
Photo 5 - Close-up of the detector side, showing the control ignition sensor.
Photo 6 - Close-up of the emitter side, showing the control ignition sensor.